0:13
Hey, hey, hey. Come on, my Heat. Heat. N. Hey, hey, hey. Heat. Heat. N. Heat. Heat. Heat. Heat.
Search this video
3:15
We came to this to reach the stars.
3:26
We came to this world to see a future.
3:36
We came to this world to reach the stars. to shape our future.
3:52
We came to feel the new Heat. Heat. N. team deathmatch. The team 5. >> You're watching TVPN.
5:02
>> Today's Monday, December 1st, 2025.
5:02
We are live from the TVPN Ultra Dome, the temple of technology, the fortress of finance, the capital of capital ramp. Time is money. Save both.
5:11
Easy to use corporate cards, bill payments, accounting, and a whole lot more. All in one place. We have a special guest.
5:18
special guest today opening the show with us, Albi from the Land Down Under.
5:22
>> Probably saw him go viral recently, but uh why don't you introduce yourself? >> Yeah.
5:27
Um so I just I actually arrived in LA on Saturday. >> Welcome.
5:32
>> Um >> yeah, I'm from Sydney or Wllingong, so about an hour and a half from Sydney.
5:37
>> Um and yeah, I've been building something called finle which is basically dualingo for life skills.
5:41
Um, and I just applied to YC as well with that post on X.
5:47
Um, >> how many views did you get on the application video?
5:51
>> I think it got like 7. 8 million. So, yeah.
5:55
>> Hit the Let's hit the gong. Pretty viral for Albi. >> Well done. Well done.
6:02
>> So, uh, give me an example of a life skill that you can learn with your app.
6:06
Yeah, I guess like um entrepreneurship especially um startups and stuff cuz like in Australia, I don't know about in the US, but school is very entry level. It's not hands-on.
6:16
I feel like it's very um >> uh just not preparing us for life.
6:20
Like you do need it for if you want to be a doctor or a lawyer or something, but some kids don't want to do that.
6:26
Um >> and like yes, you do have commerce and computer science and stuff which I am doing as electives.
6:32
um but they're not hands-on and they're very like outdated and like textbook heavy.
6:39
>> Um so I feel like actually learning uh life skills that can that you can apply now especially like with AI and everything like if you don't know how to use AI now you're sort of going to be left behind. So >> very exciting.
6:53
What are you what are you hoping to get out of your trip?
6:54
You're on summer holiday right now.
6:55
Um, well, like my exams just finished before I came, so there's still like two weeks left of school.
7:02
But >> how do you think you did?
7:05
>> I think I got I got like a B in science and like >> there you go. There you go. >> A B in math.
7:11
>> Focus focus on the game. Focus on the game.
7:14
There's really room to grow.
7:14
But >> um yeah, I guess what I'm trying to get out of it is just to like meet as many people as possible, make as many connections as possible because this trip probably won't a trip like this probably won't happen again for a while.
7:27
>> Um so yeah, that's sort of my goal.
7:30
>> What's the status of the YC application? You've submitted it.
7:33
>> Uh yeah, >> have you heard back yet? >> No, it hasn't.
7:34
It's still like >> I'll need to recommendation.
7:37
>> We got a Yeah, we can we got we got a long >> If you're a YCL watching this, please go leave a recommendation. >> Yeah.
7:42
Um but uh congratulations.
7:42
Thanks so much uh for coming by.
7:46
Uh what what is the stage of development of the actual application? The the product itself. Are you live?
7:52
Can people go download demo right now?
7:54
We're getting like beta testers.
7:56
Um but the beta should um should be launching soon, probably by like the end of this year.
8:02
>> Um >> do you have a wait list?
8:03
Are you doing email capture yet?
8:05
>> Yeah, like weight list beta testers.
8:07
We've got like a couple hundred, but yeah. >> Very cool. >> Incredible.
8:10
Well, congratulations on uh all the attention.
8:13
I'm sure you'll convert it into a lot of opportunity and uh have a great trip. >> Yes. Great to have you.
8:19
>> And good luck with the YC application.
8:22
>> Looking sharp in the suit.
8:23
>> Looking sharp in the suit. >> Amazing.
8:25
Have a good >> Thank you.
8:27
>> Thanks for stopping by.
8:27
Uh before we move on to the rest of the show, let me tell you about Reream.
8:31
One live stream, 30 plus destinations.
8:33
If you want to multiream, go to reream. com.
8:38
And uh it's been three years since chat GPT launched.
8:40
I wanted to reflect a little bit.
8:42
Everything changed or maybe nothing changed or maybe some amount of change in between everything and nothing.
8:48
You're more on the nothing changed camp.
8:50
I sort of agree with you.
8:53
I was sort of I was sort of reflecting on like, okay, Thanksgiving's happened.
8:57
It was Thanksgiving over the weekend.
8:59
You know, how different is my world?
9:01
Like there's not a humanoid robot that's cooking for me.
9:03
And also, even if we had a humanoid robot, I think that would I think Thanksgiving would be the day we let the robot sit in the closet cuz we enjoy No, no, let us cook. We enjoy cooking.
9:13
Cooking is a fun family experience.
9:15
And so, of all the things, >> let us >> Thanksgiving is like the track day of cooking.
9:20
Like even if you have the robot that does it, you still want to do it on Thanksgiving.
9:24
You don't want to cook on a random Tuesday when you're busy, you got lunch, you know, all this other stuff.
9:29
Uh Thanksgiving is the is the Nurburg Ring.
9:31
And I was doing some dishes after Thanksgiving and I felt like it was a good way to kind of like it felt like walking off the pie in a little way.
9:38
I wasn't walk wasn't walking very far just kind of back and forth.
9:42
>> And so yeah, so that that that hasn't really changed that much for me.
9:44
Um I was thinking I was reflecting more on the agentic commerce thing.
9:48
It feels like chatbt and openai they really are pushing to make revenue from agentic commerce like in this holiday season.
9:54
commerce like in this holiday season. uh and uh incredible speed of execution like clearly it's a big opportunity if you can figure out how to you know run ads commerce convert take a cut of that that's big um my experience actually demoing it it was kind of interesting like the actual product in ship is pretty good but you can see that the
10:14
walled gardens are already going up so uh one place that I like to go to for reviews of products specifically around the holidays is the wire cutter now the wire cutter their whole twist was they wouldn't rate each product, what they would do is they would pick a category and then they would just tell you what their best product was in that category. Sort of like a cluster max of of
10:32
Sort of like a cluster max of of vacuums.
10:35
So, they would give you the platinum tier uh vacuum and then a budget pick.
10:40
And so, I've always liked the Wire Cutter.
10:42
I think they do a very rigorous job.
10:43
Uh they were acquired by the New York Times.
10:46
The New York Times is currently in a lawsuit with OpenAI.
10:49
And so, if you go to Chad GBT and say, "Hey, >> and I think they're about to be in a lawsuit with David Sax, >> maybe.
10:54
maybe which we will talk about on the show in a little bit.
10:56
Um but if but but but if you go so I went to ChachiP and I was like hey okay pull a deep research report just pull everything from the wire cutter and uh and tell me every category and every product that's top ranked because then I can just scan it really quickly and be like oh yeah I didn't even remember that that category existed.
11:14
That would be a great gift.
11:16
I'll get it and I'll go through the I'll go through the wire cutter link. I'm fine with that. I'm paying ChachiBT.
11:19
I'm happy to go and use their affiliate link on the wire cutter.
11:23
That's how the wire cutter monetizes.
11:24
Uh, but it couldn't do it. It couldn't do it.
11:26
It said, "Hey, we don't we can't touch the wire cutter. Like, it's off limits.
11:31
>> You got to head over there yourself.
11:31
Get uh pop open a Chrome tab, brother, if you want to get over there." Like, that's on you.
11:38
>> Or or maybe an Atlas tab. I don't know.
11:40
But, uh, so so so that that had not really changed that much for me.
11:43
Um, but the one thing that did really change on Thanksgiving was the discourse.
11:48
Like the AI narrative has fully arrived to just family and friends.
11:55
>> You mean in in the in the home? >> Yes. Yes.
11:57
In people that don't work in technology that don't their job is not >> their favorite trough.
12:03
>> Not that more talking about is it a bubble?
12:05
Where do you think all this stuff goes?
12:07
Um the stuff that you know we've been talking about. >> Sure.
12:10
>> You're not living in a bubble.
12:10
You think the average family in America is talking?
12:14
>> I saw I saw multiple newsletters where the whole conceit of the newsletter going into the holidays was how to talk to your family about the AI bubble and how to how to talk to your family about AI generally.
12:23
And I think it's real because if you've been watching your 401k over the last year, you've seen a massive spike and then a recent selloff and if you've turned on any news or opened up any newspaper, you've been hearing about $1 trillion and you're like, "What?
12:40
like, "What? a trillion dollars that chat GBT app they need a trillion dollars to make that thing work right chat uh and so and and so it is it is a really big narrative and so I wanted to reflect on like what has actually changed over the last three years uh and
12:56
specifically in the Mag7 the Mag 7 has been on absolute tear uh just over the last three years the value as a whole has basically tripled it was uh a little under $8 trillion now it's over $21 trillion it's a lot value created in the last three years. Uh Nvidia was second to
13:12
Uh Nvidia was second to last in the Max in the Mag 7 when CHAGPT launched.
13:19
It was worth just $420 billion something there.
13:23
>> Today the stock is up over 10x. Basically, it's four $4.
13:25
36 trillion >> and up today.
13:30
>> And up despite all the chaos, >> Dylan Patel was trying so hard to bring that stock down, but he couldn't do it.
13:36
He's coming on the show at noon.
13:36
We're going to confront him about his bare posting and whether or not the market >> funny enough Broadcom is down today. >> Okay.
13:43
Why is >> that the maker of the TPU? >> Oh yeah.
13:47
Well, I mean a lot of these things it's like it's already been priced in.
13:49
I mean even when you read the semi- analysis uh uh uh piece, you know, a lot of it's like we've been writing about this for months.
13:56
People have already put this trade on, etc. , etc.
13:59
Um, but I do think that uh the the Nvidia the 10x that's happened has really created some crazy zealots and just an entire industrial complex because there are so many people who put who who heard AI they tried the chatbt thing and they were like this is big how
14:19
do I get it on this I can't buy o open AAI open AI is running away with it oh they need Nvidia chips that's the logical next step they went in Nvidia and and they got a 10x and they could have gotten a 10x on like a million, $10 million. Like there's no amount of money
14:33
Like there's no amount of money cuz it's it was it was already a $420 billion company.
14:37
So you could be you could put your entire retirement savings in it. No problem.
14:41
Complete liquidity, right?
14:42
It's not oh you got to get some SPV. It was really easy.
14:45
>> Siki Chen from Runway was saying that back I think it was 2020 2021 he he said he put an uncomfortable amount of his net worth into Nvidia. Yep.
14:51
And obviously uh >> and uh in in near cyan same same story right >> still underappreciated the the Nvidia 10-year fund all it does is buy Nvidia you just by investing in it you can't possibly sell >> God's chosen company. >> Yes.
15:09
>> That's what the I think title of the fund was. >> Oh really? >> Yeah. >> That's hilarious.
15:12
Um and so I mean yeah there's been a ton of zealots.
15:15
We're going to talk to Dylan Patel at noon about some of the zealots that have been attacking him.
15:19
Um, previously uh the world's largest company in November of 2022 was Apple and at the time they had a sizable lead over Microsoft, Amazon and Google.
15:30
Now that gap has closed a bit as the hyperscalers have grown more over the last three years on the back of the AI boom.
15:36
Uh, and it's interesting.
15:36
I mean, you can sort the Mag 7 by uh by market cap and today you get the following ranking.
15:44
Tesla, then Meta, then Amazon, Microsoft, Alphabet, Apple, and then Nvidia at the top.
15:49
And the big question, I think that's on everyone's mind and kind of underpins the horse race that we that we cover every day on this show is uh what will that ranking look like in the next three years?
15:58
Um is Nvidia really a monopoly?
16:00
Is it is it impervious to attacks from the uh from you know, different suppliers?
16:08
>> What does Broadcom have to do to get into the Mag 7? Uh >> I don't know.
16:12
Tesla's sitting at 10 on the market cap. Yeah. Uh companiesarketcap.
16:16
com which we are not affiliated with which is just a fantastic website. >> Fantastic.
16:22
>> Uh Broadcom is sitting at number six above uh Meta currently. >> Um >> I don't know.
16:27
I mean I think uh I think you know several years in the $1 trillion club like just being you know undeniable at that uh at that scale.
16:39
There's also just like a bit of branding like um some of the companies that made it into the Mag 7 were I feel like the Mag 7 leaned understandable like not that deep in the supply chain.
16:51
Even Nvidia was the deepest.
16:52
Nvidia had the least of like a consumer brand but still a lot of people used the gaming graph gaming graphics cards.
16:58
Broadcom is really tricky because there's no consumer angle whatsoever.
17:02
Consumers can buy Tesla, they can use Meta products, they can buy on Amazon, have a Microsoft, you know, operating system, they can use Google, they can have an iPhone, and they can have an Nvidia gaming graphics card. the top 10 right now.
17:16
Tesla sitting at 10, TSMC at 9, eight is Saudi Ramco, seven Meta, and then six is >> I also think you have to be an American company to be in this like mag seven or whatever the hot ranking is like Fang.
17:29
Uh Fang did not include um ne never included oil companies, never included international companies because if you go there, then you could be like, "Oh, well, let's include like the Chinese tobacco company that's worth a trillion dollars or something like that."
17:40
Like like there are some crazy there are some crazy like foreignowned companies that are if they were independent might be worth a trillion dollars because they just have so much of these assets. Yeah, exactly.
17:52
But it doesn't really count because it's just sitting there out in the out in the ether.
17:55
Uh well um let me tell you about Gemini 3 Pro.
17:59
Google's most intelligent model yet.
18:02
State-of-the-art reasoning, next level vibe coding, and deep multimodal understanding.
18:06
And speaking of that, Buco Capital Bloke has a post here.
18:09
Gemini app downloads are catching up to chat GBT and Gemini users now spend more time in the app than Chat GBT users.
18:17
People are going back and forth on can Gemini catch up.
18:19
Uh you know the model clearly very good.
18:23
Uh the big bombshell in the semi- analysis piece over the weekend was uh the this idea which I think has been bandied about before this idea that uh OpenAI has not done a proper pre-train since 40 and the 45 pre-train kind of got mothballled and so but uh there was this question about like is pre-training dead.
18:43
Seems like the Google folks said no it's not and then they went and did a pre-train and and Gemini 3 outperformed.
18:49
Uh Anthropic also pre-trained. >> Yeah. I mean, yeah.
18:53
I mean, we asked Cholto about this and he said, >> "Oh, yeah, we're still bullish on scaling."
18:59
>> And he I I think actually like Cholto kind of like in the subtext said like the >> the reason uh Opus45 was good is not because it was a new pre-chain, it's because it was RL. >> That's interesting. That was your reading. Yeah.
19:11
>> I I I feel like there's still there's still juice in the lemon of pre-training, but it's not scale.
19:15
Like we only have one internet.
19:17
Ilia was correct about that.
19:19
It's not scaling the size of the pre-train, which is what happened with 45 uh from uh from GPT4. 5.
19:28
That was just bigger, I guess.
19:28
Um but but does seem like there's little optimizations that you can do on the on the pre-training, but I don't know.
19:36
We'll have to dig into it.
19:36
Uh but but I I I think the thing that no one is debating is the fact that the Gemini 3 as a model with Nano Banana Pro with V3 is just like the actual foundational intelligence is plenty good to be dominant in the consumer AI category.
19:55
The question is can you actually get people to install the app, use it, do can they enjoy it?
19:59
Do they not churn and go back to CHBT?
20:01
I've been fighting back and forth left and right going into one app and the other.
20:06
I was getting a ton of disconnect errors with the Gemini app.
20:10
Even though the model's great and there's some really cool features, I found >> you need to catch up on the product side. >> Exactly. Yeah. Yeah. The product side.
20:16
And so a lot of people are saying like, "Oh, Gemini team should just the app team should just go and, you know, copy Chat GPT's homework and and you know, copy all these little features."
20:23
I've put out a post uh that the folks over at the Gemini team actually uh you know, did turn into bug reports and I think are working on.
20:30
Um, but it really does seem like it's uh it's it's a really it's a sprint to actually create an app that is as sticky as Chad GBT because Chachbt the app is fantastic and and very very very well designed.
20:41
And so um the uh >> yeah and there's some reporting uh from similar web is what the FT is using to uh track average user minutes.
20:51
I always find those hard to I mean it's it must be like Neielson ratings where they're like polling people or something because you can't get a pixel in open AI like you can't get a pixel into the Gemini app.
21:04
>> And are they counting user minutes if a tab is open but I'm not actually in >> and is this just desktop because that's like completely separate from mobile use >> desktop and mobile web which I again I don't mobile web. >> I I don't know.
21:15
I wouldn't read too many too much into this data specifically.
21:17
too much into this data specifically. I would I would much more look at like what are the structural advantages that we know exist and I mean with Gemini one of them is to that point about the wire cutter you know you know where the wire cutter shows up Google search results you know what you know what what company
21:33
has one bot for scraping everything Google so the Google bot is uh identifies as one entity so you can either say I'm allowing Google or not and it's a big it's a tall order to be like yeah I don't want to be in Google results And so you a lot of companies are saying, "Yeah, I'm good with Google showing up in Google results, but that also shows up in AI search results." Um,
21:55
Um, and they can and and there are things that companies can do to say, "Hey, don't put me in the Gemini, you know, like training data set necessarily, but in terms of just actually showing up.
22:07
You've seen it in the Google in the Gemini app, it says using Google search."
22:11
And so if I go to Gemini and I say, "Hey, head over to the wire cutter.
22:15
Find me the best vacuum cleaner."
22:15
Google probably can do that. Gemini probably do that. Yeah, test it.
22:19
Um whereas whereas OpenAI is is in a fight with the New York Times, whereas Google and the New York Times, like they might not love each other, but they definitely have like a like an uncomfortable truce, right?
22:32
>> A funny a funny uh Gemini integration that I that I've used is that you land in a in a hangout and you just say, "Who is this person? >> Is this real?"
22:40
You can actually do that. >> It is. It pulls up a sidebar.
22:42
You can just ask like, "Who who am I meeting with right now?"
22:46
now and it'll give you like a uh >> it's clearly who am I meeting with? What should I say?
22:50
What what should I say to them?
22:52
>> What should I ask them? >> What is my name?
22:53
What are what do they want to know about me?
22:55
What what should I tell them about me?
23:01
>> Okay, so Gemini was able to pull wire cutter recommendations. >> Yeah. >> I I don't know.
23:05
This is I feel like I feel like um >> Yeah. >> Uh I wonder Yeah.
23:10
I wonder if if if Wire cutter is actually benefiting from this in any way yet.
23:18
>> I mean, I'm assuming I'm not sure because Google hasn't Gemini hasn't rolled out the agentic commerce stuff that would actually like scrape out the referral uh the referral token.
23:26
And so if I'm if I'm in Gemini and I'm saying I'm going to do some agentic shopping or whatever and I say, "Pull me the best vacuum cleaner from the wire cutter."
23:34
It goes over and does that.
23:36
And then I land on the wire cutter and then I click that link.
23:40
That should give the wire cutter the credit.
23:42
Now, if I as a follow-up prompt go in Gemini and say, "Okay, great.
23:48
The wire cutter told me the best vacuum cleaner is from James Dyson, of course, is the Dyson.
23:52
Uh, find me the Amazon link."
23:55
Well, Gemini's probably not given the wire cutter the attribution at that point.
23:59
It might even be taking its own attribution.
24:01
I don't know exactly how it's how it's functioning right now, but I would imagine that that link does not get reinstantiated as the as the Wire Cutter uh affiliate link.
24:10
And so, um, we could see, I mean, these are all like going to be pretty existential questions for the SEO crowd, anyone who's monetizing off of SEO.
24:21
We saw some uh some screenshots that apparently site traffic to Vox properties is down 50%.
24:25
Um, and I don't know if that how how much how much of that is just the shift to social media versus the shift to >> Yeah.
24:34
How much is it their business strategy just being like, hey, we want to do more video. Yeah.
24:37
and that'll be distributed off our site for the most part.
24:41
>> I think a lot of people generally do not they consume more and more content on social media platforms.
24:46
They go from YouTube to their RSS player to audiobooks to Twitter to Instagram and they kind of bounce around from one and the other and then every once in a while they will go in and and actually land on a particular site.
25:00
>> Uh like you can if you go to tbpn.
25:00
com you can get our newsletter in your inbox every morning.
25:05
Uh, and we can you can also sign up for cognition.
25:07
They're the makers of Devon, the AI software engineer.
25:10
Crush your backlog with your personal AI engineering team.
25:12
Well, speaking of the New York Times, uh, David Saxs is going to war with the New York Times.
25:20
he says inside the NYT's hoax factory.
25:22
He calls it a hoax factory because the New York Times posted a a piece about David Sax saying that uh the the headline was Silicon Valley's man in the White House is benefiting himself and his friends and Ryan Mack uh was going back and forth with Shawn Maguire or Yeah. Yeah. Shan Magcguire.
25:42
Um Ryan Max says, "Today has been a good example of what X has become.
25:46
complaints from a subset of wealthy tech folks about a story that circulates more widely than the actual story itself.
25:53
Musk bought the platform to control the message and he and his friends are getting just that.
25:59
And Sean Magcguire says, uh, you don't get to run this headline, then write an article that doesn't validate the claim and then get away with playing the victim.
26:06
We see the we see through the ruse.
26:08
And so, uh, David Saxs has responded in full to the NYT's hoax factory.
26:14
He says, "Five months ago, the five New York Times reporters were dispatched to create a story about my supposed conflicts of interest working as the White House AI and cryptozar.
26:24
Through a series of fact checks, they revealed their accusations, which we debunked in detail.
26:27
Not surprisingly, they the published article included only bits and pieces of our responses.
26:31
Their accusations ranged from a fabricated dinner with a leading tech CEO to non-existent promises of access to the president to baseless claims of influencing defense contracts.
26:42
Every time we would prove an accusation false NT pivoted to the next allegation.
26:47
This is why the story has dragged on for five months.
26:52
Today they evidently just threw up their hands and published this nothing burger.
26:56
Anyone who reads the story carefully can see that they strung together a bunch of anecdotes that don't support the headline.
27:02
And of course, that was the whole point.
27:03
At no point in their constant goalpost shifting was NT willing to update the premise of their story to accept it.
27:10
I have no conflicts of interest to uncover.
27:11
No conflicts of interest.
27:14
Uh as as it became clear that NYT wasn't interested in writing a fair story, I hired the the law firm Claire Lockach, which uh specializes in defamation law.
27:26
I'm attaching Clare Lock's letter to the NYT.
27:27
So readers have full context on our interactions with NT reporters over the past several months.
27:32
Once you read the letter, it becomes very clear how NT willfully mischaracterized or ignored the facts to support their bogus narrative.
27:39
So >> Will says, "Hiring Claire Lock for this is sick.
27:43
Cruz missile to blow up a straw hut."
27:48
>> He's a big fan of uh of litigation. He loves litigation.
27:50
Uh well, people have been uh supportive of this broadly in tech. Let's go.
27:55
Let's go through some of the reaction.
27:57
Um, Sam Alman says, "David Sax really understands AI and cares about the US leading in innovation.
28:02
I'm grateful we have him." Brian Armstrong. >> Yeah.
28:07
Here's here's here was uh here's my takeaway. >> Yeah.
28:11
If you believe that AI and crypto are are are uh industries that we should support in the United States, then you want to have a a ZAR focused on those things that uh generally feels positively about those things and and uh wants to create the best possible environment for those industries to thrive in the US.
28:34
I think that uh there's actually a debate on both fronts, right?
28:38
Like there's people on the left that think AI and crypto are just default bad. They want less of them.
28:45
>> Uh and there's people on there's people on the right that believe that too.
28:47
But uh I think that uh ultimately uh there's arguments for why the US should lead in in stable coins which uh you know is is part of the >> uh part of why the Genius Act is is important and uh a lot of you know the AI action plan.
29:05
uh there's going to be debates on on individual points in that but in general I think uh you know creating an environment in the US where we can continue to lead an in AI is is important.
29:16
So um I think uh uh there wasn't uh I didn't see any sort of like smoking smoking gun in any of this stuff.
29:25
There were some allegations around uh around the all >> I don't think they smoke very much at all.
29:29
I think it's mostly tequila drinking. >> That's true.
29:32
>> They do all in tequila.
29:32
Uh although although JCal does tow a gun regularly. Oh yeah.
29:37
So maybe that's the smoking gun. >> He's a Texan. >> Uh yeah.
29:40
No, I I didn't see anything uh very specific.
29:44
I mean it's it's >> it's all in like they're they are super connected if you partner with them in some ways.
29:52
Like you would expect to get more of a read on where they're spending time in DC, what they're seeing.
29:55
that seems like there are clear lines on what you can share like what what turns you into a lobbying firm and what doesn't.
30:06
Uh I think that they've stayed out of becoming a lobbying firm and so they have clear clear rules on that.
30:10
Uh yeah, I think Bos distilled it uh pretty well.
30:15
Uh before we read his post, let me tell you about ADIO, the AI native CRM.
30:19
Adios, scales, and grows your company to the next level.
30:21
Uh boss said, "I don't know David Saxs, but I want more expertise in government.
30:24
Experts tend to have made money in their area of expertise, have friends in their area of expertise.
30:30
If our pe if people can't have history or friends in a field before leading it, then our leaders won't know anything.
30:37
And I thought this was a good distillation of like the core debate about like should you have someone who has never participated in in in an industry overseeing it or should you like someone who's purely academic, purely outside of it?
30:51
>> And I believe there's some readers >> Yeah.
30:53
and probably people at the New York Times that would like somebody that hasn't participated in either industry to to be running in a role like that. Yeah.
31:02
And just blanket against both industries and sort of like hold them back.
31:05
>> So so the so the reaction is interesting in the comments.
31:08
I mean first top comment is somebody like beefing with Bos over how he ran the Quest store.
31:14
It's like clearly a VR afficionado who like has a axe to grind over niche VR policies.
31:21
Uh, but the second post is what I want to get to because it actually addresses the core claim here.
31:27
And Alex says, "The the construct you're thinking of is called a council.
31:29
It's been used for a long time to allow the elected with limited knowledge on a domain to get a consensus of options from a range of experts.
31:37
This minimizes conflicts and prevents kleptocracy."
31:39
But like, isn't that what aar is?
31:41
I thought I I thought I thought I thought he Sax was a council.
31:45
Like, he's not he's not an elected official.
31:47
like the the elected official is Donald Trump, the president, and like there's a variety of folks there and then and then uh Sax is like appointed to this Zar role that is just to give his like his like he he doesn't have the right he doesn't have the ability to just like create legislation out of thin air, right?
32:06
Like he he is he is very much a >> I was trying to look up the history of Zars, right?
32:11
Uh >> it is weird is like have we always had Zars?
32:14
I know there was a whole thing was Bernard uh >> Barak appointed by uh President Woodra Wilson to head the war industries board in 1918.
32:25
The press dubbed him the industry ZAR because he had sweeping powers to coordinate wartime production.
32:31
During World War II, President Franklin D.
32:33
Roosevelt appointed several ZAR to manage the massive wartime economy, including a shipping zar and a synthetic rubber zar.
32:39
rubber zar. These rolls were essential >> synthetic rubbers are >> one of the most iconic >> stoked for that >> people don't talk about the need for our ongoing need for synthetic rubbers are no >> these roles were essential because existing government bureaucracies were
32:54
too slow to handle the urgent demands of total war uh the during the Nixon era the modern concept of thesar a policy specialist with a specific portfolio solidified under uh under under Nixon uh during the 1973 oil crisis Nixon and appointed William Simon as the energy ZAR to manage fuel shortages. Uh he also
33:13
Uh he also uh had uh a drug zar during uh the uh sort of like beginnings of of the uh war on drugs.
33:26
So anyways, uh again, I think uh unless you're just blanket against these industries, it's hard to argue uh that you want somebody that doesn't have any expertise in said industries.
33:41
>> Yeah, some of these some of these claims here here's one uh it's sort of hard to to track.
33:46
Like so he says, "Free free from those this is from the New York Times from the actual article screenshot.
33:51
Uh free of those restrictions, Mr.
33:53
Sachs flew to the Middle East in May and struck a deal to send 500,000 American AI chips, mostly from Nvidia, to the UAE, the United Arab Emirates.
34:03
The large number alarmed some White House officials who fear that China, an ally of the Emirates, would gain access to the technology, these people said.
34:11
But the deal was a win for Nvidia.
34:12
Analysts estimated that it could make as much as 200 billion from the chip sales.
34:18
And so like I I I like we we've covered the debate around export controls and should Nvidia where should Nvidia be able to sell things, but um it's never been an openandshot case in my mind.
34:32
It's never been like, oh, it's so obvious that the UAE is completely off the table. >> Yeah. >> I don't know.
34:38
Yeah, it was I mean it was also just like painting painting the friendship between Sachs and Jensen uh as like something that that felt wrong was was a little bit rough considering it's the most valuable company in the world. Yeah.
34:53
>> One of the most important AI companies potentially the most uh important AI company if you just go by uh weight in the uh in various uh indexes.
35:01
So >> I don't Yeah, I don't I mean, it's like it's clear that he doesn't have Nvidia bags directly.
35:07
Like that's completely debunked.
35:09
So, so you have to do these like 25 different steps to get to some sort of conflict.
35:14
Um, it's a lot of like, you know, >> I read this and I think like this is >> if you're if you're the average New York Times subscriber, >> Yeah.
35:24
>> this is probably that you were they were probably like very excited by this story, right? >> Yeah.
35:30
Yeah, I mean a lot of I I think a lot of people um are are definitely like uh yeah just riled up by the All-In podcast.
35:37
>> Charlie in the chat says Allin Pod about to be an all-time after this article.
35:39
Do you think it's possible that David uh and Jason coordinated to get this hit piece done to grow allin even further?
35:50
They said we're at such an insane that was thing.
35:53
Yeah, Jason Jason said a bunch of I mean Jason made a lot of good arguments uh about this, but one thing was he was like we would be smaller if we what was it he was like he was like we would be bigger if we didn't talk about politics and that seems crazy to me.
36:09
I feel like politics is like the ultimate TAM expander in the history of podcasting and media broadly. >> Yeah.
36:14
The audience for political content is like 10 times less.
36:17
>> I would I would think so.
36:17
I I I do believe that that Jason loves talking about TAC and like I think he's I think he's an OG. He's he's an OG.
36:22
He's say he's said that multiple times.
36:24
Um but but I would be shocked if if if uh if politics was not a was not a TAM expander for uh for podcast broadly.
36:32
And then the other thing is that you said that they lost money on the all-in events.
36:38
I don't know how that's possible.
36:39
Like those events obviously they're like big budgets, but you know I would I would imagine that that like the sponsors can and the ticket sales they're not cheap tickets, right?
36:46
Um I would I would imagine that they'd be making money off that. I certainly hope so.
36:51
I mean, they've been running this thing for 5 years.
36:52
They It's incredibly valuable in the ecosystem.
36:54
They should be able to capture some value there.
36:56
>> Maybe they set up their own data center to sort of manage >> they're just underwater.
37:00
Yeah, we we decided to bring it back >> podcast production on prem and we we ordered we ordered 100,000 black blue really has us by the balls. >> It's rough.
37:11
Uh Martin Scral here says, "The Sachs piece illustrates the exact problem with the New York Times.
37:15
Voters specifically want this type of person, not a bureaucrat who has never worked a real job. in a con street.
37:20
So yeah, that's so the the issue and the reason I think this article was written is that New York Times v uh subscribers specifically want this type of article. >> Mhm. >> Yeah.
37:34
Um yeah, Whiskey Titans going back and forth here.
37:37
Did you miss the entire part of the article?
37:38
This isn't a quote, we can't have businessmen in government.
37:42
This is a we can't have the government officials who host government summits and sell access to the president for $1 million via their podcast business.
37:49
Um, and Martin Scarley says, "I doubt it was Sachs who wanted to sell $1 million passes."
37:56
And Whiskey Titan says, "I agree with you.
37:58
Uh, I'm sure it wasn't, but letting Jason run rampant until Susie Wild steps in isn't a great look.
38:03
I happen to think Sax is doing fine at this particular role, but I also understand the general public feelings like there's a lot of graft.
38:10
The New York Times isn't the right conduit for that argument, though, and they're going back and forth.
38:15
The timeline truly is in turmoil over this.
38:17
Uh Dan Primac had a good take.
38:19
He had a whole uh breakdown of this, which I think was interesting.
38:24
Um he said, let's let's kick this off.
38:27
But first, let me tell you about fall build and deploy AI and im AI video and image models.
38:34
They're trusted by millions to power generative media at scale.
38:38
So Dan Primac said, "Lots of people are sending me the New York Times story on David Sachs.
38:42
Outside of the all-in sponsorship proposal, which feels oblivious at best, corrupt at worst, I'm not seeing much in there that's new, at least to those who've been following."
38:56
Dan Primac says, "As an aside, it's true that Sachscraft still have a ton of AI investments.
39:03
Thing is, all tech investments at this point are AI investments.
39:06
It's kind of like internet investments at this point.
39:09
If you invest in tech startups, you de facto invest in AI startups.
39:12
And Jason says, "We've lost money on the event."
39:16
The NIT knew this and deliberately published false information.
39:18
And and uh Dan Primac says, "They included statement that you lost money on it.
39:25
What did they print that was false?
39:25
They were somehow make that that we are somehow making money in this or some gain."
39:31
And Dan Primac says, "Just rere reread. Just re reread.
39:34
doesn't claim that All-In made money.
39:36
Said you tried to generate revenue via $1 million sponsorships, including for VIP reception that didn't end up happening, but ads that you don't know what But ads that you don't know what sponsors ultimately paid or that it doesn't know what sponsors ultimately paid included the statement that you lost money. Am I missing something? And Jason says, "Mr.
40:00
Sachs has raised the profile of his weekly podcast all-in through his government role and expanded its business. Um, confused.
40:07
I thought you were talking specifically about the White House AI summit pieces.
40:11
Dean Premik talking in general.
40:12
Don't know how you would would not quantify Sachs role in White House defraised all-in profile at least among normies.
40:20
As for role in biz expansion, guess you could stake your claim there.
40:26
I completely disagree with this.
40:27
I feel like I feel like the All-In podcast put the White House on the map.
40:31
I feel like a lot of people were like they found out about the White House and about the US government because Exactly. Exactly.
40:39
Because of the All-In podcast.
40:39
They were listening to all podcast and they were like, "Wait, wait, wait.
40:42
You're telling me >> there's people in Washington DC and they run this whole country.
40:47
>> They create >> they're in charge of the rules. >> Yeah.
40:49
They create sort of laws and framework over how >> our country should operate, which which industries we want to, you know, support >> and grow.
40:59
>> You're telling me, you're telling me that there's a group of people and one of my besties is they're running it. This is amazing.
41:05
I got to learn more about this.
41:07
I got to figure out what a bill is.
41:08
I got to figure out what how a bill turns into a law. >> Chat, what is a bill?
41:12
Jason says, "If anything, going deep into politics has been a net negative for allin, at least in my opinion.
41:19
We would we would be growing faster and wouldn't have lost some percentage of our left-leaning audience if we'd stuck to tech, markets, science, VC, etc."
41:26
That's an interesting take.
41:28
I I I still think that politics. It made it so big. >> Well, yeah.
41:34
And it made it made the content polarizing, which but I think that polarizing in media is good.
41:41
>> You actually get more attention.
41:41
uh not necessarily good from all points of view, but good from a pure just like reach.
41:49
>> I mean, yeah, I was looking at the I think the I think the uh the the the ratings are like the amount of viewers for uh for like CNBC is bigger than Bloomberg by like a pretty significant margin because Bloomberg's like extra wonky and CNBC is a little bit I mean it's like literally called consumer business news.
42:09
Like that's what the C stands for I believe.
42:11
Um, and then and then you have Fox, which is even more like Fox News is political and it's much bigger ratings um than CNBC or Bloomberg.
42:21
And then ESPN is like by far the biggest. Yeah.
42:23
Because it's like sports. Everyone loves sports.
42:24
And so uh like maybe that's the final that's the final form.
42:29
They should go full poker and then full sports.
42:31
It should just become Sports Center competitor. >> I could see it. That might be the way.
42:35
>> Uh AB says, "I only learned about the Trump about Trump because Chimoth endorsed him." >> Yes. Exactly.
42:42
I had never heard of this guy.
42:45
>> Who who >> the vodka and social media entrepreneur?
42:49
>> He's running for president.
42:51
>> Okay, so Dan Primac is weighing in again concluding it.
42:54
He says this, the New York Times story was mostly a nothing burger, at least for those familiar with the situation.
43:00
As for hoax, the story itself as published isn't being disputed.
43:04
Obviously, the New York Times had info questions that Sachs lawyers answered, and disproven info wasn't included.
43:11
That's how journalism works.
43:13
The real complaint seems to be about the headline.
43:15
Quote, "Silicon Valley's man in the White House is benefiting himself and his friends."
43:19
I get the complaint, but that's not but it's really a matter of interpretation, not true, false, hoax.
43:27
Imagine if you had a friend and they went to the White House and they didn't try and benefit you.
43:30
You'd feel you'd feel like >> you might not be friends with them.
43:34
might not be friends with Sachs and Trump and the Trump White House are pursuing let them cook AI policy. I like that.
43:41
That they believe will help us win the AI race and that the rewards outweigh the risks. Others disagree. Yeah, this is so true.
43:47
It's like there there is no like oh like we now know the correct way to win the AI war.
43:52
Like we we we know that there's a correct way. It's very obvious.
43:56
It's like no, everyone's debating this constantly even inside of tech.
43:59
Uh and and Sax has one view that I think has actually played out pretty well considering that he's been anti-doomer, anti-fast takeoff, more industrial capacity, more you know, opportunity to grow GDP.
44:12
You know, there there are some elements of his takes that are a little bit more like TBDs, like what what actually happens to jobs over the long term, how does it manifest in GDP growth over the long term?
44:24
But so far I think he's been correct and and I think that's what Dan Primick's saying here.
44:30
He says only time will tell if Sax is correct.
44:32
What we know for sure though is that his deregulatory policies should help VC funds his those runs by his friends those run by strangers etc.
44:39
Thus the headline is defensible albeit pushing an agenda.
44:45
And that's and that's the timeline in turmoil folks.
44:47
Let me tell you about graphite.
44:48
dev code review for the age of AI.
44:50
Graphite helps teams on GitHub ship higher quality software faster.
44:56
Uh I can't read this AI amlicus this name.
44:59
Uh this Ilia interview will be compulsory viewing for any future student trying to understand what misallocation of capital looks like in real life.
45:06
See I I completely disagree with this take.
45:08
Um people were going back and forth on this.
45:10
Um we talked about this a little bit over the uh the over the holidays but uh fleeting in bit says can you say more just that he doesn't have any business direction or something else?
45:20
And uh and the the original poster says, "These are my intuitions, but for what it's worth on the micro level, he just seems to drift in a sea of possibility and not the kind of person."
45:30
>> See, I originally read this as uh as as >> the misallocation of capital that I've seen is like the the 10th, 11th, 12th foundation model lab that has like a hundred to a billion dollars that is just like kind of iterating on what Ilia already worked on.
45:52
and doesn't necessarily like if they just do >> if they just create a model that's not like there's going to be incredible value in that.
46:01
Meanwhile, >> I'm like, okay, you take the guy that that that whose work led to Chad GPT and you give him a few billion dollars and let him, you know, continue to iterate and and uh he's not just like, you know, firing a single can, you know, multi-billion dollar cannon and hoping he hits a target.
46:18
It's like this incremental research that uh I think is still one of the best shots at like developing the next paradigm, whatever comes after LLM.
46:25
So, >> um I I read this I I I think that your your reading of this was right, but I initially read it the other way and I was like, yeah, I do think this is, you know, somewhat somewhat bearish on the on the the incremental uh large language model lab. >> Yeah, I don't know.
46:42
I mean, I can I can kind of steal man both.
46:45
Uh we're gonna have uh Julian uh on the show in when is he coming on?
46:48
At uh or or we're having uh Vincent from Prime Intellect come on the show.
46:54
And uh I was talking to him.
46:57
We'll get more information from him.
46:58
He's going to be on at uh 140.
46:59
But uh Vincent was explaining that that more and more uh more and more companies and different business processes, they do need specific training runs.
47:09
They do need the the skill sets of a foundation model lab, but they're not but there's a lot of uh business to be done that's not purely AGI seeking, not purely paradigm shifting.
47:20
Uh so I do think that there's there's some value there if the business can be run well, which is a big if, but there is a there is a path where a thinking machines or or one of these companies is going to go and do specific reinforcement learning, specific model development for uh a specific company and task that can work out.
47:38
It's a very different business than searching for the next paradigm doing science and maybe you shouldn't even call it a lab because you're not really even trying to do foundational science necessarily.
47:49
You're more productizing >> company.
47:51
>> Yeah, it's a business which is great. We love that.
47:52
Uh what's what's interesting about Ilia is that um >> when we talked about this like it is a venture style bet like let the scientist go experiment maybe it will work out.
48:04
It's extremely high risk probably a zero but if it works it's huge right?
48:07
So the expected value is still high.
48:09
What's crazy is that we're doing a venture style bet at growth scale and it's just massive amount of capital uh for something that I think I think the consensus here is that it's it's either he solves it and it's incredibly valuable and leaprogs everything and is just amazing or it's just you do get lost and you get a sea you get lost in the sea of research and ideas and you never really produce anything.
48:32
So uh I love I love the high-risk uh bets.
48:34
I I just understand why people are saying like what at that scale that's a lot of money that's a lot of money.
48:41
Uh but that but that has been happening internally at Google for a long time.
48:45
They probably burned a lot of money on research projects hasn't been that big of a deal because they had the engine for it.
48:51
And if the if the investors are uh significantly diversified, they should be fine.
48:57
>> Y >> um anyway uh what what else is in the timeline today? Finn.
49:00
AI, AI, the AI that handles your customer support, the number one AI agent for customer service.
49:06
Uh, we did get a good meme.
49:06
We got a couple good memes.
49:10
>> Cody says, "When my wife asks what we should eat for dinner, but says no to my first two suggestions."
49:16
>> We are back to the age of research. I like it.
49:19
And then uh when she asks what I want for dinner from Beas Lord, the answer to that question will reveal itself.
49:26
I think there will be lots of possible answers. Very true.
49:28
Uh, it's a great great new meme template. I like it.
49:33
When my husband asks how many Amazon packages are still on the way, the answer to that question will reveal itself.
49:39
I think there will be lots of possible answers, but I think that's actually true.
49:43
Like, if he creates if he creates some new AI, like there's a bunch of different ways to monetize it.
49:49
We know this is a we know this is a fact.
49:50
Uh, of course, Ilia is is now joining the ranks of Yan Lun and Rich Sutton and Andre Karpathy of sort of industry legends that are uh more or less saying that scaling is over and LLMs are dead.
50:04
Um, you know, on the other side, Schultto is saying scaling maybe not over. So, we'll see.
50:13
>> This is uh >> what >> this post is.
50:16
Yeah, this post is great.
50:18
Scaling is over and LMs are a dead end. A you're sweet.
50:20
Scaling is over and LMs are a dead end.
50:23
Hello human >> human resources.
50:25
I love this meme template because it's like Yeah.
50:27
Yan Lun has been saying the same thing.
50:30
>> He says Yan says, "For the record, my current BMI is 24. >> This guy rocks. Very funny."
50:39
>> Uh I thought he would dropped the the meta uh tag on X by now.
50:41
Um but uh I guess he's still >> Oh, he's still wrapping them. Didn't he leave?
50:48
Oh, he's like he's like reporting to leave. >> Okay.
50:51
He's like on on his way out more or less.
50:54
>> Uh another 1 billion to SSI.
50:54
There's a bunch of this in in the SSI bucket.
50:57
Um let me tell you about profound.
51:01
Get your brand mentioned in chatbt.
51:02
Reach millions of consumers who use AI to discover new products and brands.
51:06
Um, of course we are having Dylan Patel on the show in 12 minutes and we should do a little bit of a run through of the drama on the timeline.
51:18
The timeline was in turmoil.
51:19
Um, lots of people uh very, you know, upset with semi analysis's latest post. >> How dare you?
51:27
>> How dare you take a They took a swing at the king, which was the name of their article.
51:32
They said TPUv7 Google takes a swing at the king.
51:35
The king is of course Nvidia.
51:37
Uh and they are asking is this potentially the end of the coupe anthropics?
51:42
They're talking about anthropics 1 gawatt TPU purchase.
51:48
The more TPU uh meta SSI XAI openAI anthropic buy, the more GPU capex you save, next generation TPU V8.
51:56
Uh and they're going into what the what the battle between TPU and uh the next generation GPU out of Nvidia will look like.
52:07
And uh this uh this upsets some people.
52:10
There's a lot of folks who are long Nvidia.
52:12
Either they have invested in Nvidia, they made a lot of money in Nvidia, or their their whole business is tied to Nvidia or AMD even.
52:18
Uh and so >> or they bought the local top a month or >> potentially.
52:25
There's a whole bunch of reasons.
52:27
Uh you could also just disagree with this and you could just think that, you know, uh that uh semi analysis their takeaways are wrong.
52:34
But I think it's a thought-provoking article.
52:36
I think there's a lot of data here.
52:38
They're extremely thorough and I think that they do leave you with uh a lot of new information that you can, you know, do with what you want.
52:47
And I think in general the response to this article was very positive, but there were some folks who were very upset by it and went uh and went all over the >> place and on accounts that uh that put a noun and then capital as their name. >> Yes.
53:03
>> And suddenly they're an expert on >> Yes. Yes. Yes. Yeah.
53:05
It was it was a little odd seeing the credentialism come out from the anons because like uh I I don't think we should get in the two can play that game uh camp. It's a little bit rough.
53:18
Um but there's there's a little bit of interesting stuff in here.
53:23
Uh I want to read through some of this.
53:25
Let's let's kick it off with the with the opening of the semi analysis article.
53:28
The two best models in the world, Othropics Claude OP 4.
53:29
5 Opus and Google's Gemini 3 have the majority of their training and inference infrastructure on Google TPUs and Amazon's Tranium.
53:38
Now Google is selling TPUs physically to multiple firms.
53:41
Is this the end of Nvidia dominance?
53:44
The dawn of the AI era is here and it's crucial to understand that cost structure of AIdriven software deviates considerably from traditional software.
53:55
chip micro architecture and system architecture play a vital role in the development and scalability of these innovative new forms of software.
54:00
The hardware infrastructure on which AI software runs has a notably larger impact on capex and opex and subsequently the gross margins in contrast to earlier generations of software where developer costs were were were relatively larger.
54:16
Consequently, it is even more crucial to devote considerable attention to optimizing your AI infrastructure to be able to deploy software.
54:24
Firms that have an advantage in infrastructure will also have an advantage in the ability to deploy and scale applications with AI.
54:33
And as I say, we've long believed that the TPU is among the world's best systems for AI training and inference, neckand-neck with king of the jungle, Nvidia. 2.
54:40
5 years ago, we wrote about TPU supremacy, and this thesis has proven to be very correct.
54:46
TPU's results speak for themselves.
54:49
Gemini 3 is one of the best models in the world.
54:51
And uh there's a very funny bit in here. I need to find it. Uh saving Oh yeah, here.
54:58
So, uh this is this is a very spicy line in here.
55:01
He says, "OpenAI hasn't even deployed TPUs yet, and they've already saved 30% on their entire labwide NVIDIA fleet."
55:09
This demonstrates how the Perf per TCO advantage of TPUs is so strong that you already get the gains from adopting TPUs even before turning one on.
55:21
And so basically what he's what he's explaining is that because of the competitive dynamic between Nvidia and Google with the CPU now uh you can use TPU as a stocking horse. Yeah.
55:34
And say hey if you don't cut your prices invidia we know that you have really high margins >> or not even cut prices but encourage an investment. >> Exactly.
55:42
And so that's what they're >> Nvidia would Nvidia would rather invest back into your business instead of cutting prices. >> Yes.
55:49
And so uh says we think the more re more realistic explanation is that Nvidia aims to protect its dominant position at the foundation labs by offering equity investment rather than cutting prices which would lower gross margins and cause widespread investor panic.
56:06
Below we outline the OpenAI and enthropic arrangements to show how Frontier Labs can lower GPU total cost of ownership by buying or threatening to buy TPUs.
56:15
And so OpenAI, Nvidia, the uh you know it was uh $22 billion per gigawatt the rest of the system.
56:22
So it's a $ 34 billion uh billion dollar per gigawatt expense to Nvidia, but Nvidia is doing effectively an equity rebate of $10 billion per gigawatt in investment.
56:38
And so how that works out is a 29% partner discount.
56:41
Anthropic has similar math but a little bit higher at 44% partner discount because Microsoft is paying for a piece of it.
56:48
And so it's an interesting it's an interesting thesis and uh it's unclear exactly like well you know if if if the claim that the the investors will panic if it was actually just lower gross margins.
57:01
Well, if you say the quiet part out loud like this and you have uh you know you you do the math to show that there is basically a discount that margins might be coming down uh because of competitive dynamics.
57:16
Uh does that wind up uh resulting in investor panic?
57:18
I mean certainly it didn't today. Isn't Nvidia up today? Right. >> Yeah.
57:22
Um, Nvidia is up 1% adding a casual, you know, what, 10 trillion or 100 billion or something? >> 10 billion. Quadrillion. >> Yeah. G gigajillion dollars. >> Yeah.
57:35
Just I mean, uh, I I And again, uh, we said this earlier on the show, but Broadcom is down almost 4% today, which I would have expected it to be the other direction given that, uh, to actually buy TPUs physically, you need to go through Broadcom. Yeah. >> Yeah.
57:54
So, a lot of people are going back and forth on, you know, you can semi analysis be trusted because they're writing about uh about, you know, Nvidia and and uh Dylan.
58:06
I think some people didn't understand that he was joking. Zephyr here has a post.
58:10
Uh Dylan is being tongue-in-cheek, but he's not wrong.
58:14
Nvidia was extremely dominant for the last three years.
58:16
As we saw in the stock, it's up 10x over the last three years.
58:20
New competitors will cause a reduction in market share and margin compression, but TAM is big, so revenue profits won't go down.
58:27
Uh 75% of GM is just unsustainable.
58:31
Hyperscalers will also use the cheap TPU's threat to extract better deals from Jensen.
58:37
Priority access for Ruben Fineman or discounts on GPUs.
58:40
Jensen called Altman and initiated the $10 billion deal after he saw the information about uh the information article about OpenAI testing TPUs.
58:48
And so, uh, this is in reaction to that, uh, that that point about OpenAI hasn't even deployed TPUs yet and they've already saved 50%.
58:58
>> Um, there's a a decent post here from just another pod guy.
59:01
just another pod guy. They say Dylan speedr runninging through all the learnings of sellside research industry capture pissing off IR execs gatekeeping info based on client tier difficulty scaling beyond single star analyst distorted MSN representation of your notes eventually spending too much time
59:17
marketing versus researching amazing biz content though obviously Dylan would would push back on he did on a lot of this stuff you if you actually read through the entire article um >> there's nothing in the article should in this article should be that surprising because so much of the article is just referencing old semi-analysis research. Some of which they did a you know sort
59:39
Some of which they did a you know sort of uh before the payw wall, some of which they did under the payw wall.
59:43
Uh but it felt felt like a kind of a culmination of everything that they've been saying for a really long time.
59:49
And I think that part of part of I think the surprise here is just how much um how much faster this conversation has really come to a head than people may have expected.
1:00:01
I think I think the at least um at least like surface level on the timeline.
1:00:07
I think people felt like the TPU threat was maybe like a 2026 2027 conversation versus being like it's a part of these buying discussions right now and and negotiations. >> Yeah. Yeah.
1:00:19
Um, the other buried lead in the article was of course about uh pre-training.
1:00:25
So there's a snippet in here.
1:00:27
Open's leading researchers have not completed a successful full-scale pre-training run that was broadly for a new frontier model since GPT40 in May of 2024.
1:00:38
2024. And you know this this is it's it's it's so interesting that this like if this was wrong you would imagine that there would be a whole bunch of reaction from open AAI people or like proxies or surrogates right people quote and be like that's just not true wow something
1:00:55
else is cooked but the fact that I haven't seen anyone re respond to this and say like oh this is this is wrong like we actually did uh not that not that like that's the north star for what the business is like the business's job is to create profit. Right? It's not to, you know, complete Right?
1:01:09
It's not to, you know, complete successful full-scale pre-training runs. That's not the goal.
1:01:14
Uh that's just something that they might do in service of making a better model, making a better product, but ultimately, uh it's whatever the customers want. Yeah.
1:01:23
>> And if the customers are happy with 40 level base pre-train and a bunch of reasoning on top, um that's fine.
1:01:27
So, um what what else is in the back and forth?
1:01:33
people are uh also I mean it really uh it does make me happy that we didn't go deeper into ranking people uh because the the it does feel like when you create a list of tiers and rank a bunch of people uh you're just creating a big bucket of enemies down at the bottom of like people who want you dead because you rank them low.
1:01:52
Um but uh I'm sure we'll get into the discussion of uh of cluster max and uh and what what each what what how people are interpreting cluster max because there's a whole bunch of ways to read it.
1:02:03
Like one way to read it is like which which stock should you buy, right?
1:02:07
But like that's not necessarily the read.
1:02:10
The other the other the other read is like which product is the best to work with as a customer.
1:02:17
But it's like what customer are you? Yeah.
1:02:19
There are some that are in the lower tiers that are fantastic for very specific use cases.
1:02:23
Like this is the nature of every business.
1:02:25
Uh like one of the one of the one of the Neoclouds that was particularly upset with uh with Dylan uh is is in a very niche market, but if you're if you're in that niche market, it's probably a great product.
1:02:37
It's probably great for you.
1:02:37
If you're if you satisfy like this specific list of criteria and you don't need these features, you're probably fine then.
1:02:45
Um but um it's uh it's a lot of fun.
1:02:47
People are going back and forth.
1:02:49
They're also debating whether or not uh Dylan is uh is independent given that he lives with uh Schultto from Anthropic.
1:02:56
And >> we got to ask him why why he has roommates.
1:03:00
Not even I'm not even concerned about a conflict. >> It's roommate gate.
1:03:04
>> Yeah, it's roommate gate.
1:03:04
Uh but uh >> what about this other tinfoil hat post from JCON?
1:03:10
My theory is that Meta deliberately leaked the story to the information about uh Google's about acquiring Google's TPUs.
1:03:17
For Meta, it's a classic risk-free power play.
1:03:19
The moment Jensen Wong reaches wind, catches wind of Meta using Google silicon, Nvidia is likely to rush in with an investment.
1:03:28
They might even be negotiating as we speak.
1:03:30
This allows Meta to secure capital and shift from burning their own cash to potentially getting discounts or effectively buying Nvidia chips with Nvidia's own money.
1:03:38
Plus, if they actually do secure Google TPUs, they solve their compute shortage. It covers all bases.
1:03:42
I wonder when other hypers uh hyperscalers will catch on to this magic wand.
1:03:47
All you have to do is hint at using TPUs.
1:03:49
And >> but the issue the issue is is how how many red flags would be would be waving if Jensen was like, "Yeah, we're investing $20 billion in meta.
1:03:58
We're very excited about uh we're very very excited about Meta and owning a piece of of uh >> Yeah, it seems very very odd.
1:04:08
>> So So he's in a position where I I don't know what kind of leverage Jensen has around uh in those conversations with Meta because he doesn't want a discount and he it's not like OpenAI where he can just announce an investment or an anthropic etc.
1:04:24
So how does any type of like rebate actually happen is the question. Yeah.
1:04:29
Well, before we bring in our next guest, let me tell you about Turbo Puffer serverless vector and full text search built from first principles and object storage.
1:04:38
Fast, 10x cheaper, and extremely scalable.
1:04:40
Um, let's uh let's read through some more uh some more TPU stuff to set the table.
1:04:46
So, Clive Chan says, "I keep seeing stuff about TPU.
1:04:49
Has anything materially new happened?
1:04:51
There's no evidence Google has ever trained Gemini on non-TPU hardware going back to preGPT models like BERT.
1:04:59
TPUs predate Nvidia's own tensor cores.
1:05:02
Anthropic and character and SSI and Midjourney have long used TPUs.
1:05:04
I'd be surprised if Meta weren't looking at them.
1:05:08
Nvidia's moat has never been deep for the big labs.
1:05:11
C OpenAI deciding it could do better than CUDA and investing in Triton instead.
1:05:15
Regularly edging out CUDN on benchmarks.
1:05:18
Uh there's nothing magical or structural about any of this.
1:05:23
Just good engineers doing good work.
1:05:25
TPUs are not that much more efficient than GPUs and small performance per watt difference are dwarfed by whether Meta has the right kernels and systems engineering to pull talent to pull it off.
1:05:35
Both Nvidia's and Google's modes are small and we are still at the point where individual good engineers can flip the entire balance.
1:05:41
Why why was this not priced in?
1:05:44
This is all super old public info.
1:05:46
I I have a feeling that uh that this uh Clive Chan, who I guess is over at uh was it Tesla and then OpenAI is uh is is a little bit of like first time in the public markets, first time realizing that the people who trade this stuff are not necessarily like on the super inside of the labs actually uh actually understanding yeah the decisions that are being made inside the labs like it's a completely uh separate ecosystem.
1:06:11
Uh and that's why uh organizations like semi analysis exist.
1:06:17
And I believe we have Dylan Patel from Semi analysis in the re waiting room. Let's bring him in.
1:06:22
Dylan, how are you doing? >> I'm doing fantastic. How about yourself?
1:06:26
>> You know, I saw I saw the meme image that you guys uh put out there for me, so I had to wear the tank. Let's go, >> dude.
1:06:34
We need a bigger bigger uh screen for that bicep. We'll we'll work on it. >> Let's go.
1:06:40
Uh, where in the world are you? >> I'm in Florida.
1:06:44
I was spending Thanksgiving with my family here.
1:06:46
Um, I'm trying to chill out a little bit.
1:06:49
It's nice to have the family pamper me a little bit because I broke my foot a couple weeks ago. >> I'm sorry. Do that.
1:06:56
>> How'd you break your >> tripped over at TPU?
1:06:59
>> Family family reunion playing football in Texas.
1:07:01
We're American as we can get. >> There you go. There you go.
1:07:03
Um, well, uh, we were just running through a little bit of the the TPU article.
1:07:08
Can you uh can you actually set the table for me on like what do you think is new about it versus what has semi- analysis already been saying and this is more just like tying everything in a bow. >> Yeah.
1:07:23
Half of the article is just referencing recent >> we've been saying this for two years.
1:07:26
We've been saying this for one year >> and even referencing con Google's own content about the TPU dating back >> even even further.
1:07:34
So >> yeah, I would say I would say the majority of this piece was if you're if you're a client, it's already been pretty much all published.
1:07:41
Um but it hasn't been tied together.
1:07:44
It hasn't had a narrative around it, right?
1:07:45
Cuz when we think about like what we put out as on the paid side versus what we put out on the newsletter, right?
1:07:50
Um our clients sort of get uh you know what changed, what happened, uh here's the numbers.
1:07:55
Um that's about it, right?
1:07:58
We don't explain the technology that much because our clients are sophisticated, right?
1:08:02
they're either in the industry or uh and and or they're finance pros who don't give a [ __ ] about the technical stuff.
1:08:08
Um and so it's it's either of those two, right?
1:08:10
Um and and so we we're just explaining here's what's happening. Here's the change.
1:08:14
Here's the numbers, right?
1:08:15
So for months we've been saying Google's selling TPUs.
1:08:17
For months we've been saying, hey, here's TPUv7 versus Blackwell.
1:08:20
Um we've even put out updates on here's what we think TPUv8 is versus what we think Ruben is.
1:08:25
what we think Ruben is. And so generally it was making it into a narrative and explaining the technology and the corporate I would say politics or uh dynamicism around it right so that's you know I think I think there has been bits and pieces put out by other other folks right I think the information has done
1:08:41
great reporting on some of the stuff after we did but in the public space I think um you know for as an example right like um so so other people have put out bits and pieces surrounding this but they haven't put out the full picture um so so as far as like what's new it depends on where you sit in the stack. Uh but but you know Anthropic and
1:08:57
Uh but but you know Anthropic and and and Meta and folks like that have been talking to Google about buying TPUs for many months, right?
1:09:04
>> Um whereas whereas people externally are you know last week when Gemini 3 was launched or two weeks ago, people were just learning that TPUs are trained training Google's models, right?
1:09:13
So it's it's where are you in the information spectrum, right? >> Yeah, totally.
1:09:17
So, uh, on that information spectrum, uh, the finance bros, they can probably just like if they if they read into this, oh, bullish Google or bearish Nvidia or whatever, like they can kind of trade in and out as as they please.
1:09:30
But what on the on the more technical side, like are people using semi- analysis research to understand like, okay, I'm a NeoCloud.
1:09:40
What do I want to rack for next year?
1:09:43
Maybe I need to be putting in a TPU order.
1:09:44
Is that is that how people interpret your research?
1:09:46
like what happens on the technical side of the house. >> Yeah.
1:09:50
So, so as far as like some of the paid stuff we do, we have one model called the TCO model, right?
1:09:54
Which is uh calculating the TCO of all these uh different hardware performance um building up the entire cluster cost, you know, breaking out into like a dozen plus different things whether it's storage or networking and breaking down the cost of everything.
1:10:06
So there we put out research on TPUs because as soon as NeoCloud started getting offered, hey, you want to buy TPUs? Yeah.
1:10:14
>> We're like, okay, we need our own groundup model.
1:10:15
So when you're negotiating a big contract, what you do is called a should cost, right?
1:10:19
You you go and calculate what it costs for the company versus what it costs for me to deploy.
1:10:24
Um and then you like think about like, oh, what's the margins they have?
1:10:28
What is ridiculous to offer them? What is not, right?
1:10:30
Because everyone always wants to know like, hey, what margin are they making off of me?
1:10:33
Can I push that down a little bit?
1:10:35
Um what is what is ridiculous to demand in a negotiation versus what's not?
1:10:39
So we've already been working, you know, through this TCO model.
1:10:42
model. We've put out four different updates on the TCO of uh TPUs V7 and V8 because there are NeoClouds out there um as well as labs who are purchasing TPUs that are using that to understand what's the cost now you know Enthropic I will say just already knew and figured it out because they've hired so many Google
1:10:58
people but other labs are also looking at it right yeah >> um and so so you know when you when you say hey on on the cost side of things on the technical side of things right there's a lot of network engineers now out there who have never deployed Google hardware that are now like okay I need to figure out how to do text, right? Like, you know, so there's there's
1:11:13
Like, you know, so there's there's people who have DM'd me that are like, "Oh, we've been, you know, as you know, we've been thinking about deploying Neil clouds, but your material on this is technically better and teaches me more than Google's own material, right?"
1:11:22
So, it's like this is this is helpful to people on multiple factors. >> Yeah.
1:11:27
What what about uh the software side?
1:11:29
Uh Google's built their own internal stack to compete with CUDA.
1:11:31
Uh how much of that are they going to actually give to their customers who are buying TPU?
1:11:38
buying TPU? because that feels like you it feels like potentially you could overrotate on oh well Gemini 3 is really good but why is it good is it just because of the hardware or is it also Google's incredible prowess multi- data center training all this fancy stuff that they have that they won't be giving
1:11:56
you when they sell you the TPU >> yeah so so that's the interesting thing is some of the stuff software will remain closed source but you can still use it right and then some of the software um they are trying to open source aggressively and then some of the software they're never going to get out there anywhere, right? So, it sits in
1:12:09
So, it sits in three kind of buckets, right?
1:12:10
Um the interesting I guess newer thing that we did in the piece was um we looked across all these different open source AI uh software, right?
1:12:19
Whether it's PyTorch, whether it's VLM, whether you know, all these different um open source uh libraries and we calculated and counted up how many Google uh commits there were, right?
1:12:29
And you can see there's a chart in the article where the number of commits that Google's doing on TPUs has exploded over the last handful of months, right?
1:12:35
as they've decided to shift their strategy, sell GPUs externally.
1:12:39
They also recognize software has to be open for this, right?
1:12:41
Um, you know, only the gigab brains that like anthropic can figure out how to do everything themselves, right?
1:12:46
It's it's those people outside of anthropic, you know, types that that need a bunch of open source software that builds on top of it, right?
1:12:53
Um, and what's interesting is when you look at like, hey, Nvidia, you know, the biggest argument that Nvidia doesn't really make for GPUs, but they should, is that, you know, about 40% of the software that's open sourced is actually just from China, right, on on CUDA, and that's the CUDA mode, right?
1:13:08
It's like 40% of the software is just like open source stuff, whether it's people committing to VLM or PyTorch or or all these other libraries, right?
1:13:15
Byte dance open sourcing stuff, DeepSec open sourcing stuff.
1:13:16
Um and and and and and Google, you know, they they don't have people open, you know, Anthropic is not going to open source software.
1:13:22
So Google needs to catch up not just by, hey, here's all the software we have internally, let's open source it.
1:13:27
They also need the ecosystem to build a ton of software on top of TPUs.
1:13:30
And so that's the that's the real big uh challenge there.
1:13:34
Um and and there's an element of software there that Nvidia's happy to open source.
1:13:37
Um and customers of Nvidia are happy to open source that Google will never open source because it's it's you know, Google Cloud is selling the TPU.
1:13:44
Gemini is the one actually using it and developing a lot of the software and these two groups are not always going to be aligned. >> Yeah.
1:13:52
Isn't that uh like I mean what are the other kind of uh just problems with uh Google becoming an actual like seller of TPU?
1:14:01
It feels like uh there's obviously an opportunity because Nvidia has high margins, there's demand, it's a great chip, but culturally, structurally like Google tries a lot of different things.
1:14:11
they have a lot of advantages, but occasionally like they fall flat on their face with just like they can't even get an RSS reader out or something like that.
1:14:18
Uh, so like like are there other risks to the TPU not really finding its footing for reasons that aren't just the the laws of physics? >> Yeah.
1:14:30
So, so the biggest challenge I see with them is it's everything is non-standard, right?
1:14:33
Google for years, they they developed liquid cooling first, right? Sure. uh for for AI computing.
1:14:39
They deployed rack scale architectures, right?
1:14:41
Everyone's talking about GB200 rack scale architecture.
1:14:44
Google did it first with TPUs, right?
1:14:46
But when they did all of this stuff, they didn't give a crap about, hey, you know, this has to go in 50, 100,000 different people's data centers, right?
1:14:54
>> Um this has to go in my data centers that I designed myself.
1:14:56
So everything is super vertical.
1:14:57
The entire liquid cooling supply chain is super vertical.
1:15:00
Entire the the racks aren't even the standard uh width, right?
1:15:02
So when I look at like a data center, it's like the door the loading bays because they're so much wider.
1:15:07
The Google racks are like three times as wide.
1:15:09
It's like maybe it might not even fit into the data center like physically like through the doors.
1:15:14
Um so there's like all sorts of random like I wouldn't say random, it's Google from first principles design stuff. >> Totally. Yeah. Yeah. Yeah.
1:15:19
But but if you're Neo cloud and you're like the hot thing is going to be TPU next year or the year after and I want to be able to sell into that market.
1:15:27
Uh it's not just flip a switch, drop in, replace with TPU.
1:15:31
you have to maybe build a whole new building like it might be that significant.
1:15:36
>> Let >> right or or do or or like knock down some walls and then like you know I need to I need to go get liquid cooling not from Dell and Super Micro and and HPE who I've who service me already.
1:15:44
I need to go get it from some random supplier who's only ever sold to Google.
1:15:48
And usually they're sitting across the table from like some gigabrain engineer who has a team of 20 people working on liquid cooling instead of like you know my one guy who does liquid cooling procurement and negotiations and like also does procurement of like network stuff. Yep. Yep.
1:16:04
>> There there was a there was a tinfoil hat theory floating around that Meta leaked their TPU in Meta leaked their TPU interest uh to try to gain some ne uh sort of leverage over maybe uh some negotiations with Nvidia.
1:16:16
I don't know if you uh see any possibility in that.
1:16:22
But how do you think those conversations are going?
1:16:24
uh Jensen doesn't want to discount uh and compress his margins, but at the same time, he can't do this kind of like equity rebate thing.
1:16:30
If he if he took a big position in in Meta, he'd be very suspicious. I'd be very concerned.
1:16:36
>> I totally get the opening eye investment.
1:16:38
That seems like it makes much much more sense than saying, "Hey, we're going long Meta, you know, is a $4 trillion company." >> Yeah.
1:16:47
At the end of the day, right, like TPUs have like a set of maybe 10 customers, right?
1:16:50
Because you have to be super sophisticated. Yeah.
1:16:52
Um, and so what really is challenging here is is, you know, Meta Meta looks at the numbers, you know, it's like, okay, open I'm getting 30% off because they're paying they're investing in me as a result.
1:17:04
Obviously, they get equity, but they're investing in me and I get 30% off on these GPUs as a result, right? Meta, you can't do that.
1:17:09
So, so Meta, they're I don't think that they're just negotiating, right?
1:17:13
Like, you know, is is are they just negotiating with Nvidia when they buy AMD?
1:17:16
No, they're any engineers, right?
1:17:19
They're developing all the software.
1:17:20
they're actually deploying uh Llama 405B was exclusively on AMD for a number of months, right, for inference, right?
1:17:28
Um so so when you look across, hey, is Meta just like playing around trying to negotiate? It's like no, no, no.
1:17:33
Like they're they're looking out for what is best, right?
1:17:35
And Meta is power constrained and TPUs are currently way more power efficient.
1:17:38
Meta is compute constraint.
1:17:40
Um and TPUs are potentially higher performance per watt and higher performance per dollar, right?
1:17:46
At least that's what we believe for TPUv7 that it is.
1:17:47
So, they'd be dumb not to look at it, right?
1:17:50
And they have the time, they have the people, they have the team.
1:17:53
Now, Nvidia at the same time has to play the game of chicken, right? Yeah.
1:17:56
Sure, they could discount the pricing somewhat.
1:17:58
Um, and because what's funny is Nvidia is more vertically in integrated than Google is when selling hardware, right?
1:18:04
Google has to pay Broadcom who pays TSMC, whereas Nvidia gets to pay TSMC directly, right?
1:18:10
There's this vertical integration challenge where Nvidia could drop the price a little bit and they'll be fine, but they don't want to, right?
1:18:14
You know, the whole point is you charge the highest price possible.
1:18:17
highest price possible. And then the last thing is they've got this like um you know they've got this view about antitrust right you you don't want to cut deals for specific customers because that looks bad right instead you want you know right now Dell pays the same
1:18:32
price for a GPU as Gigabyte as meta >> now the networking hardware there's different pricing because there's a lot more competition and Nvidia can cut a lot more there but on the GPUs themselves Nvidia's pricing is very fair right fair in the sense that they're making a shitload of money off of everyone you Yeah. >> Uh how talk about kind of Jensen's
1:18:51
>> Uh how talk about kind of Jensen's leverage uh that he has around Reuben allocations as some of these customers start to uh at least consider TPUs. >> Yeah.
1:19:04
So as far as like next year's TPU deployments, it's pretty set in stone for the vast majority of the volume, right?
1:19:10
Anthropics got a bunch and then there's some sprinkled elsewhere.
1:19:12
there's some sprinkled elsewhere. Uh but as we go into 2028 where Google can actually ramp um you know the flip side is Ruben is also ramping and at least based on our research looking throughout the supply chain um you know over a year ago when open started their trip team they poached like 15 Google people overnight right in one week like someone
1:19:31
I knew I heard was like oh yeah I'm joining open I text like another three people I know and they're like oh yeah I'm also joining open like what the [ __ ] um so so Google's a lot of their best TPU engineers have left right they also have a ton left and so that what that's done is, you know, chip timelines are so long, that didn't affect TPV7, that's affecting TPV8. At the same time,
1:19:47
At the same time, Google's trying to diversify their supply chain, get from not just Broadcom, but also MediaTek.
1:19:51
And so, Google's got a real challenge on TPUv8 in that uh it's good.
1:19:56
It's an improvement, but then when you go look at what Nvidia is doing with Reuben, Reuben is so much better because Nvidia is just pedal to the floor, paranoid as [ __ ] Uh we we have to be the best and we have to be way way way better than everything because how much better I am than everyone else is my margin, right?
1:20:13
And so Nvidia, Nvidia has the sort of like at least currently we think Nvidia is going to be so much better that they'll be fine and they'll be able to maintain margins.
1:20:20
Right now things can happen.
1:20:22
Reuben can delay or TPUs can delay and the position looks better or worse, right?
1:20:26
Um there's a lot of unknowns to go through, >> but as far as like what is Jensen's leverages, look, I'm going to make the best hardware and plus my software advantages and I'll be able to continue to be dominant and dominate the market, right?
1:20:38
right? Um there's there's curveballs I could go which is like oh Google software they could open source enough software that actually their software ecosystem is not far behind Nvidia maybe they don't want to right or hey they could um you know they could execute everything and Nvidia has a three sixmonth delay now all of a sudden
1:20:54
they're a lot more competitive right um and and so all these things are still open questions but it's it's Nvidia can play the allocation game as well of course right um hey I'm going to give all of the GPUs initially to companies that probably could buy uh TPUs but that ends up all the AI labs and hyperscalers, right? Um at least, you
1:21:10
Um at least, you know, like Meta, right?
1:21:12
And bite dance, people that would actually be willing to buy TPUs.
1:21:15
Um and then you end up with this like weird situation where okay, well that's like 75% of the GPU market anyways when I look at the the AI labs through the NeoClouds, right?
1:21:23
Um when there's, you know, Nebus and Iris Energy and all these other, you know, Coreweave and all these folks are buy are deploying for OpenAI anyways, right?
1:21:31
Um, you know, this this sort of ends up being like, well, sure, I could stiff like some people on the allocation, but at the end of the day, everyone who was a potential customer for TPUs uh is sophisticated enough to be where they were going to be on the beginning of the allocation anyways, >> right?
1:21:51
>> What what uh how are you framing ClusterMax these days?
1:21:52
Is it is it for customers who want to buy services from Neoclouds?
1:21:58
Is that the primary goal of cluster max?
1:22:01
because I feel like some people look at it and they're like, "This is a buy rating.
1:22:04
This is a sell rating on the stock."
1:22:09
>> So, so the funniest thing is like Cluster Max V1, the title of it was Cluster Max, how to rent a GPU, right?
1:22:15
Because we discussed all of that and then and then in Cluster Max V1, I believe we put Irish energy and underperform, right?
1:22:20
At the same time, the research side of the business, uh, we explicitly were like, "Dude, they've got these data centers.
1:22:26
It doesn't matter if they suck at running GPUs.
1:22:29
They've got these data centers. They've got this power.
1:22:31
got this power. if you just value them on a watts per you know how much money they could make it's it's a long so like at the same time as like um Jordan right Jordan he's running cluster max is like Iris kind of sucks uh and it was other people in the technical team before him you know it's like it's like Jeremy who's running the data center side and I
1:22:48
think he's been on TVPN is like dude Iris energy is a stock right so it's like it's it's kind of like you know it's like what what what the technical side of the house does versus what the you know research side of the house does yes they talk to each other right Jeremy did ask the team like, "Hey, what do you think of Iris Energy? I think it's a I think it's a log."
1:23:04
And and the team working on Cluster Max is like, "I don't know."
1:23:07
Like, you know, it's it's it's a bad cloud.
1:23:09
And it's like that doesn't matter.
1:23:10
So, ClusterMax has nothing to do with the stock, right?
1:23:12
Um, now obviously there's going to be some correlation with how good is a stock versus, you know, who's going to want to rent from them. Yeah.
1:23:19
them. Yeah. Um but at the end of the day right like cluster max is the the goal purpose sole purpose and what we explicitly say in there is it's for people renting anywhere from like you know hundreds of GPUs to you know right below the AI lab scale right the AI lab scale there's different considerations
1:23:35
um but in that range tens of thousands of GPUs all the way down to hundreds of GPUs that's who we're targeting plus we're saying we're giving a bunch of feedback for people to make the cloud ecosystem better >> the unsung hero between cluster max v1 and v2 is that we move the bar up Right? You know what it required to be in gold
1:23:49
You know what it required to be in gold >> like was was much more.
1:23:52
What it required to be in silver was much more because everyone improved so much, right?
1:23:56
And as as we continue to like increase the requirements, make it harder and harder to >> got keep moving the goalpost, >> right?
1:24:05
People keep improving the ecosystem and actually, you know, this is this is the funny thing.
1:24:08
It's like cluster max is evil.
1:24:09
It's like when I when we look at the quotes and we've got hundreds of quotes on clustermax.
1:24:12
ai, all these companies are like, "Dude, I love this.
1:24:15
this one specific bug that this Neocloud had, they fixed it as soon as you wrote about it. Yeah. Right.
1:24:19
Or like, hey, help me understand the reliability, help me understand this or that.
1:24:23
People are like, "Love clusterbacks."
1:24:24
Um, and and you know, altruistically, like I think we're generating billions of dollars in value just from hey, like all these clouds are more efficient and there's less failures and it's easier to get your workload running on any random GPU cloud and the market is more efficient.
1:24:38
Um, now I'm not making any money off of that.
1:24:40
How am I making money off of cluster max?
1:24:41
I'll be very clear is people who hire us to do due diligence, right?
1:24:45
So, people who want to acquire a NeoCloud, people who want to sign a massive massive deal that's not just like thousands of GPUs, but tens of thousands of GPUs, and then lastly, it's people who want to um you know, invest in NeoCloud.
1:24:58
Those are the three areas where we're making money off of quote unquote cluster max, but not really.
1:25:03
We're not selling ratings.
1:25:03
We're not, you know, you know, we're in fact like a customer will do a consulting project with us or want to want to buy some research from us and I'll explicitly put in our Slack share Slack or I'll send an email to the CEO like, "Dude, just so you know, the people working on this are not the people who are doing cluster max rating, right?
1:25:19
Um, you know, the people who are buying, you know, the research on like these data centers are there and this is the power ramp or here's the accelerators or here's the TCO.
1:25:27
That's not the people doing cluster max, right?
1:25:29
And I don't care about, you know, whether you buy it or not.
1:25:33
or not. I you know at the end of the day Google and Amazon and Microsoft are way bigger customer than you know Flipstack and like you know those kind of companies right and yet one some of those are ranked in silver and some of those are ranked in platinum
1:25:47
and gold and that's because technically what's what matters not hey you know obviously when we talk about who buys our research the biggest companies in the world are going to pay me more than the mid-size companies in the world. Okay, question from the chat.
1:25:56
Okay, question from the chat.
1:25:57
>> And the price has discriminated based on that.
1:25:59
>> Uh, would you change the rating of a Neocloud if Shalto promised to do the dishes for two weeks straight?
1:26:07
>> You know, there was an argument.
1:26:07
I saw someone was like, "Who does the chores?"
1:26:10
And it's like, "Brother, we we live together by choice, you know, we we pay someone to come once a week.
1:26:14
If you cook something, you do your own dishes."
1:26:16
But like, you know, um, frankly, we're we're working so much and I think like, >> you know, I think I think Dores Ces has ordered pizza from the same spot three nights in a row before, right?
1:26:26
Like it's it's it's Is being an adult man with roommates underrated.
1:26:36
>> Um, so I'm I haven't lived with people in years and then when I moved to SF. >> So you came back. >> This is crazy.
1:26:40
I moved with I moved to SF uh this year, you know, I'm like, "Oh, you know, I should live with friends just so it's more fun."
1:26:46
Um, and the first house kind of fell apart, so I moved into this house with these guys and we've been talking about it for months. Um, I love it, right?
1:26:52
It's like, look, we we we we have, you know, if you think about, oh, what if we all rented our own places that were good and then we pulled that budget together? We have a nice place. Yeah. Right.
1:27:01
And then in that place, we have plenty of space for ourselves.
1:27:04
We we pay for someone to come and clean once a week, right?
1:27:07
So, at the end of the day, what is what is the negative here?
1:27:10
Is like, >> well, we're living with our friends, but we have enough space to where like >> and the beauty is if you if you do bunk beds, you have more room for activities. >> Exactly.
1:27:23
>> Anyway, no, no, sorry, sorry.
1:27:23
Actual question from the chat.
1:27:25
Uh, when is TPU going on inference max? We got to know.
1:27:30
>> So, we're working on it, right?
1:27:30
We we're we're working with Google um technical folks.
1:27:34
Um, you know, funnily enough, actually, um, we triggered a security warning for this Google engineer.
1:27:38
Uh, Kimbo went to a a Jax conference, right? Jax is is the opposite.
1:27:43
It's like PyTorch for but for TPUs. It's the most simple.
1:27:47
It's it's Google's own internal thing, right?
1:27:48
Um, that people do use externally.
1:27:50
Um, he went to this PyTorch or this Jax conference.
1:27:52
A Google engineer presented something.
1:27:54
He's like, can I get the slides?
1:27:55
They send it to him and then Google security alert like locks him out of his computer because he sent us like a some technical like information.
1:28:02
said like for 3 days the guy can't work and he's freaking the [ __ ] out and I'm like I I emailed Jeff Dean.
1:28:07
I'm like bro this is like do not fire this guy.
1:28:09
He sent me stuff that you presented at a public conference and he's like oh okay yeah yeah I'll get that fixed but anyways like um we're working with we're trying to you know implement it.
1:28:16
Uh we have access to some TPUs.
1:28:19
>> The software stack is different right? Yeah.
1:28:21
>> You know just >> so you have to so you basically have to rewrite or reimplement inference max like like the code that actually >> I won't say I won't say it's that much work like as much as like completely redoing inference max but there's a ton of work. Right.
1:28:31
So we're moving as fast as we can internal target is this year.
1:28:35
Um you know then the obvious question is is like I I feel like inference max is my north star for TCO relative in AMD versus Nvidia land.
1:28:46
Uh there was a bar chart of TCO for GPU versus Nvidia.
1:28:49
It looked like it looked like TPU was doing really well on that chart. The bars were very low.
1:28:57
Where did those numbers come from?
1:28:57
Do you have confidence in those numbers or do you think the numbers will change once you actually get TPU on Inference Max? >> Yeah.
1:29:06
So, Inference Max shows performance TCO, right? You know, it's great. Great.
1:29:09
Like, you know, like guess what?
1:29:10
Like, um, you know, TCO of like a Raspberry Pi is incredible.
1:29:13
It's like five bucks, right?
1:29:15
Um, you know, versus versus a GPU is $50,000.
1:29:17
Performance divided by TCO is what matters.
1:29:20
divided by TCO is what matters. So that bar chart is saying look TPUs are cheaper and at least on quoted specs you know now now let's make some assumptions around utilization and in the in the article we explicitly said look we don't know what the utilization is it's going to change customer to customer um here's
1:29:34
a range worst case it's like a little bit worse than GPUs best case it's way better than GPUs right um and and so inference max will tell us what the actual performance is in inference um because we don't know yet right um currently the open source software for TPUs is not good enough for good enough for us to just take the open source software and say that's the performance, right? Because that's obviously like not
1:29:53
Because that's obviously like not real, right?
1:29:55
Anyone who like is actually buying TPUs is going to spend engineering hours to work on it.
1:29:59
And so we're trying to work with Google to get a real performance number that is achievable by people.
1:30:03
Um, you know, and and will be upstreamed into the open source software because this is an in progress thing, right?
1:30:09
No one cares what TPV7 can do today.
1:30:11
It's about what it does in six months.
1:30:12
Um, and so, you know, obviously we don't want to be, you know, today TPUs, if you're using VLM are worse performance TCO than GPUs without a doubt, but the target is moving very fast.
1:30:23
And, you know, there's a ton of like lowhanging fruit for us to implement before we actually put a number out there, right?
1:30:28
Um, and so where does Google sit there? We'll see.
1:30:32
Um, I I personally believe the TCO side of things, the total cost of ownership is based on what we know on supply chain, right?
1:30:38
How much do uh how much do the chips cost?
1:30:40
How much do the racks cost?
1:30:41
How much do the liquid cooling cost?
1:30:42
how much does the memory cost, how much do the cables cost, etc. , etc. , etc. , right?
1:30:46
That's based on our estimates up and down.
1:30:47
So, I think the TCO side of things we're pretty confident.
1:30:50
Um, it's the it's the performance side of things where we don't know, right?
1:30:53
Um, there's a wide range and that's what we sort of tried to state in the article, right?
1:30:57
Performance is a wide range.
1:30:59
>> Uh, can you can you uh explain more about Google and Broadcom's relationship?
1:31:04
Max Hodak from from Neurolink and Science was was asking on the timeline last week why why have Broadcom as a middleman.
1:31:11
Couldn't couldn't uh Google do the design and and place the orders from TSMC themselves, but but what's your read on on that relationship and how how durable it is? >> Yeah.
1:31:23
So, when you think about chip design, there's a few different stages, right?
1:31:26
There's defining the architecture and then there's actually like implementing that architecture onto a process technology.
1:31:31
There's laying out that architecture into gates in the on the chip and then there's like the whole supply chain side of things, right?
1:31:40
Negotiating contracts, getting allocations, etc.
1:31:42
>> That takes like 18 months, right?
1:31:45
>> Isn't that like an 18month process basically?
1:31:47
>> Yeah, 18 months or more, right?
1:31:47
Um >> I would say I would say actually like Nvidia is is faster side and Google's on the slower side just because you know Nvidia's been doing it for longer.
1:31:56
They have a bigger team, right?
1:31:58
have a bigger team, right? Um and they they you know but at the same time Intel has the biggest chip design team and they move even slower than that right they take like four years at least that's what they did a year or two ago we'll see what the new CEO can get into the you know you know reorgate right um but as far as like Google you know when
1:32:13
they first started the TPU it was a very few people and they relied heavily heavily heavily on Broadcom to do everything right they just defined the top level architecture and Broadcom did everything I said below right negotiating with supply chain figuring out proc uh figuring out how to lay out the gates everything right As time has moved forward, Google has taken on more and more of this. Right now, they use,
1:32:30
Right now, they use, you know, they've talked a lot about Alpha Chip where they use AI to help floor plan the chip, right?
1:32:35
Once you have the architecture, how do I physically lay it out onto the chip, right?
1:32:40
Um, they've done more and more and more there.
1:32:42
They haven't taken over everything yet, but that's that's sort of the point.
1:32:46
But Google Broadcom has this like super big advantage, right?
1:32:49
Nvidia, they they acquired Melanox, you know, call it five, six, seven years ago. Huge acquisition.
1:32:53
Who's the biggest networking company in the world?
1:32:56
Broadcom right Broadcom is the biggest networking company in the world and you know when you talk about AI it's it's the AI it's the architecture of the actual processing elements it's memory um which you're buying from you know you know the memory companies right and Samsung and Micron and then it's
1:33:12
networking right when you try and boil it down to the most simple things and software right the networking side of things is so important and the let's say technical competence of everyone around the world besides Broadcom and Nvidia in networking is so Oh, or rather it's just not as good as them. They're actually
1:33:27
They're actually good, but it's like Broadcom and Nvidia are just so good and Broadcom is better than Nvidia in many ways at networking that you know when you think about what is Google doing, yes, they're defining how the network topology is, but when you're talking about the physical networks, you know, how how packets get transferred, all these different things.
1:33:45
Um, Broadcom has heavy heavy influence there.
1:33:47
So, to this day, right, Broadcom is still charging margins like they did three, four years ago, even though Google has taken up more and more of the work.
1:33:53
Um, but at the same time, Google can't leave until they figure out how to do the networking and supply chain themselves or with a partner.
1:34:00
And so what are they doing on TPUv8 that is potentially a distraction that's slowing down their execution is they're working with MediaTek, right?
1:34:08
MediaTek at times has helped Cisco with their network chips.
1:34:11
MediaTek has uh a lot of work on some of this networking stuff.
1:34:14
They're nowhere close to Broadcom, right?
1:34:15
On revenue, right, for you that's that's that's one metric on technical competence, you know, that's another metric.
1:34:21
I I think Medatech is good, right?
1:34:23
But like they're just nowhere close to Broadcom.
1:34:24
So now Google is having to work with, you know, I don't want to say subpar vendors, but uh inferior vendors to Broadcom, >> and that's just to increase their margin on TPU8.
1:34:35
>> Um I would I would even say their angle when they started this project was never we're going to sell TPUs externally.
1:34:38
It was dude, we're paying, you know, a 3x markup to Broadcom.
1:34:42
Um and half the cost of this chip is memory.
1:34:45
Like what the [ __ ] are we doing? Right?
1:34:46
[ __ ] are we doing? Right? um you know at the same time it's like well sure physically the cost for the networking is not that much but what value does the networking bring is you know sort of Broadcom and then Broadcom's also doing the like game theory not science of like well you can't really leave us so we're going to charge you what we think is
1:35:02
fair or what we think we can charge and Google's like oh no we're stuck to you right so media is taking way way way less margin they're not passing the memory through them right and so you know this this ends up being like hey that's a huge advantage for them flip side is like well they've They've they've got to engineer all this work that Broadcom was doing. Instead of
1:35:20
Instead of working on a way better architecture, they've got to work with a worse vendor, right?
1:35:25
Objectively worse, although MediaTek, like I said, is very good.
1:35:27
Uh to try and implement TPUs uh more directly with TSMC with less Broadcom sort of in the middle. >> Very helpful.
1:35:37
>> And and Google, you know, because it's risky is going down both paths, right?
1:35:41
They're continuing to work with Broadcom on TPV8 and then separate TPV8 project they're working with MediaTek, right?
1:35:47
cuz they can't risk, you know, fine whatever 30 points of margin, 40 points of margin, 50 points of margin.
1:35:50
I can't risk the TPU being late because ads runs on that. Gemini runs on that. >> Yeah.
1:35:55
Uh can you can you give any takes on the Nvidia's $2 billion investment in Synopsis that that got announced this morning?
1:36:03
I don't know if you you saw it. I'm assuming you did. >> Yeah.
1:36:06
So, in a time where, you know, let's say the two biggest chip makers, Broadcom and Nvidia are making more money than ever and everyone else in the supply chain and all the hyperscalers are trying to design more and more chips.
1:36:16
Everyone's everyone's sort of working on that.
1:36:17
You've got you've got the the EDA vendors are at the lowest possible valuations or lowest valuations that they've had.
1:36:24
They're still very expensive, but lowest valuations they've had on a earnings multiple basis for a long time there.
1:36:29
And and and this is on the eve of hey like objectively are there going to be more chip designs or less chip designs in five years? A lot lot more.
1:36:38
Right now the flip side is AI chip design is coming.
1:36:40
There's 20 plus companies doing AI chip design.
1:36:42
And it's a we've got a really long article coming on that soon uh that will sort of explain the landscape.
1:36:47
But AI chip design is going to shake up everything.
1:36:51
And so the question is like >> AI this is AI AI chip design correct >> AI helping chip design whether it's for AI chips or for like power chips. Okay. >> Yeah. >> Got it.
1:37:02
>> Got it. Um and so so the question is like you know Nvidia Nvidia has a lot of tools internally right the the dirty the thing about EDA is that there's three companies that own 95% of the revenue but at the same time Google and Nvidia and Broadcom and all these guys also design a lot of their tooling internally
1:37:17
although they are massive customers of all three vendors right so it's kind of like an oligopy where the customers also contribute a lot um and so Nvidia's whole goal here is like how do I get every EDA flow working on GPUs because today a lot of it is running on on FPGAs a lot of it's running on CPUs. Um, and
1:37:32
Um, and AI AI chip design is going to get a lot more AI influenced.
1:37:36
How do I get everything working on GPUs?
1:37:38
Um, in terms of like the operation of it, even if it's helping people design, not GPUs, right?
1:37:44
Um, and and I don't have enough engineers to work on all the software.
1:37:48
They've open sourced a lot of software, right?
1:37:50
Like KU litho, it's it's software for lithography, right?
1:37:52
And they've got all this software up and down the chain all the way from lithography to laying out chips and all this other things.
1:37:59
They just want to make it all run on GPUs.
1:38:00
Um, and and so that's that's what their goal here is, right?
1:38:04
And now they've given Synopsis a huge huge they're buying Synopsis at the lowest valuations that Synopsis has ever had with all this cash that they were going to give away in dividends or buybacks anyways.
1:38:13
Um, and and they're getting Synopsis to now make GPUs first class, right?
1:38:18
And so I think this is a win-win um for Synopsis and Nvidia.
1:38:23
>> Well, we could go way longer, but I know what's on your calendar. You got to hit the gym.
1:38:27
Thank you so much for coming by and chatting with us.
1:38:31
This is really really helpful.
1:38:33
Have a great rest of your day.
1:38:35
Enjoy the holidays with your family. Uh great catching up. We'll talk to you soon. >> Cheers. >> Goodbye. >> See you guys. >> See you.
1:38:42
>> Uh let me tell you about public.
1:38:42
com investing for those who take it seriously.
1:38:45
They got multiasset investing and they're trusted by millions.
1:38:46
Uh we have Ro Kana in the reream waiting room.
1:38:52
Let's bring him in to the TVP Ultradome. Ro, good to meet you. Welcome to the show. How are you doing? >> I'm doing well.
1:38:59
You guys have become quite the celebrities in my district.
1:39:03
Everyone is tuning in to your podcast. >> I'm glad to hear it.
1:39:06
I'm glad to hear it and we're and we're happy to have you on the show.
1:39:09
Thank you so much for the time. >> What was that?
1:39:11
>> Are the all-in guys are the all-in guys jealous or do they do they respect you?
1:39:15
>> I think that they have left they have they're in the stratosphere.
1:39:17
They have left the you know the the scraps.
1:39:22
>> Yeah, we're picking up the scraps compared to them.
1:39:24
I think I I think like any great Silicon Valley startup, you're nipping at their at their heels. I >> It's funny.
1:39:30
We we we had we had a New York Times piece around us.
1:39:31
It was a very nice uh you know, just like here's what TBPN is doing.
1:39:35
Just kind of an explainer piece.
1:39:37
Uh David Sax, of course, uh got a little bit more of the investigative journalism treatment. Got five reporters. We only got one.
1:39:43
And so I think that tells you about the relative importance of the shows.
1:39:45
Um but anyway, I'm sure we'll get into that.
1:39:48
I would love for you just to kind of set >> normally normally when our guest joins join and they're wearing a suit, we we say thank you.
1:39:56
But I think this is I'm assuming it's one of your daily drivers. So yeah.
1:40:01
Uh >> well, you know, I I'm not back in my district.
1:40:04
They it'd be the only way I'd lose my seat is if I started to show up with a suit to like the place was there.
1:40:11
But in DC with the uniform, >> it's the uniform in DC.
1:40:14
But I was hoping you could you could sort of uh take us through a little bit of the prehistory since it's the first time on the show.
1:40:21
Just explain uh how you wound up in this position, a little bit of your backstory, and then um obviously there's so many hot topics that I want to talk about in uh artificial intelligence and tech broadly.
1:40:32
And I want I want your opinion on on everything that's going on, but I'd love to kick it off with a little bit of like how you wound up in Congress. >> Sure.
1:40:41
Well, I uh am the son of immigrants.
1:40:43
My parents came from India in the late 1960s.
1:40:48
My grandfather spent four years in jail alongside Gandhi as part of the Indian independence movement.
1:40:53
And that really inspired my love of public service.
1:40:56
Uh when I came out to Silicon Valley, I I had a professor, Larry Leesig.
1:41:01
He said, "If you care about policy, go go out to to to Silicon Valley.
1:41:05
That's where the interesting things are happening.
1:41:07
That's where the big things are happening."
1:41:08
So I went out uh and I ran when I was 27 against the Iraq war and I got killed.
1:41:13
I got crushed 71 to 19 but came to the attention of uh folks as someone willing to stand up for uh against the war.
1:41:23
And then I uh worked uh as a as a tech lawyer.
1:41:29
I supported President Obama.
1:41:29
I got to go work for President Obama.
1:41:31
And then I wrote a book about uh what we needed to do to build new manufacturing across this country in 2012.
1:41:37
What we needed to do to uh really have the modern economy in different parts of the country.
1:41:44
>> You're one of the the the first American uh beginning of the American dynamism movement.
1:41:48
Would you would you say >> Yeah.
1:41:51
I I I was going to say to President Trump, he stole all my ideas in terms of manufacturing, but you know, he >> Well, that's good. That's good though.
1:41:58
That's good though then, right?
1:41:58
We uh just want >> No, no.
1:42:00
Look, I I I support the American dynamism movement.
1:42:03
I I'm a fan of sort of what Mark Andre wrote in a Wall Street oped about like how do we not make masks in America?
1:42:09
How do we not make basic things in America?
1:42:11
Uh when my parents came to this country in the 60s, we were the place to be. We were humming.
1:42:17
We were brimming with confidence.
1:42:18
Kennedy said to go to the moon.
1:42:20
Uh and and uh you know, my first book was about why manufacturing still matters.
1:42:24
I think it was a colossal mistake to let China eat our lunch on so many key industries, especially now with rare earth metals and magnets.
1:42:30
I mean, we should have a Manhattan project to do that in the United States or New Zealand, Australia, Chile.
1:42:36
But, you know, so I after my time in the Obama administration, after I wrote this book, I said technology is is going to shape so much of the future of this country.
1:42:45
I have a vision of how we can make sure that it uh helps everyone in my district and around the country.
1:42:52
and maybe I have something to offer to to Congress.
1:42:54
So, I ran against an incumbent again, lost again.
1:42:58
California is a machine dominated state.
1:43:00
It's very hard to break in and I persisted and won on my third try. >> There you go. >> Third times a charm.
1:43:07
>> And so, for this year, uh, in 2025, how would you frame, you know, your top priorities?
1:43:14
There's this weird there's this weird disconnect between, uh, that we've been tracking on like how relevant is AI?
1:43:21
It's it's so dominant in tech and yet uh if you talk to somebody at Apple, they'll be like, "We didn't want to focus on AI this year at all.
1:43:28
We wanted to focus on battery life because that's what helped us sell phones and AI was actually not a driver of iPhone sales, for example."
1:43:36
Uh it's it's a it's a deeply pervasive discussion point.
1:43:38
And yet it's not necessarily uh and yet it's it's widely used, but also widely hated.
1:43:47
It's such a unique technology, but uh just in terms of political priorities, what's been on the top of the stack for you this year?
1:43:55
>> I want to answer your question on AI, but obviously in the last few months, what's been a highest priority is getting these Epstein files released.
1:44:03
Thomas Massie and I passed the Epstein transparency act.
1:44:06
It was my bill passed 427 to one, 100 to zero in the Senate and Donald Trump signed it.
1:44:11
Most urgently, it's about justice for these underage girls, over a thousand victims who were raped at Epstein's Island.
1:44:18
But it's also about this kind of idea of elite impunity that these rich and powerful people, I call them the Epstein class, don't play by the rules, which you and I have to play by and people are tired of it.
1:44:29
And it also is a story of how how in the world you get some things done in Washington.
1:44:32
how how did a Bay Area progressive congressperson end up getting Donald Trump to sign his bill and getting 427 uh people in the house to vote for it and 100 senators?
1:44:42
So that has been uh the immediate priority.
1:44:45
But what I say to folks >> So are are you optimistic that uh the American people will ever get a like a truly cohesive narrative on the Epstein story or will it be our generation's JFK assassination?
1:45:03
I'm confident we're going to get far more than we've had so far.
1:45:05
The release is now mandated by law December 19th or December 20th.
1:45:11
>> I think more names are going to fall.
1:45:13
You've already had some high-profile names uh fall because of their affiliation with Epstein covering up for him or being inappropriate.
1:45:19
I there are going to be other names that come out.
1:45:24
Now, >> do I think that it's going to satisfy everyone? No.
1:45:26
There's always going to be some sense that we didn't get a full justice, but it's going to be much better than these women who were denied justice for decades, which was not partisan.
1:45:36
I mean, they were shafted by a justice system that didn't work.
1:45:40
And there are a lot of rich and powerful people who got away with it.
1:45:44
>> But look, I what I tell people is that AI is going to matter even more than anything.
1:45:48
And and to your point about Apple, it's not AI literally as just AI as Grock or Chat GPT or a technology that detects patterns and can predict the future based on patterns.
1:46:01
It's more that AI has become a symbol for a technology revolution that people know is changing everything about their way of life and the economy and where they feel like they don't have control that they don't have a full say uh in what that's going to mean.
1:46:18
They don't have a full say in what that's going to mean for their kids in terms of having good paying jobs and they're unsure if their kids are going to have as good a life as their parents had.
1:46:26
They don't know what that's going to mean culturally for them as citizens.
1:46:30
Are they going to have the same sense or are they just going to be manipulated by algorithms?
1:46:36
>> And they don't know what that means culturally as their kids are on phones in school and and and becoming uh sort of uh creatures with machines.
1:46:43
And so this whole concept of how technology uh is going to uh be uh something that empowers people and that people feel comfortable about as opposed to fearful of.
1:46:58
uh is the challenge in my view of our time.
1:47:01
Uh, and you know, I've I've gotten attacked from some people in Silicon Valley saying, "Oh, it's kind of a lite."
1:47:06
And I was like, "No, I'm not a lit.
1:47:08
Of course, I believe AI can do a lot of great things in medicine, in uh coming up with new disease and lowering costs, but uh I I don't think we can be uh oblivious uh to uh people's concerns about keeping jobs and keeping uh social cohesion and making sure their kids are have going to have a good economic future."
1:47:30
And so I've tried to be thoughtful about how we adopt AI uh how we adopt technology in a way that keeps the American dream alive and and benefits folks.
1:47:39
And I mean that's such a wide remit how we adopt AI technology because you can see it implemented from a chatbot that you know some random person uses or you know kids are using AI all the way down to you know deeper in some the bowels of some enterprise software product that you know no human was ever interacting with to begin with and then it's just streamlined a little bit with uh some some AI dropped in the middle of some big system.
1:48:09
Um, how are you thinking about uh creating some sort of taxonomy around AI?
1:48:14
Do you do you like a divide between generative AI and more traditional machine learning workloads?
1:48:22
Do you see a divide between consumer and B2B applications, self-driving versus what happens in a chatbot?
1:48:29
Like how are you thinking about actually breaking apart that problem?
1:48:34
because there's so much there when we say AI.
1:48:39
>> I would say that the key distinction is is AI going to enhance human capability or eliminate human beings.
1:48:48
>> That is the >> the distinction and that we need to figure out as a society how we get more AI that is enhancing human beings as opposed to just eliminating them. Mhm.
1:49:01
>> Let me share two thoughts on this.
1:49:01
Uh both of people who influenced me.
1:49:05
Steve Jobs described a computer as a bicycle for the mind.
1:49:11
>> He didn't say computers would eliminate the mind.
1:49:13
He just said it would make the mind go really faster and better. Yeah. >> Right.
1:49:17
And my view is how does AI do that? >> Yeah. >> And then Darren A.
1:49:20
Smogloo won the Nobel Prize at at MIT uh has this idea of total factor productivity. Sure.
1:49:27
Let me try to explain it simply.
1:49:29
If you just had AI replacing human beings and those human beings then becoming uh not productive, not only would you have frustration in our society, right?
1:49:38
I mean, who wants to just get a check without contributing?
1:49:42
People have pride, but you also wouldn't actually maximize total production because you have all these people who could be doing things who are not productive and or not being able to earn a living and spend money.
1:49:55
And so what he says is that there is some savings of that for consumers and for shareholders if a technology just eliminates uh labor.
1:50:03
But the best technologies like electricity, like automobiles don't just uh eliminate people.
1:50:11
What they actually do is they increase people workers ability to produce that they are technologies that increase human capability.
1:50:19
And so you have the benefit of the synthesis of the technology and the worker and that that is actually what transforms lives and he calls it total factor productivity.
1:50:28
And so my uh ideas around this has been how do we do that?
1:50:33
How do we make sure we just don't eliminate four million commercial drivers?
1:50:36
How do we make sure that the adoption of things is actually uh making us more productive and that it's being done in with respect to to to to workers and and and and capability. >> Okay.
1:50:49
But so so let's let's make it more uh specific because I I agree at a high level with a lot of that.
1:50:55
Uh but let's talk about like a specific role uh or job like truck driving.
1:51:00
Uh you've generally come out against uh uh uh or or have concerns around AI based job displacement with uh and with long haul trucking and truck drivers.
1:51:13
On the other side of that, if I'm a uh if if I'm running a trucking company and uh I want to deliver deliver the best possible service for my customers, it's possible that uh AI uh would be able to support that.
1:51:28
How how what kind of like policy do you think is uh uh right in order to create uh you want some guard rails around around the industry how AI should be used in in trucking?
1:51:42
I'd love to kind of understand more.
1:51:46
>> Yeah, I would say have a human in the loop.
1:51:48
And so, uh, what does that mean?
1:51:51
Uh, when I on a plane, you know, a lot of it is automated, but we still have a pilot there.
1:51:56
And I'm glad we have a pilot.
1:51:58
I I wouldn't want to just fly in an automated plane.
1:52:00
And so, does this mean that a truck driver's job may become uh more appealing?
1:52:04
Because right now, as you know, we have a shortage actually of truck drivers and more demand.
1:52:10
But if they have a a assist from a technology that maybe allows them to rest more, that's less taxing.
1:52:16
They're there for the edge cases if something is possibly going wrong.
1:52:21
They're there uh to deal with maintenance.
1:52:23
They're there uh to make sure that you have loading and unloading happening.
1:52:28
We can reimagine what the role uh of a truck driver uh is going to be.
1:52:33
And we can certainly have a temporary uh view that for the next five years that you should have the driver there.
1:52:39
Now that doesn't mean that at some point uh there may not be uh jobs or certain parts of things that don't require a driver, but it doesn't seem unreasonable for five years to say you want a driver in the loop and let's rethink the the types of of of jobs that that that will be.
1:52:57
And if we need the government to be helping invest in in uh in in these in the developing of this technology, fine, but do it in a way that's going to be complimentary with drivers.
1:53:09
>> Yeah, this is kind of happening already with Whimo where there is a human in the loop and it's but you know the ratio of TA operators to cars on the road is potentially higher than one one right now you know according to some reports.
1:53:25
Uh but over time I think the Whimo team expects there to be fewer and fewer humans in the loop over time.
1:53:30
The question is how fast does that happen?
1:53:35
Um and you're sort of proposing um maybe try and make that as gradual as as a as a process as possible because I mean you go back to like the elevator operator used to be a human now we use buttons and no one's really missing those jobs.
1:53:50
They phased out over time.
1:53:50
Um, I think the main thing is everyone is concerned about rapid job displacement.
1:53:56
Not necessarily the if if I told you your grandson can't uh can't be a truck driver, you'd say, "Oh, you know, he'll find a different job."
1:54:05
Uh, but if it's like every truck driver out of out of the job next year, that's obviously much more dis uh dis like disengaging to the US economy.
1:54:16
Is that how you think about it in terms of just timelines more than strict rules forever?
1:54:22
I think that's thoughtful.
1:54:22
There's a famous economist who once said in a gender time, jobs for the father, not for the son.
1:54:28
And by that, uh, he meant, look, we've got to make sure that people in their 30s, 40s, 50s, 60s have have jobs.
1:54:34
That doesn't mean that, uh, that's exactly what their kids are going to do or their grandkids are going to do.
1:54:39
But a lot of these human in the loop legislation, we're talking about five years.
1:54:43
We're not talking about 15 years.
1:54:46
Uh, and we're talking about uh, roles uh, evolving, right? Right.
1:54:49
I mean, it may be that there these evolve and then there's less of a need to to to hire folks uh down the line and you you have a natural uh transition of folks, but you're taking people who are workers and making sure that they're productive and they have a good life.
1:55:06
Let me explain why I think this matters.
1:55:09
Phone operators, which people often give an example, and Alex uh Alexis, who's at Chicago, had a great point about this.
1:55:18
That was 2% of the workforce.
1:55:18
Commercial drivers are 10% of the workforce.
1:55:20
You already have an anger in the country of so many people displaced by globalization, displaced by the concentration of wealth in some areas.
1:55:31
And you really want to throw into this mix a rapid uh mass job loss displacement of and then what?
1:55:36
Just compensate them and and have people stay at home and just get a check.
1:55:41
Like is that the society that we think is going to be productive or do we rather uh figure out how they have uh some role and some say in the transition be managed in in a way that is uh uh that considers their their uh their interests as well.
1:55:59
And that's, you know, and I get that this it's a good nature debate and people say, "Okay, Connor, you're adding uh some uh costs to to to the issue."
1:56:09
And if all you cared about was shareholder profits and minimizing consumer costs as your only holy grail and you didn't care about jobs and you didn't care about communities, then people have a legitimate uh critique of me.
1:56:23
But I would argue that that was the mentality during globalization and it's what's led to so much of the polarization not just of our politics in the United States but in the western world led to things like Brexit led to anti-immigrant sentiment and maybe we should consider jobs and communities not as dispositive but as a factor just like we consider consumer costs and shareholder profits. >> Yeah.
1:56:47
What's your what's your look back on how the Uber story played out?
1:56:53
Because that was a weird moment where there was a big push back from the taxi cab drivers.
1:56:58
Those jobs still exist, but they're just way less profitable because the medallion system has kind of been undone.
1:57:06
But if you want to make a if you want to make money uh driving someone, you can, but it's you're making less money.
1:57:14
Like do you think that we should have handled that differently if we could run back the time or do you think it happened slowly enough that it was actually okay and delivered enough value to the consumer?
1:57:25
Because Uber is one of those weird examples where the amount of, you know, taxi cab like activity, ride sharing activity, it 10xed and and more people take these these rides than ever before um in the taxi era.
1:57:38
And yet it did have remarkable uh impact on the market structure of that industry.
1:57:49
>> Well, I'd be hypocritical for saying I'm against Uber.
1:57:50
I take Ubers all the time. >> I'm the same way.
1:57:55
>> A expose, you know, next time I get into an Uber.
1:57:58
The uh but but I I'll say this, we we should have done more for the medallion owners, right?
1:58:03
I I tried actually in New York.
1:58:06
actually in New York. I this is beca before Zoran Mandani became Zoran Mandani when he was an assembly member he was really focused on a lot of these taxi drivers who had lost their medallion value yeah and were underwater and what could we do to to compensate
1:58:21
them and I had actually reached out to Jamie Dman who tried to do something uh to his credit uh through uh through through JP Morgan and it ended up not working out but you know we should have as a government done more to help those folks who who had medallions who lost all their value. And that's an example
1:58:39
And that's an example of something where we could have been more proactive.
1:58:42
And then there's a huge debate about Uber drivers and whether they're getting enough value and have enough say over uh their their lives.
1:58:48
I I would argue that that we need that.
1:58:53
And I'd argue we need national health insurance.
1:58:54
This is the biggest uh biggest area where if you're not going to be employed at a as a traditional employee, it would really help if people didn't have to buy healthcare on an exchange that has soaring premiums.
1:59:06
So, there are better things we need to be doing to help that Uber driver.
1:59:10
But do I do am I glad that there is a technology like Uber? Yes, I am.
1:59:15
I think it has uh created jobs and it has uh made life easier for many people.
1:59:22
uh you brought up uh mom Donnie.
1:59:22
He made a post I think it was yesterday or the day before that a bunch of Silicon Valley types were agreeing with which was uh you don't you don't you haven't seen that very often.
1:59:34
It was around basically around SMB deregulation making it easier to get a small business off the ground.
1:59:42
Is that should that be a more important conversation at in in in every state and region?
1:59:48
I feel like I growing up in California, uh I've seen so many businesses like try to get off the ground and you end up seeing like a finished uh like a finished restaurant that's just has its door closed because they're waiting on some some permit or something like that and it's obviously hard enough to start a restaurant uh and it seems like oftent times local governments uh can get in the way.
2:00:11
uh do you think that needs to be just a bigger part of the conversation as you know given that uh starting a business is a great way to insulate yourself from uh at least some job displacement risk with with AI? >> Yes, it does.
2:00:26
And look, Zoron became famous in part with his halal video where he was basically saying it takes too much regulation to have a halal stall and we need to streamline that.
2:00:37
And so, uh, I believe yes, we need to make it easier for people to to start a small business, to be their own business owner.
2:00:44
That's not just making the permitting easier, it's also making sure people have access to capital.
2:00:49
A lot of times that's a barrier.
2:00:50
But I'll tell you one thing that I think is often a a blind spot for uh, folks in my district. I love small businesses. I love entrepreneurs.
2:01:00
I think that there's a lot of people who want to build wealth. >> Completely agree. Completely agree.
2:01:06
We love small businesses here, too. >> But here's the butt. >> Okay.
2:01:12
>> Most Americans, most Americans are not going to go just start a small business.
2:01:16
Like this this idea that every person in Bucks County, Pennsylvania, where I grew up, or Western Pennsylvania, should start a startup or build a business.
2:01:26
Like my dad never did that.
2:01:26
He had a middle- class life.
2:01:27
He worked for the same company for 30 years.
2:01:28
And there are a lot of people who just want a decent job.
2:01:32
and they just want a job that can support a family.
2:01:34
And there's nothing wrong with that if they want to be in manufacturing or they want to be a nurse or they want to be a child care provider.
2:01:43
And so sometimes our rhetoric becomes like why can't everyone become an entrepreneur?
2:01:47
It's like why can't every become a politician?
2:01:49
Now maybe an entrepreur is a better life, but like a lot of people just don't want to do that and they still want to have the American dream.
2:01:56
And so all I'm saying is let's think about how to help small business owners, but let's also think about the four million people who are drivers and like what is their life going to look like?
2:02:06
Uh and uh and it's important to have that balance. >> Yeah.
2:02:11
>> Give me some lessons from the recent trip to China.
2:02:12
I'm fascinated by how they're dealing with AI.
2:02:16
Um are they doing anything right?
2:02:19
Are they moving even faster?
2:02:21
Do they have a solution to the job displacement uh problems?
2:02:23
Uh, is there anything good or maybe risky that you found going out there?
2:02:29
What were your takeaways? >> Three takeaways.
2:02:32
One, one-third of the AI talent is in China. >> What does that mean?
2:02:37
Uh, that means it it would be totally counterproductive to ban Chinese students from coming to the United States or Chinese entrepreneurs for coming to the United States.
2:02:46
We want to have uh that talent come to the United States because we still have a better uh ecosystem for capital and for investment.
2:02:56
>> Second, uh we need to make sure that we're developing the talent in uh in in AI here in the United States and investing in STEM and making sure that we're uh encouraging uh the the local development of that.
2:03:10
A third and this is the most important important.
2:03:12
Guess how much uh youth unemployment is in China? >> It's nearly 20%. It's really high.
2:03:20
>> You guys are you guys are too smart on this stuff. >> It's really high. >> 20%. >> That's crazy. How is that possible?
2:03:24
I feel like can't they just go build more bridges and create more jobs?
2:03:28
I thought I thought it was a command and control economy. I don't know. It's always >> enough.
2:03:32
They have enough to empty skyrises. I think >> maybe. I don't know. Yeah.
2:03:36
What was what was your take away from that?
2:03:37
And as I describe it to people, you can't build dating apps in China, right?
2:03:40
Like so, you know, the people who have these degrees. >> Is it banned? >> Yeah.
2:03:47
I mean, they they it's such a directed economy.
2:03:49
They want everyone to like make stuff, manufacture stuff, not do not do things that they would consider frivolous, right? >> Sure.
2:03:56
Sports app, a music app, all the cultural stuff that we do that improves consumer life or thinks about consumer needs.
2:04:04
needs. uh and uh you know so you're a someone who gets this fancy education in college and then they're like okay go uh work uh at a factory and just like we've undervalued people who want to work at factories in America we should be having more trade schools and more respect for
2:04:21
factory workers they've undervalued people who don't want to work at a factory and the reality is like you should have both choices so these people they they they're there and they don't want to go necessarily to build a bridge or necessarily to uh build a next uh factory of robotics. And the it was
2:04:39
And the it was hilarious because I would talk to the premier Lee Chang or others and they'd say, "Well, it's a voluntary unemployment problem.
2:04:47
These are just folks they they should be getting doing these jobs."
2:04:50
But what if in America we said, "Okay, you know, as as one of the news rooms when when they're they were being laid off said to someone, go become a an electrician."
2:04:58
Well, that's as offensive as telling a steel worker to become a coder.
2:05:01
to become a coder. like you know people do things and they want to do what they aspire to do and China is a command directed economy that has overvalued manufacturing doesn't have that diversity we do our problem has been the opposite that we undervalued making
2:05:18
things we undervalued the trades and so what we need is sort of a balance for America to have manufacturing but also this incredible ecosystem of the service economy uh which can employ people where China can't and that's ultimately why I bet on America. I'm also one pointer
2:05:32
I'm also one pointer sick of this argument that let's just go be like China.
2:05:37
They're where they're going to eat our lunch. Really?
2:05:39
You know, the Chinese model is a crony communism.
2:05:44
Like, okay, Gigi Ping gets rich and a bunch of people who are running these companies get rich and the rest and then you have 20% unemployment and you have consumer welfare declining and and look at how most people live.
2:05:57
They don't live in nice houses with, you know, two cars.
2:06:00
So like I don't want China as a model and I'm not going to compromise every American having economic security just because we're chasing China. China is not the model.
2:06:11
America needs to be more like America of how we built America in the 1940s50s and >> I completely agree.
2:06:18
>> Quick couple uh want your takes on a couple things.
2:06:20
Uh housing affordability.
2:06:22
I think a lot of people agree right now that uh housing affordability is is sort of like upstream of a of a lot of the problems that that we're facing as a country.
2:06:32
How what what's your current stance on on how we can impro improve affordability at at kind of the local level and at the federal level? >> I'm a Yimi.
2:06:42
I'm an abundance guy on housing.
2:06:45
>> We got to build far more housing in California.
2:06:49
You know, I I I don't endorse people who are sort of zero housing people in in in my district.
2:06:55
We got to realize that aesthetics matter, but economic uh equality of opportunity matters more.
2:07:02
And you can't have five trillion dollar companies in my district and expect to live live like where the valley of the hearts delight.
2:07:08
Like if you got that many companies, you got to have housing near transit and dense housing uh to make sure that people can live there and that it's not just a place where wealthy people can live and that the working and middle class is getting shafted.
2:07:20
We also need to stop private equity from buying up single family homes.
2:07:24
People say, "Oh, this is a red herring."
2:07:26
No, it's not a red herring.
2:07:28
In some places, uh they have bought up too much single family homes.
2:07:32
So probuilding uh pro- streamlining making it easier to build and uh having zoning reform uh and and stop private equity from buying up these single family homes.
2:07:43
>> What about international?
2:07:45
>> Will will we will we make progress at least in California on those issues in the next 10 years? >> Yes.
2:07:53
Because I I I think people realize we didn't make enough progress over the last 10 years that this is a failure of California policy.
2:07:59
uh and whoever is elected the next governor, I can't imagine it won't be on a abundance agenda when it comes to housing.
2:08:06
And it's not going to be, okay, let me do it at the last year.
2:08:10
Try to do something of of an eight-year term.
2:08:11
It's going to be day one.
2:08:13
How do we start to do things that it's going to build more housing?
2:08:17
So, I I I think it's been a wakeup call for uh for California. >> That makes sense.
2:08:21
Uh any quick comments on the current state versus federal AI regulation?
2:08:27
We didn't get to touch on that earlier.
2:08:30
Uh and and you had some comments uh recently on um SB 1047, the the bill in California, but what's your updated view on on uh where regulation should be happening?
2:08:47
>> Well, look, ultimately we need a federal regulatory framework, but the way you get good federal legislation is having legislation in the states. That's federalism.
2:08:55
And uh I don't understand how you would have a moratorum on having state legislation when federal legislation right now looks bleak.
2:09:03
The prospects of it are bleak.
2:09:06
Uh it is such an unpopular position even among Republicans.
2:09:11
So my view is uh build a consensus that you can have thoughtful uh regulations at the federal level uh and work on that.
2:09:20
Don't stop states uh from uh from regulating.
2:09:23
And this idea that okay, you're going to stop all the growth.
2:09:27
I mean, my district is $18 trillion of value.
2:09:30
We've got five companies over a trillion dollars.
2:09:32
East of the Mississippi, there's not a singleion.
2:09:38
>> You know, California's undefeated. It's so good.
2:09:44
>> You talk to folks in like Bucks County, Pennsylvania, where I grew up, and they're like, "Come on. Come on.
2:09:47
They producing more wealth than ever before."
2:09:51
Like what we want to know is how is how are our kids going to fit into this? Yeah.
2:09:56
>> And I I just think that that I wish more tech leaders, you know, who sometimes gets it is a Jensen Wong has talked about this.
2:10:02
I'm like, yeah, I mean about how do we create economic development opportunities in places that have been left out?
2:10:10
How do we make sure that everyone comes along on the AI revolution?
2:10:14
I I just think it would it's in tech companies interest to embrace this in in a similar way as uh the economic royalist embrace the New Deal eventually.
2:10:23
I mean you can't have just a capitalism that is only working uh for some with with large chunks of the country suspicious and and left out. >> Yeah.
2:10:34
I I just worry that we don't know the shape of what we're regulating yet.
2:10:38
like the unintended consequences of social media took 5 10 years to develop.
2:10:44
I mean, two years ago, we were reflecting on this.
2:10:46
People were worried about AI killing everyone and creating the Terminator.
2:10:51
And then what wound up happening?
2:10:53
Well, it wasn't really political misinformation.
2:10:54
It was much more people chatting with it for a really long time, going crazy, uh, you know, maybe overbuilding, maybe risk in the debt markets.
2:11:04
Like, the risks were very hard to predict.
2:11:06
there were risks, but it wasn't exactly what we thought.
2:11:10
And so I'm always I I'm I'm a little bit like hesitant about like, you know, maybe there should be regulations, but how when will we be confident that we know how to regulate it? Is it right? Is now the right time? Do we have clarity?
2:11:25
Because a lot of the stuff it stands on, you know, we already have fair use.
2:11:27
We already have copyright protections.
2:11:29
And so a lot of it can be enforced through the courts, I would imagine.
2:11:32
Um, but of course if if new problems come up, they need to be resolved and that's the way we resolve them in in a democratic society. >> I think that's fair.
2:11:41
The places I focus on are jobs. >> Yeah.
2:11:45
>> And uh American citizenship >> and I agree with you on the jobs part, but it just feels like the jobs we haven't seen a collapse and and even people building the AI technology are like this is going to put everyone out of jobs and that's good.
2:11:58
And then the people that hate the technology are saying it's going to put everyone out of a job and that's bad.
2:12:02
And it's kind of crazy because they all agree that the jobs are going away.
2:12:05
And yet what do you get when you actually look at the jobs figures?
2:12:09
It seems like we still have jobs.
2:12:11
Like it seems like we we actually can't delegate to the AI and I can't just say, "Hey, you know, trucker, like I I want the AI to handle this one."
2:12:17
It just it's just the technology is not there yet. And will it be a year?
2:12:21
Will it be 5 years, 10 years, 100 years?
2:12:26
There's a whole bunch of incentives to say it's coming right now.
2:12:27
uh and it's hard to get a read on and predicting predicting when things will happen is is you know fortunes are one and lost on that on that alone.
2:12:37
>> Totally agree with you.
2:12:37
John Maynard Kane said we'd all be working 15 hour work weeks and he was on more about economics than any of us.
2:12:42
So, you know, it's hard to predict, but I think what we can do is when you look at Darren A.
2:12:49
Smoggler who says, "Well, why don't we have a neutral tax code so we're not uh incentivizing depreciation of investment and technology and automation over hiring people?"
2:12:59
I mean, there are things we can do that make it that we we we prioritize having people in the loop.
2:13:04
And then there are things we can do in our social media environment that protect us as citizens and kids.
2:13:10
Two things are like, let's eliminate bots, right?
2:13:11
Elon Musk talked about doing this on X and there's still a ton of bots, but a lot of the bots that use AI are in my view uh hurting our democracy and then let's protect kids from some of the harms on social media. Yeah.
2:13:24
you know, so yeah, I guess I you know, I'd love I love sparring with folks and I appreciate sort of the criticism I've gotten from the tech folks for the the the tweets on on AI and and and drivers, but I guess what I would hope for tech people listening to this is don't resist uh uh every form of of of of regulation and and sort of dismiss people's anxieties.
2:13:49
instead be part of how we get smart regulations and how we answer people's concerns because if 70% of the American people believe the American dream is dead and have a concern about AI.
2:13:58
Like the answer to that uh for anyone who's been like in a relationship is not to dismiss it and say they're dumb.
2:14:04
It's to say, okay, how do I address that anxiety so that we can move forward?
2:14:10
And I guess I I I guess my hope would be that uh there'll be more tech leaders uh like that.
2:14:17
Victor Pang is one who was the former leader at AMD.
2:14:18
I mean, there's some people who are thinking in that way and I I I I think it's in Silicon Valley's interest to have that kind of view. >> No, that makes sense.
2:14:26
>> I think you really you really freaked people out with there should be a tax on mass job displacement.
2:14:31
>> Well, there is a tax on the profits, right? Like we we tax profits.
2:14:33
So, I mean it there there's a question of like maybe we adjust that, but it's it's all these are all dials that already exist.
2:14:43
We're just discussing how we turn them, I would imagine. Yeah. I don't know.
2:14:46
Well, a lot of times, you know, this is one thing different for me than other politicians is I I toss ideas out there.
2:14:54
If I think there's good push back, then I adjust my views and I I'm like a politician like this.
2:14:58
I just talk like I talk to someone someone over a drink over at a bar, you know, I and everyone else is like so scripted.
2:15:04
Oh, you can't put out an idea because, you know, maybe it'll come back 10 two years later on Face the Nation.
2:15:10
I just don't think that's what our politics are.
2:15:12
I It's like put your ideas out there.
2:15:13
What human being doesn't have some ideas that are dumb?
2:15:17
Like maybe maybe Einstein didn't or something.
2:15:19
Most of us, yeah, we put up good ideas, we put up bad ideas.
2:15:22
We think >> I love I love that.
2:15:23
I love that approach and it and it's and it uh certainly sparks a conversation >> and it certainly fits with what we've done here today. This was really fun.
2:15:29
We really appreciate you coming on the show and just like going all over the place and just talking through all this stuff. It's fascinating.
2:15:35
I'm learning a ton and uh we really appreciate you taking the time to come talk to us. >> Yes.
2:15:40
Thank you so much for coming on.
2:15:41
>> Well, you guys are doing great.
2:15:42
Seriously, you're you're elevating the conversation in Silicon Valley and it's an honor to be on and I look forward to being >> Yeah. Yeah.
2:15:49
We'd love to have you back on the show and and go way deeper on all of these and I'm sure uh by the time the next time you're on uh all the data points will be different and and we'll be looking at and we'll be staring at new problems and they will require new solutions and new discussions.
2:16:01
And so, thank you so much for taking the time to come talk to us.
2:16:05
>> I appreciate your approach. >> Thank you. >> Have a great day. We'll talk to you soon. >> Bye.
2:16:10
Uh before we bring in our next guest, let me tell you about numeral. com.
2:16:12
Let numeral worry about sales tax and VAT compliance compliance handled so you can focus on growth.
2:16:19
Uh our next guest is >> where do I have it? >> Jonathan. >> Jonathan Swerlin. >> Function. >> Hey, sorry.
2:16:29
>> We were in, you know, a political quagmire. We were in the swamp. >> We were in the swamp. >> We went to the swamp.
2:16:33
We don't normally go to the swamp.
2:16:35
Normally we talk about series B's.
2:16:37
We talk about large series bees.
2:16:39
He did get us going though.
2:16:39
He was telling us how much value has been created in in his district. It's in the trillions.
2:16:45
It's in the tens of trillions.
2:16:46
And we were just, you know, foaming at the mouth about the market caps.
2:16:50
Uh and then we said a bunch of other stuff.
2:16:51
But thank you so much for coming on the show.
2:16:53
Uh for those who aren't familiar, uh introduce yourself, introduce the business, tell us what's going on. >> Absolutely. Great to be here.
2:16:59
Now you're climbing out of the swamp.
2:17:01
We're going to talk about something a little less swampy. >> Thank you. >> We talk about health.
2:17:05
It's great to see you guys.
2:17:08
>> Well, I mean, health is health is like honestly more political than politics. >> It can be.
2:17:12
It can be, but uh but uh this conversation won't be. >> It's funny.
2:17:17
We say we actually say that biology is bipartisan though.
2:17:21
>> And I and I like this idea of everybody can agree that nobody likes to suffer. >> Yeah.
2:17:28
>> You know, and and everybody can agree that preventable death shouldn't happen. >> Yeah.
2:17:34
So it it it comes at but of course the nuance of how you get there can become political because who's gonna pay for it, right?
2:17:41
>> Oh that but also just >> well that or or the uh well this diet is >> this diet is rightwing that diet.
2:17:46
Oh working out that's a right-wing thing or like oh this is leftwing and like you know different ingredients became politically charged over the last few years.
2:17:55
>> My powder is better than your powder >> for sure.
2:17:57
And sometimes there's political influencers on the right and the left who actually have the same supplier and then they put different branding on top of it and they sell that.
2:18:05
That that's a fascinating rabbit hole to go down.
2:18:06
But anyway, we're not here to sell supplements.
2:18:08
Uh let's talk about the business.
2:18:11
Uh what what you know it's funny.
2:18:12
It's like I'm not leftwing, I'm not rightwing. I'm the whole bird.
2:18:16
Otherwise, you fly around in circles is kind of the idea with all this. >> I love it. I like it. I like it. >> The whole bird. That's great. >> The whole bird. The whole turkey. >> So yeah.
2:18:23
Uh take us through the shape of the business these days.
2:18:25
What what's the value prop to consumers?
2:18:27
Uh what's the progress been? How big is the company?
2:18:31
Kind of set the table for us.
2:18:33
>> Okay, so simple value prop is get on top of your health.
2:18:36
It's time you open your health.
2:18:37
So what does that start with?
2:18:37
It starts with a new platform. It's $1 per day to join.
2:18:42
And the platform includes twice a year comprehensive lab testing at over 2200 locations, Any Quest Diagnostics around the country.
2:18:48
You go and you test everything, heart, hormones, liver, kidney, thyroid, cancer signals, you name it. up and down.
2:18:53
All of that data goes into a platform into an app that explains you what's actually happening inside your body.
2:19:00
And these are the things that you would not get in a physical.
2:19:04
This is like a true true deep look.
2:19:06
And what function has created is this entirely new standard for your health that every year for the rest of your life, you know that you're well on top of whatever's happening inside your body.
2:19:16
You're seeing how it's changing over time.
2:19:17
You're making sure that you're getting well ahead of disease.
2:19:20
You're doing everything you can to feel your best.
2:19:22
So that's that's the the value proposition and that's what function delivers right now.
2:19:26
We started with lab testing. Yeah.
2:19:28
Because that's that's like that's the most impactful data.
2:19:32
70% of medical decisions are based on lab testing.
2:19:34
And recently we acquired a company you might have heard about this called Ezra.
2:19:40
>> And Ezra is an imaging business. And so Ezra does.
2:19:44
And what has been amazing for us, we've gotten FDA cleared AIs that have reduced the time that it takes for somebody to get an MRI. Okay.
2:20:00
>> So why does that matter?
2:20:00
One, nobody wants to be beside an MRI machine typically. >> Yeah.
2:20:03
>> Two, it also massively reduces the cost >> and it picks up the efficiency. >> Yeah.
2:20:08
And so you what we've actually done is we've introduced lab testing became one of the largest most powerful lab testers in the country and then we went into imaging on the imaging side bringing down the cost and what you're seeing actually emerge as this new standard for health.
2:20:20
We took the most impactful parts of the health system for capturing your data and we we packaged it up into something that's really simple to understand and really affordable for for many many people.
2:20:33
talk about how talk about how M MRIs were used historically.
2:20:38
Are these things that that get done when uh like you're act you have like acute pain or you have an issue and and then you're doing it and this feels like >> Terry ACL. >> Yeah.
2:20:48
Or this this feels like kind of flipping it and saying using it as like preventative uh preventive care. Is that the right read?
2:20:55
>> That that's the right read.
2:20:55
Not just preventive, I would just say I would just say responsible because this idea of preventive is great, but it's also what might be happening right now that you don't even know about, >> right?
2:21:07
And so the word preventive and the word early are a little tricky for me because the word early, it's like why is it early detection?
2:21:12
We just call it detection.
2:21:15
Can't we just get rid of the word early?
2:21:16
word early? Um what MRI does is traditionally it allows somebody to look inside the body but to do that it's been really really expensive to get an MRI you you know tear your ACL something like that like you basically um you you have to spend thousands of dollars to
2:21:34
look inside your body and that's the way insurance is set up and that's the way MRI is set up but what MRI can do is it can look at every single organ and look for tumors that are 2 cm 2 millimeters It can look for stroke risk, aneurysm risk, endometriosis, hernas tears, everything. So, if you actually want to
2:21:52
So, if you actually want to understand what's happening inside the body, an MRI is an incredible way to do it.
2:21:58
But it's been so arcane and so difficult.
2:22:00
It's never actually been architected and set up to look at the body and get well ahead of things.
2:22:05
And it's usually been, oh, you go in a hospital, you broke something, you have an issue, you go look at this one particular area.
2:22:12
In function's case, you can actually look at most of the body through an MRI.
2:22:15
and you can detect cancers early, you can detect aneurysm, stroke risk, etc.
2:22:19
Um, and you can do it for $499 and you can do it across almost 200 locations by the end of this year.
2:22:28
There's never been anything like this.
2:22:30
This is the first time in history this has been possible.
2:22:32
It is the first time in history it's been possible geographically from a cost perspective, um, technologically and culturally it's changing.
2:22:40
People are realizing this.
2:22:42
What I was alluding to before, it's a really important point is a new standard of health is emerging and that standard includes twice a year comprehensive lab testing.
2:22:50
It takes a 10 15 minutes each time you go in, you get your whole body tested, you find out what's actually happening inside.
2:22:56
And the second thing is now a quick MRI every year.
2:22:57
And if you do it, what you're doing is you're actually creating a baseline for your whole health.
2:23:04
And you're seeing how things are changing over time.
2:23:05
You're catching velocity.
2:23:07
You're seeing bad trend lines.
2:23:09
And you're also just flagging critical issues as well as finding out what you can optimize and what can be better in your life.
2:23:15
And what's crazy to me is the current standard like the status quo. We've all done this.
2:23:20
We've all gone into the doctor's office.
2:23:22
They test you for like 20 things.
2:23:24
You get a phone call in 3 weeks. You're good to go.
2:23:25
John, Jordy, see you in six months, a year, two years, whatever.
2:23:30
And you move on with your day.
2:23:32
And that's just this episodic once in a while very narrow perspective on your health that's gone. But they miss.
2:23:39
They're not looking at cancer and they're really not even looking at heart disease, the two leading causes of death, let alone metabolic dysfunction, hormonal issues, thyroid issues, and function looks at all that. I give you a crazy stat.
2:23:50
A new study just came out.
2:23:53
45% of people that were hospitalized for their first heart attack did not have what is considered high- risk cholesterol.
2:24:01
>> That should be terrifying. Why?
2:24:01
Because if you go to a doctor's office today, a regular old physician's office for a checkup, you get your LDL checked, right? You guys have done this. Yeah. >> Yeah. >> Okay.
2:24:13
>> That marker was born in the 1950s.
2:24:13
It's older than my father. >> Vintage. >> It's vintage marker.
2:24:22
>> Some people would say Lindy.
2:24:22
Some people would say that's Lindy.
2:24:24
Okay, let's just steal man for a minute.
2:24:26
They might say it's Lindy.
2:24:30
So, so look, um, there is no world where any top cardiologist say, I'm just going to rely on LDL cholesterol.
2:24:38
Basically, what every top cardiologist tell you, let's look at APOB, let's look at LB, little A, let's look at lipid particle size.
2:24:46
For most people, they don't, those words are far into them, >> but they should be.
2:24:49
I mean, it's this off guard stuff, right?
2:24:51
But so there are way better ways to look at the heart, but we're relying on something that's back to the 1950s as status quo.
2:24:58
And what function has done is for hundreds of thousands of people now, we've actually delivered a new standard of health that includes twice a year testing >> of everything that looks at your heart.
2:25:11
It's looking at your kidneys, your thyroid, your hormones, everything.
2:25:14
Women don't have to go to their doctor and ask for horn panel and get chased around.
2:25:18
instead they can actually get a look at what's going on with their hormones.
2:25:23
>> And then on the cancer side, real quick, cancer side, >> 400 you're you're four times more likely to survive cancer if you catch it early.
2:25:31
But right now, status quo is you have to wait till you have symptoms to catch cancer. >> Yeah, it's crazy.
2:25:37
>> There's no way that's okay.
2:25:37
I don't want that for my family.
2:25:39
So, and now there's technology where we can actually with an MRI as well as with a grail test that we test for many, many, many thousands of people, we can actually get way ahead of these things.
2:25:50
So, I can talk about >> Yeah. Yeah.
2:25:51
So, I'm I'm sold on the product.
2:25:53
Uh I think it's I think it's hard I think it's hard hard not to be.
2:25:56
It's it's the best kind of like value offering I think in health like period.
2:26:03
Uh, and I was sold obviously uh when you were raising uh preede uh back in the day, however been two two and a half years ago or something like that.
2:26:10
Feels feels like forever ago.
2:26:12
Can you give us like a I'd love to get your view on an update of like the the market uh structure.
2:26:19
A lot of companies have seen uh you guys weren't uh function wasn't the first uh first lab, you know, testing and and and health platform like this to exist, but your guys' execution and the growth.
2:26:33
I think you're one of the uh uh at least uh growing faster uh than than a lot of the the fastest growing AI companies that we're seeing out of the last year.
2:26:43
um give us an update on like the shape of the market, how you see the market evolving because like I was saying, a lot of people are trying to like ride uh ride your your coattails.
2:26:54
Um but uh I'm curious for for an update there.
2:26:59
You know, we the the market is realizing that the word consumer health has been this like dirty word for 20 years or something.
2:27:07
And it's not it's it what it is is it's premised on the most primary thing that we experience as human beings is our biology.
2:27:15
It's our it's our life experience.
2:27:17
And what's the LTV of your health, right?
2:27:21
You'd be willing to pay anything for health.
2:27:24
It's the most valuable thing in the world for you.
2:27:26
And so we're finally in a place where we can actually see technology and products broadly applied to health.
2:27:34
And so it's it's you're looking at a TAM that conservatively is $7 trillion in some >> give it up for 7 trillion TAMs.
2:27:43
John, hit the hit the hit the size gong for a $7 trillion TAM. >> Had to had to. Anyways, continue. >> I love it. I love it. No.
2:27:56
So, so, so look, this is this is people have been spending absurd amounts of money on their health through these massive service platforms like insurance companies and big health systems.
2:28:06
And finally, people are saying, you know what, health happens outside of the doctor's office, and I'm taking it into my own hands.
2:28:14
And what we're doing is we're bringing scientific and medical rigor directly into a a platform that people they themselves can sign up for.
2:28:23
They themselves can manage and so they can make decisions for themselves.
2:28:25
And that gets them way ahead of disease as you know as you saying before like this is this is not a it's not a trivial space.
2:28:34
I think it's it is the best it is the most anticipated service for AI.
2:28:40
AI. it is the best application of AI in the world is to our health because that is the major experience and so of we're we're we're surprised that that the the category and all the all the competitors aren't it's not that it's not bigger
2:28:55
that more people are jumping into this like we we know that people are going to try to ride these coattails but but um we just we are our head is down and our focus is in how can we deliver as much value per dollar for each one of our members. We have hundreds of thousands
2:29:08
We have hundreds of thousands of members, soon millions of members.
2:29:10
We have been growing really fast because we're at we're actually delivering something to somebody that has real real substantial value.
2:29:18
And at a time when a lot of technology can do a lot, it's it's like what are we really paying for?
2:29:25
And it's like >> where can where can people get started?
2:29:28
You mentioned it's a dollar a day. Um correct.
2:29:30
Does it take me through like the customer flow?
2:29:32
Is it just a website and then I go to the lab?
2:29:35
explain how people can get can get going.
2:29:39
>> Okay, so it used to be $999 when we started.
2:29:42
It was manual and it was per day. >> $1,000 per day. No, $1,000 per year. >> Sign me up. >> He said per year.
2:29:50
Don't Don't worry, John's just messing with you.
2:29:54
>> So, it started at a,000 bucks per year.
2:29:56
Then we worked really hard to to bring up the efficiencies and tech.
2:29:57
Got it down to 499 >> and and a couple weeks ago we announced it's now 365.
2:30:02
It's like when actually >> $365 per day because health health is an everyday thing and it's an understandable price and when has healthcare actually been deflationary. >> Yeah.
2:30:12
And then so so uh go to Quest Labs probably twice a year.
2:30:17
>> You go to functionhealth. com. >> Yep. >> functionalth. com. >> functionalth. com.
2:30:21
>> You just sign right up. >> Yeah.
2:30:22
>> Right there in theuler.
2:30:22
You sign up for your lab appointment. You show up at the lab.
2:30:27
You get your blood drawn. Urine collected.
2:30:28
You walk out 10 15 minutes later. >> Cool.
2:30:31
In 24 hours, results start pouring in. >> Yep.
2:30:35
>> Now, now your app is live. >> Yep.
2:30:36
>> And all the data is coming in. It's making sense of it.
2:30:38
And you test every six months. >> Six months. >> Got it. >> Exactly.
2:30:42
>> I think people I think people one of one of the reasons people underestimated this kind of category it was as it was emerging is so many people got burned on like DNA testing, open DNA, DNA like the 23 and me is you test it once >> then there's like zero incentive to retest, right?
2:30:56
You just you you did 23 and me, you have have the data. It depends.
2:31:00
Are you working on your DNA or not?
2:31:01
Have you been modifying your DNA?
2:31:04
>> If you modify your DNA regularly, you should probably be testing your DNA regularly. >> You never know. >> You never know.
2:31:09
>> I might have I might have rewritten my entire DNA.
2:31:11
All of it from start to finish.
2:31:14
Every base pair is different now. Sign me up again. I'm ready to go.
2:31:19
>> We have to study you if that's the case.
2:31:21
We're gonna have to We're going to have to bring you in.
2:31:23
>> John John needs to be studied.
2:31:23
Honestly, got ridiculous.
2:31:28
What does 25,000 diet cokes do to the human body? We're going to find out.
2:31:34
What is 500 diet cokes a year?
2:31:37
>> A dollar a day on function and four diet and and at least four diet cokes a day for John.
2:31:42
Uh we're we're actually running we're running a split test.
2:31:44
We have the exact same uh lifestyle.
2:31:45
We we show up at the gym every morning. We work out. We prep the show. We do the show.
2:31:52
We hang out with our family.
2:31:52
We just we're going to do that forever.
2:31:53
But John drinks Diet Coke and I drink uh >> Mattina Yerba Mate podcast in a Clint can from Andrew Huberman of course.
2:32:00
Uh and uh we're going to find out. Yeah.
2:32:05
Yeah, we're going to find out.
2:32:05
Well, thank you so much for taking the time to come chat with us.
2:32:07
We have a small bit of breaking news I want to get to before our next guest.
2:32:10
So, we will be seeing you soon. Oh, one last thing.
2:32:12
Uh give us the numbers on the last fundraising round.
2:32:16
I want to ring the gong for real. >> Yeah, let's do it.
2:32:18
Series B, $298 million raised, $2. 5 billion valuation.
2:32:25
But look, look, the the thing to think about here is that's basically a dollar for every American adult. >> There we go.
2:32:31
>> And so what what that is is that's a that's a vote on your health.
2:32:33
That's not just >> I love it. I love it.
2:32:35
Well, thank you so much for taking the time to stop by.
2:32:37
We will talk to you soon here, Jonathan.
2:32:41
>> Long live TVPN for for for the next one for the C. Come in person.
2:32:46
We got a seat here for you.
2:32:48
I'd be honored to have you in person. >> Be great. >> Let's do it, brother.
2:32:51
>> We'll talk to you soon. >> Great to see you. >> Goodbye.
2:32:54
>> Let me tell you about Vanta.
2:32:54
Automate compliance and security leading.
2:32:55
It's le Vant is the leading AI trust management platform.
2:33:00
Also, if you're running a NeoCloud, you got to get on Vanta because that's one of the criteria for cluster max. I'm not kidding. Not making this up.
2:33:07
this up. SOCK 2 compliance is a big factor in in in actually making it up the tier rankings for cluster max because of course if you're training on customer data you need sock 2 compliance you need the the whole process anyway the the the uh breaking news that I wanted to get to really quickly is Josh
2:33:26
Kushner is partnering with OpenAI OpenAI uh he says um we are excited to announce a strategic partnership between OpenAI and Thrive Holdings through our partnership OpenAI will become an equity holder in holdings and collectively we will set out to deliver frontier technology to our customers. For
2:33:43
For decades, technology has has transformed the world's largest industries from the outside in.
2:33:49
We believe the AI paradigm will be different in that some of the most profound transformations will now occur from the inside out.
2:33:56
We view the businesses we that we own and operate as the right reward system to build, test and improve industry specific products and models. So the race is on.
2:34:06
Is it inside out or outside in transformation? What's going to happen?
2:34:12
These are the new fast takeoff, short timeline, long timeline.
2:34:16
Are you an inside out guy or an outside in guy?
2:34:18
This is going to be the defining debate over the next couple days. >> Get ready to lock in.
2:34:23
We'll be covering it here.
2:34:24
We'll probably have some people on who are digging into this, investing in this, getting uh you know, have long takes, short takes, who knows?
2:34:30
But I want to get to the bottom of what uh this outside in versus inside out trans transformation will look like.
2:34:38
We've been digging in a little bit talking to some folks who are building companies uh buying companies.
2:34:44
>> Taylor says uh deal guy yuga.
2:34:46
>> This is the deal guy yuga. It's happening. It's happening.
2:34:47
Well, before we bring in our next guest, let me tell you about Figma.
2:34:52
Think bigger, build faster.
2:34:53
Figma helps design and de development teams build great products together.
2:34:58
We have Crisal Valenuela from Runway in the Reream waiting room. Let's bring him in. How are you doing? Good to see you again.
2:35:05
Thank you so much for taking the time to come talk to us on such a big day.
2:35:09
Uh kick us off with an intro a reintroduction on where the company is today and then the news.
2:35:14
I'd love to know about the news. >> Yeah. Yeah.
2:35:18
Thank you for having me again. It's been a while. Uh yeah. So big big news.
2:35:21
We just released uh our latest Frontier model, Gen Runway Gen Gen 4.
2:35:26
5, as >> it's a model we've been working on for like quite some time.
2:35:29
It's the best video model right now in the world, which is a pretty remarkable fit. >> Yes.
2:35:35
>> So, I think it's um it's it's it's pretty good.
2:35:39
It's pretty fun to play with.
2:35:42
>> That's that's my audio. I'm adding not not that. But perfect timing.
2:35:48
>> But but but let's play some of the video.
2:35:51
I want to see uh this uh the the demo videos that you put out, the examples, and I want to ask you a bunch of questions about it because uh it's a it's an extraordinary claim.
2:35:59
Google is a serious company.
2:36:02
They have a very serious asset in YouTube, and I'm fascinated by uh so first give me give me the uh the news uh video arena leaderboard.
2:36:12
That's the that's the ranking that you're using. How is that scored?
2:36:16
How does that actually work?
2:36:18
So it's a it's kind of like a a way of crowdsourcing like performance.
2:36:20
You basically ask people in the internet to vote against two videos and it's anonymous.
2:36:25
So you vote left or right and then as you keep on voting you accumulate more votes.
2:36:29
Um once you vote you can see like who you voted for but before beforehand you don't know.
2:36:33
before beforehand you don't know. Um and so over the last couple of months we've been working for like this entirely new way of I would say training both uh video models and image models in such a way that hopefully we thought it would like out compete uh others in the arena
2:36:48
and and we got results a couple days ago and yes we managed to basically out compete uh all other video models including both Google and OpenAI which is which is a very remarkable feat if you think about the scale of resources like um I think it's the era of Ilia was saying this is the era of research again and I agree. But it's also the year of
2:37:04
But it's also the year of efficiency.
2:37:06
Like really good, really focused teams with highly efficient like you know mandates can get really far. >> Um and so yeah. >> Yeah.
2:37:16
Tell me about what you optimized for here because uh Sora seems it's an incredible model and it it was for like a minute like whoa really mind-blowing.
2:37:28
Then I feel like I kind of developed an immune system for it and I can clock a Sora video and it feels like Sora was very much trained on Tik Tok almost or vertical vertical social media video.
2:37:40
And so what have been the breakout Sora videos?
2:37:42
It's been a lot of uh dash cam footage and uh doorbell uh nest camera footage and facing videos the model dramatically.
2:37:52
they have degraded the model a lot.
2:37:54
Whereas V3, it felt like it had uh it it it had a little bit of the Hollywood polish, but it was more like Michael Bay when I looked at it.
2:38:02
It looked very saturated. It was cool.
2:38:04
It looked good, but what you went for it feels a little bit more I want to say cinematic, even though that's kind of an overused term, but talk to me about what your goal was or even if you if even if you have a goal when you go into a training run like this. >> It does.
2:38:21
So I think there's an explicit goal and an implicit goal.
2:38:23
I think in a way all models, specifically video models that are more visually like clear or like perceptible have some sort of personality behind it.
2:38:32
And I think that personality reflects a little bit both the point of view of the company and like the way you want to train the models in the first place.
2:38:39
models in the first place. to to your point like if you want to make like like consumer slob and like quick like sharable stuff you're going to train the models just from the ground up very differently that for the stuff that we're trying to do which is a much more professional like high quality very
2:38:54
controllable set of like tools uh and so a lot of what you're like basically outlining is I would say the personality of the models in somehow also reflects the personality of the companies like if you're trying to sell ads you're going to do a very different model from from if you're trying to bake creative tools. Um, and so I don't think there's one
2:39:11
Um, and so I don't think there's one single recipe or one single ingredient.
2:39:14
It's more of a just like taste.
2:39:14
Like I think that word gets thrown a lot in research today, just taste.
2:39:20
>> And I think taste is both the research like what do you want to work on like having vision like having okay I want to pick this specific problems I want to work on and this is how we're going to solve them and this is what we've learned over time.
2:39:30
That's one form of taste and the other one more aesthetically is like what things look good like and that's on a construction site.
2:39:40
>> This is actually very pure taste. That's pure taste.
2:39:43
>> Look at hilarious, >> right?
2:39:45
Look at the motion of the donkey moving like the camera, the angles like the amount of data creation our team of artists and like filmmakers and like people have spent.
2:39:53
It's not it's not it's not like trivial to be honest.
2:39:56
I think that's also the taste component like shots like this.
2:39:59
It's like >> some of this is horrifying.
2:40:02
I mean I guess that's the point.
2:40:05
>> Really had to summon the demon on this one.
2:40:10
Are have you been inspired by anthropic at all?
2:40:12
It feels like somebody could put you in the anthropic for video bucket and that like they're just like extreme focus on code and and ignoring everything else.
2:40:21
And meanwhile, your competitors like are putting a lot of resources towards this, but they're not betting their entire business on it in the way that you are. >> Yeah.
2:40:31
I think it's a it's like a a mercenaries versus like visionary type of like I would say bet.
2:40:34
It's like you want to have people who who who feel like very committed to the vision long term and the way you do that is like you're very focused on like the culture and like that culture eventually shines in the product.
2:40:46
I think entropic has also that you can you can tell like who works there and like how they think and it's it's all very co cohesive in a way.
2:40:55
>> I think we spend somehow a similar amount of time like doing that in a way and and I hope you can tell via the models themselves that like that personality comes across nicely as well.
2:41:04
Um yeah and and I agree like you don't that at the end will be perhaps the most defining part of the companies that like stay in the long run like I think if you just throw money at the problem you're not going to get too far to be honest. >> Yeah.
2:41:18
Um what went into the actual training run?
2:41:22
Are you at a are you at a scale now where it's a meaningful capital investment to build a model like this?
2:41:30
We saw the scaling paradigm change from like, you know, maybe it's hund00 million to do a big frontier language model run.
2:41:37
Then we were talking about billion dollar training runs, bigger and bigger training runs.
2:41:41
The results are remarkable, but has it been a remarkable amount of investment to get here or are there more efficient ways to actually uh get to a frontier result without spending frontier money?
2:41:57
Yeah, I mean uh it's definitely not cheap like this is not like traditional SAS like you know like so you definitely have to spend more more money more resources but uh I think we've proven that like we are not spending tens of billions of dollars to get there and to like overcome the challenges and look to
2:42:14
be honest like the model is not perfect there's a lot of things we're going to improve and we're going to fix and we're going to do larger training runs and we do more over time but it's kind of a I would say the expense the most expensive thing is like the natural intuition the team builds around what kind of works and what doesn't work. It's kind of go
2:42:29
It's kind of go back to the idea of research states like you can't throw money at it.
2:42:32
You just have to spend enough time.
2:42:34
We've been working on Rhino for almost a decade.
2:42:37
>> And so there's a lot of you've learned over time about what works and what doesn't that informs a lot of the efficiencies on training.
2:42:43
And yes, like expensive models will like you'll need more money to train larger and bigger models.
2:42:48
Like if this is the worst the models will be, imagine them in like two years.
2:42:53
Like you're going to get there by training larger models for sure, but also knowing how to train them in the first place.
2:42:58
And that's the part that I think is hard to quantify per se.
2:43:00
Um, and and what I'm really excited about is is not only what the models can can do, but also like the efficiencies are not only on training, but on inference.
2:43:08
Like this is a price point that's very comparable to our previous models.
2:43:12
So, it's actually very usable and hopefully you'll be using it in real time very soon.
2:43:18
And so, that level of I would say efficiency at inference level, we haven't yet seen it.
2:43:22
And and and I think we're we're going to get there very soon. >> Yeah. Fascinating.
2:43:26
I mean, some of those videos are very, pretty remarkable.
2:43:28
Uh your unlimited plan uh is includes uh 2,250 credits monthly.
2:43:38
How much video can one actually generate with that?
2:43:44
>> Well, technically unlimited. >> Okay.
2:43:46
I was confused because it said it said there's still like a a credit credit system, but >> No. So So we have a queue.
2:43:51
There's like we we have like compute and you there's a queue and you get into the queue and you generate as like the queue becomes available.
2:43:57
If you just want to generate like fast, you pay for credits.
2:43:59
So, but but depending on how anxious you are with like your generations.
2:44:03
So, it's a measurement of how how fast you want it.
2:44:07
Um, but eventually you can just literally generate unlimited.
2:44:08
It's it's by the way I think no one else has a plan like that.
2:44:12
It's pretty a pretty good deal.
2:44:14
>> What is uh uh like what are the length of generations that are that are most commonly being done today?
2:44:20
And is that a metric that you track?
2:44:22
Like are people consistently is it is it like a 20 second scene that's the most common today?
2:44:29
And are you trying to get to two minutes or two hours?
2:44:32
Like how do you think about >> duration?
2:44:36
So well technically you can do you can do like arbitrary durations if if you want it but like the average scene duration in like a short film or a movie is like actually two to three seconds long at the most and that's actually been trending down like the scene right the scene itself it cut is like two to three seconds long on average.
2:44:52
Um, and so when when you actually when when people mean like I want to join a 45 minute like long thing, you don't want 45 minutes of like one camera like fixed.
2:45:02
>> You want like scene cuts and world and you want the character like a shot, a medium shot, a long shot and you have, you know, >> like that's a different problem from like creating one continuous long sequence.
2:45:14
So the long one continuous long sequence for me is less interesting than the like multi-shot approach where you can create much more compelling like narrative work.
2:45:23
And I think we're not that far away from that being a reality where like you can generate consistent narrative work like really good visuals really good stories like with the level of quality of the videos that we're seeing right now here but they're all tied together in a way that just makes it feel like cohesive with each other you know. >> Yeah.
2:45:41
>> Um and so that's a different problem I would say altogether. >> Yeah.
2:45:44
Uh there was some debate on the why doesn't the cursor for video exist yet?
2:45:50
Do you have any any thoughts there?
2:45:53
>> What's the cursor for video?
2:45:54
>> Basically a nonlinear editor like a Premiere Pro, a Da Vinci Resolve, an Adobe After Effects for video, cursor for video, like replacing the actual bones of the software that the that the editor that the video creator uses.
2:46:08
Uh there's been a couple apps that have spun up uh Runway originally the reason I was using it back in the day was for green screen uh for for Chrome basically.
2:46:20
>> Um it was fantastic for that and it feels like that uh building a canvas building NLE uh that feels like one potential pathway to victory.
2:46:28
Uh it's way it's also very difficult because you can't just fork VS code.
2:46:33
There are no leading open-source NLE's.
2:46:35
On the flip side, uh when you if you wanted to play nice with Adobe, you could be a vendor all the way Nano Banana is now vended into Photoshop and that could be a solution and um you know there's a variety of ways to to win.
2:46:50
I'm interested in hearing your approach.
2:46:55
>> Yeah, that's definitely definitely an interesting question and uh by the way, shout out for you for being an OG on runway since what, like 20 2019.
2:47:02
>> Yeah, something crazy. I love you. >> Yeah.
2:47:04
>> Yeah. So um so so uh well my two thoughts are first uh the art of like NLE and editing and film it's an art and there's a lot of like pacing and like details that are very nuance and specific it's it's it's about granular
2:47:19
details and it's hard for I would say model to or assistant to automate that level of like decisions that's on a purely NL side right but I would say at least for us more interestingly is the question of like do we need an N analy in first place, right? Like, do we
2:47:34
Like, do we actually need this primitives?
2:47:36
If you think about nonlinear editing, this idea that you're like stacking frames of video against each other and like you're cutting them.
2:47:45
>> Before it was with physical racers and now we have deter racers.
2:47:46
You're cutting things together.
2:47:50
>> My bet is that you probably won't need like anise like that whole paradigm will feel like a fax machine like in a few more years.
2:47:56
And so I feel that's somewhat what's happening with like the the the the Devons and the clogged codes and the codeexes of video.
2:48:05
Uh I just I do wonder if there's going to be an intermediate step.
2:48:10
Um or maybe it'll just be absorbed by the current NLE.
2:48:12
I mean I'm sure that's what your customers are are using, right? >> Yeah. I I don't know. We'll see it play.
2:48:19
But I'm not I'm not too fond of like, you know, pushing like better versions of Analis out there.
2:48:24
I think there's there's something around how you make videos and how you interact with this AI systems that just naturally allows itself with different primitives.
2:48:30
And if you think also about the fact that very soon you'll start to see this happen in real time.
2:48:37
>> Like when you make real time like narrative work or videos or experiences, however you want to call them, like you don't need to edit things async because you're generating on the fly and you're have people interact with them.
2:48:48
And so it changes, that's what I'm saying.
2:48:49
It changes the nature of like those things in the first place.
2:48:52
and and there's a transitional period where like you'll you we're seeing like NLE being augmented with AI, but I think it's that's transitory.
2:49:00
I don't think it's going to pay out like in the long run. >> Yeah. Yeah.
2:49:03
No, I think >> uh has has Hollywood capitulated yet? What's going on there? We we had uh it's funny.
2:49:10
I've been hearing I've been hearing more and more about Sunno from from not just uh guests and friends of the show, but just like random people out in the world.
2:49:18
It sounds like every single musical artist now is like using it in some degree, even if they're not willing to talk about it.
2:49:24
Uh what what is the the the case in in traditional Hollywood and entertainment?
2:49:31
You can't exactly hide that you're using AI video.
2:49:35
Uh it's basically out in the open immediately.
2:49:37
Uh and there's just so much like so much negative energy that gets focused on it.
2:49:42
Uh specifically from people that are within the industry, >> you know.
2:49:46
You know, I think like the the negative energy is like the water problem with AI, you know, like it's it's kind of this this unrealistic like and very noisy, not representative sample of what's actually happening within the industry.
2:49:59
If you go to LA, if you speak with the agencies, with the talent, with the filmmakers, with the studios, with the production teams, they're on board on AI like years ago, like months ago, like they're fans, they're using it, they understand it.
2:50:12
Of course, there's pockets of people who are like like more advanced than others, >> but I would say that the the narrative publicly hasn't yet catch up with that.
2:50:21
I mostly because might some people might not want to speak about it or it's much more interesting to say like all the negative things that to think to say the positive things.
2:50:29
Uh I would say Hollywood already has overcome that and they've they've they're pretty much on board.
2:50:35
I would say gaming companies are now where Hollywood companies were like a year and a half ago or two years ago.
2:50:41
So that's I would say an industry who's now catching up more to what AI can help them and how they can use it.
2:50:46
Um so yeah, I would say the some of those naries are a bit fake to be honest. >> Yeah.
2:50:53
>> Well, thank you so much for taking the time to come on the show busy day.
2:50:54
We appreciate it and uh I can't wait to play around with the new model.
2:51:00
We have a benchmark here, Bezelbench, where we try and recreate a very complicated shot from uh that we shot practically with uh a bunch of different watches uh with our with our intern or gap semester, Tyler Cosgrove.
2:51:15
Uh and it has a very the shots very long. It pulls out. It twists around.
2:51:20
It's it's a pretty complex shot.
2:51:22
And that's our current benchmark and we'll be testing and we'll let everyone know how it goes.
2:51:25
But thank you so much for taking time to come chat with us.
2:51:28
>> We'll talk to you soon update. Goodbye.
2:51:31
>> So, >> uh, let me tell you about Julius.
2:51:32
ai, the AI data analyst that works for you.
2:51:38
Join millions who use Julius to connect their data, ask questions, and get insights in seconds.
2:51:41
We have Vincent from Prime Intellect in the Reream waiting room. How are you doing? Great to see you.
2:51:48
It's been too long since last weekend. >> Thanks for having me. >> Congratulations.
2:51:53
uh master of really figure finding the one day that we're not live to launch your new news.
2:51:59
Uh tell us what happened on Wednesday.
2:52:02
The one day that we were off of streaming. >> Yes.
2:52:08
Um so excited to give you a rundown.
2:52:10
So basically for for the broader context with prime kind of like our broader goals really creating um open frontier models um and infrastructure for everyone to create them and um last week we released intellect 3 which is basically really like a scale up um towards um scaling RL and post training and creating like a sort model um especially for like more agentic tasks.
2:52:34
Um so basically what we did is we took GLM and and did a whole SFT stage and RL stage to create kind of like a state-of-the-art um 100 billion parameter um MOE model and really kind of like that whole infrastructure um is is kind of quite a challenge like from um like the RL environments to the broader like code sandboxes and the whole stack to do post training.
2:52:58
>> That's basically what we built over the last half year.
2:52:59
I think Will Brown came on the show to unpack some of it on the verifiers and environment side.
2:53:04
Um so basically that's kind of like what we released last week and really proved that kind of like we got performance um at 100 billion scale that um thus far in the open source only 300 to 600 billion parameter models like deepse for example achieved before.
2:53:20
So basically getting to better performance actually at a much smaller scale.
2:53:23
Um and I think in general it showcases that like um open models are starting to catch up.
2:53:29
Obviously I think quite interesting um is in in general seeing the trend that um not just with our model but also more broadly with other releases like deepseek today um and over the weekend that actually um they're also on par with like the closed models now.
2:53:44
with like the closed models now. And I think really our goal is so was almost like a a preview release but already sort of is um we basically released like a early checkpoint and we're actually scaling it much further um also on more like a genetic capabilities but
2:54:00
basically really like making it um sort of across like a range of task and really I think the foundation of this which is quite interesting is that we created this environment hub where anyone in the world can create one of these RL environments which we ultimately then included in a training run. So basically um different people in
2:54:14
So basically um different people in in the open source contributed actually to the RL environments that we trained on for those for this model.
2:54:23
>> So yeah give me a concrete example of like this shift of businesses that need to you know buy a model that has been trained in a specific RL environment.
2:54:35
You know we've heard the example of like uh someone's creating a clone of Door Dash and they're figuring out how to do Door Dash orders agentically.
2:54:40
Uh what else are you seeing?
2:54:43
What are some other good examples of when a business uh would pull this off the shelf from all the different opportunities from all the different APIs that are out there and create something I guess semicustom for a specific business use case?
2:54:58
Like what are you seeing out there? >> Yeah.
2:55:01
So I think what's interesting is like there's I think two buckets basically.
2:55:05
There's a bunch of these people like creating our environments for the labs like the Doish clones etc.
2:55:11
So basically to push really capabilities.
2:55:13
capabilities. So I think we're in this paradigm right now obviously where ultimately like scaling RL is the main way on on how these models improve right like we've seen it with OPOS or um with GBD5 and Gemini like there was mainly like I think a scale up in RL >> um but basically what we are seeing are
2:55:29
two things is like on the one side is like there's a lot of demand for these RL environments but then the other side RL is very sample efficient so you can take an open model and um and and then really create an RL environment for the specific use case you care about and scale capabilities for that. So I think
2:55:43
So I think good example of this was for example cursor with composer like that was like what's what's widely believe known it's like to be a scale up of an open source model and RL environment was cursor like like they basically just gave it like the the tools and and the things within the harness and application of cursor itself. Yeah.
2:56:02
So they trained basically that model um and like really on on getting really good at using cursor and I think we'll see the same play out like across all the applications where basically the the broader theory is um like every application every company will be an AI company or AI native and will um have an opportunity to really um post- train and use RL to make the models work specifically on on their application.
2:56:28
So even if you take examples of like say a Figma right like if they want to uh make make their platform agent take really they need to create an RL environment around Figma and post train on on that environment to be able to serve that within Figma like kind of like out of the box like the closed models won't be perfect at like really navigating and and making those applications agentic.
2:56:50
So I think that like that's the broader theory.
2:56:51
like that's the broader theory. I think really it's also it's like it's so like the the capital requirements are are much much lower than I think the big labs want to believe you like in essential it's like you can for like hundreds of thousands of dollars like post train a model right it's like to to be much better on your application and
2:57:08
then also to you are able to like the model >> that's one weird trick post train a model for 100k and create a better so I mean that that's basically what you're saying is that is that if I'm Figma as an example and I could use a frontier model that's really expensive expensive and beefy and it knows everything about it knows some stuff about Figma but it also knows about the Roman Empire. uh I
2:57:26
also knows about the Roman Empire. uh I can go and RL on just my particular application and have a smaller model that's fine-tuned on open source you know uh open source model and and get better performance than with the big beefy you know do everything omni model is that right >> exactly and I think really you get
2:57:47
better performance but also at a lower price point potentially right because you can really specialize the model to be extremely good for your use case so I think you could see this with like cognition posting their own model with like um cursor postering their own model composer and composer is also it's like
2:58:02
it's much cheaper to serve it's much faster like same for for the model cognition was building so I think what we're seeing and and we we've started to work with like dozens of customers on like helping them basically do post training and RL yeah >> I think um >> we're basically starting to see a huge pull in terms of like enterprises
2:58:19
realizing that like if they want to get a specific capability um RL is the way to get it and and ultimately enables them quite capital efficiently like to to train those models and surf those models and then really get to like a point where even in deployment like all the interactions from the user help improve the model. So I think with the
2:58:36
So I think with the cursor example like every for example cursor tap interaction every yes and no that the user gives to the model is updating the model every two hours.
2:58:44
So it's what like talks a lot about like online RL >> yeah like like they're basically retra like continuously training the model in two-hour interval and pushing updates every two hours to first of app.
2:58:54
every two hours to first of app. So basically every user using purser for the last two hours is um is being post trained on so to speak like with kind of like an online RL loop and I think that's something which we'll see more and more that basically applications will do their own RL their own post training
2:59:11
>> um actually then and that's like really how we unhubble basically towards AGI where it's like the question is like why haven't we say automated um >> like specific valuable knowledge work yet and I think the answer that also like Shelter was speaking about for example on on the example of like autom
2:59:28
automating text and accounting for number right it's like no one has really created RL environments post trained on them and then serve the model in the application where the end user is and then ultimately the the end user's interaction with the agent can improve the model further right um so I think
2:59:43
that's really the paradigm that we see play out which I think is really a paradigm of like thousands of models or like millions of models that like basically continuously improve and where actually the applications uh win to some extent through distribution like ultimately they own the the end customer um interaction, right? Where it's like
2:59:58
Where it's like even the cursors and and um cognitions have like an advantage there over folks who basically just model providers and who don't interact with like millions of developers and I think we'll see the same play out like across all the different applications.
3:00:12
different applications. Um and it's something like for example s talk about also in the context even of like copilot and Microsoft right like they own distribution they can like create the cursor for excel for like powerpoint or other things right and then pull strain on all those interactions so I think
3:00:27
we'll see this like play out I think across like all the different verticals and I think it's like a border trend of just like every company needs to become AI native right like >> and own also the keep owning the distribution like they don't want to give all of it up to the to the big HR labs. >> Yep. That makes sense. >> Yep. That makes sense. >> Makes sense.
3:00:42
We got a question from uh our intern Tyler.
3:00:44
If we can shoot over there.
3:00:47
>> Um yeah, I guess I I I saw you guys talk about this a little bit online.
3:00:50
Um but is there any like point of you guys uh training your own base model? Yeah.
3:00:56
So basically I think one interesting release in this context was like um today we actually released uh like we supported RCI in in their base model release uh which is like kind of like catching up to um the the Chinese base models.
3:01:11
So basically we we supported them in training um a small base model which achieves like pretty sort of results.
3:01:19
So we released that I think like an hour ago with them and uh and we're actually now like ramping up with them to towards like a much bigger uh base model.
3:01:28
Um so fully like pre-trained from scratch.
3:01:31
So we actually just have like 2,000 B300s going live I think yesterday uh to ramp up like towards like a much bigger >> um pre-train.
3:01:41
And I think like it's really I think like the broader pattern is like since kind of like Llama had some reorgs and changes and misrel became sort of like a for deployed European enterprise play or something.
3:01:52
I think there's really no one left outside of China right now to go end to end in the model stack.
3:01:57
I think um others like reflection I think are trying to um also pick that up but I think there's very few players I think outside of China.
3:02:04
few players I think outside of China. So I think that's our broader goal is really is like serving um like the the world more global um globally but also like the west and the US um with like an end toend pipeline right it's like from data to pre-training to m training to post- trainining like the full stack and making that accessible to like enterprises and people who are like
3:02:21
training their own models so I think like there's a huge I think pull where a lot of enterprises or even like like sovereign like nation states etc like they can't train on Chinese open models but they also they can't rely on closed models um so I think there's a huge gap in the market right now that we're trying to fill of of really like serving kind of like that uh whole segment. >> Do you have anything else, Jordy?
3:02:42
>> Do you have anything else, Jordy? >> No, this is great.
3:02:44
>> Um I I want to know one last question about you know what what will the market structure look like in uh maybe a year or two around like implementing these RL environments for companies because when I when I see you know you say every every company is an AI company.
3:03:02
I I believe that's somewhat true.
3:03:05
Uh and I believe every tech company, maybe every founder-led tech company under 10 years might be able to say, "Okay, yes, we're going to go and train fine-tune a model and bu turn our our application into an RL environment."
3:03:19
But if I'm, you know, the Coca-Cola company, uh, you know, I might not be at that level of like going and building RL environments for every business process.
3:03:28
I'm probably more of a buyer of this AI as SAS almost.
3:03:30
Uh, so how do you see that kind of breaking out?
3:03:36
How do you see like a truly legacy, you know, non- tech company adopting a fine-tuned LLM or an RL or an RLED model? >> Totally.
3:03:49
No, I I think there's like early adopters and later like later adopters.
3:03:52
I think Coca-Cola might be more like the late adopter and might not need to adopt it early on, but I think they are adopting it just like in less obvious places, right?
3:04:00
It's like ultimately I think they're initially just like using the AI tools that use us for example right in a sense where it's like say customer service right like is a like perfect example of like where you get a lot of gains out of post training and then like they might like like basically the AI native customer service platforms might use us to post train using Coca-Cola data. >> Sure.
3:04:21
>> Um to serve them a better like model.
3:04:21
>> Um to serve them a better like model. So I think what we'll see play out I think is really um just like making like a lot of that like so accessible to your point that it feels more like using SAS >> where I think like one element of it is like we are like launching also like our whole like like RFT platform basically and and offering to make it extremely
3:04:41
like easy and plug and play but then there's also like a for deployed element right where you can outsource a lot of that stuff to our team and I think the other element is like really like we're walking the walk in in terms of like making our own thing kind of like agentic and autonomous that you could basically just use like an autonomous AI researcher to do all of it for you, right? Like that you basically just like
3:04:58
Like that you basically just like plug it into your system and like the AI even like creates AI for you. Yeah.
3:05:02
Like like and I think like like I think that's the next paradigm is really making a like making in general training models like fine-tuning models, post training models like as accessible as VIP coding is today, right?
3:05:15
In a sense, it's like I think with VIP coding, like literally every human on Earth is able now to like code some stuff up and I think we'll see the same play out with AI over the next 12 months and that's one of the big things that we're playing into.
3:05:26
We're kind of like pushing towards like autonomous AI research where AI can do most of it for you.
3:05:32
>> Well, thank you so much for taking the time to come and talk to us on the show.
3:05:36
Congratulations on all the project >> and we will talk to you soon.
3:05:39
>> Great to see you, Vincent. >> Goodbye. >> See you guys. Have a good one.
3:05:42
>> Let me tell you about Privy.
3:05:42
Privy makes it easy to build on crypto rail.
3:05:46
Securely spin up white label wallets, sign transactions, and integrate onchain infrastructure all through one simple API.
3:05:50
And I'm also going to tell you about adquick. com.
3:05:52
Out ofome advertising made easy and measurable.
3:05:54
Plan, buy, and measure out of home with precision.
3:05:56
Our last guest of the show is Ben Hilac.
3:05:59
Did you do the Jaguar rebrand? >> That's him. Ben, welcome.
3:06:06
>> And we'll follow him forever. >> How are you? How are you? >> Grab a seat. Hang out. Good to see you. Oh, you brought hats. Fantastic. Thank you.
3:06:14
Please uh grab a seat, introduce yourself, introduce the company. What's the name? >> Yes.
3:06:19
So, my name is Ben Hilock.
3:06:19
Um >> yes, >> let's take a second for the flow. >> Fantastic.
3:06:26
>> This is kind of like a vintage Silicon Valley flow that you don't somewhat of a lost art. >> I appreciate it.
3:06:30
You guys have great hair as well.
3:06:32
You know, I th a lot of pressure.
3:06:33
Um you'll notice I'm not wearing a hat today and it's because I did notice actually.
3:06:37
Yeah, kind of I discovered a blow dryer I think around uh 9 months ago, 10 months ago.
3:06:42
So that was a big never been the same since.
3:06:45
Um but yeah, my name is Ben Hil as you guys know.
3:06:48
Uh I'm the CTO of a company called Raindrop.
3:06:51
So really simply put um we monitor agents in production.
3:06:55
So uh we were building a product ourselves uh probably around 2 years ago now which was like a coding agent and um we realized that there was just this huge gap of like if you're using Sentry if you're using traditional analytics um you know uh they're covering like the things the users are clicking and almost everything that's happening in your product if you're making an agent is just not covered.
3:07:18
So you just have no idea what's going on.
3:07:20
>> These agents are going absolutely wild.
3:07:22
They're going they're going haywire.
3:07:24
>> You know what's been insane?
3:07:24
I think one of the things that's been like really kind of critical to our growth in the last couple months has been realizing that um as agents get better this problem gets worse.
3:07:33
So that that was not necessarily intuitive to us in the beginning.
3:07:37
You know you think like oh well agents are going to get better maybe this problem becomes less important but it's like actually as they become more capable they can use more tools more valuable. >> Exactly.
3:07:45
So, for example, if you take a company like Replet, it's like, you know, maybe a year ago or two years ago, um, uh, or when they first launched, um, you know, you couldn't quite get as far, right?
3:07:56
Maybe you could just get like a personal website or something.
3:07:57
And, and so if it messes up at that point, it kind of gets stuck.
3:08:00
It's like, okay, maybe it's not the end of the world.
3:08:03
>> But now with Replet, you're able to build just like real applications, like people are building real production applications.
3:08:08
So now, if you get to a point where it gets stuck, something goes wrong, suddenly it's like it's a real issue.
3:08:12
Um, so that was not intuitive.
3:08:14
uh before but >> so uh agent's a pretty overloaded term at this point.
3:08:19
Uh I I think of >> you know when I fire off a deep research report in chat GPT that's an agentic workflow uh to some customer service agent that's happening completely behind the scenes and the customer might not even know that they're dealing with an agent.
3:08:36
Uh and then there's coding agents.
3:08:38
There's a few that you mentioned.
3:08:39
Are are you uh dividing the market and trying to focus on an early landing zone first or do you want to do all of those? >> Yes.
3:08:45
So, we focus on essentially uh and I will say I agree the word agents overloaded.
3:08:50
We're very hesitant to use it for a really long time and then we realize it actually matters of course.
3:08:54
>> So, we focus on products that have some sort of user input and some sort of uh assistant output eventually.
3:09:00
So, that that's sort of our focus.
3:09:02
So what we we're not focused on is for example like uh we're not going to focus on like specific like ML pipelines or things like you know maybe like translating text or like summarizing text even it's like we want to see like the user the user is sort of like has some sort of request the assistant is responding to that request.
3:09:18
Um and we do uh map essentially everything that happens in between that initial uh you know user input and to what the actually what the assistant actually responds.
3:09:29
>> Uh and then what's the go to market for you?
3:09:31
I mean, it's been a little crazy, actually.
3:09:34
We've had a lot of inbound.
3:09:35
So, some of our biggest customers have been inbound.
3:09:37
Um, a lot of it has been like when we first launched, I think, uh, uh, like I guess this was like 6 months ago or seven months ago now.
3:09:44
Um, agents weren't as big of a deal.
3:09:46
And so I think in the first month or two, we had a lot of customers that were like, "Okay, like I have evalu." >> Sure. Sure. Sure.
3:10:12
uh how are you thinking about the you know target like the best type of customer?
3:10:18
Are you segmenting it by size?
3:10:20
Do you want to go enterprise upfront because they're implementing agents at scale or are you more likely to see immediate results at the startup that just kind of gets it and they can hop on really quickly?
3:10:30
Like how are you thinking about prioritizing if you are at all?
3:10:35
>> Yeah, it's a really good question.
3:10:35
I think that we really look at the entire range and I think that we see and have always seen startups as being a really core part of keeping our company healthy. >> Sure.
3:10:45
>> Um you know I I heard a while ago that like Post Hog has this metric where they look at like what percentage of YC companies in every batch are using them. >> Sure. Sure.
3:10:53
>> Um and so that's why we started with startups like uh they're always they're able to move faster.
3:10:58
So for example like when a new model comes out just give actually a very specific example.
3:11:02
So GPD5 um introduced intermediate reasoning right there.
3:11:05
It was kind of one of the first models to do this where like it's going to make tool calls.
3:11:08
It's going to look at the results of those tool calls, think about it, and then make more tool calls.
3:11:12
Take that, think about it, write, you know, more tool calls.
3:11:16
>> It sounds small or subtle, but actually it kind of means that you know if you architected these uh your system, your pipelines in the wrong way, you just couldn't use that like and and it and it really helped.
3:11:27
Um, so where startups will just like d just throw everything out the next day, right?
3:11:31
And they'll they'll ship a whole new thing in a week.
3:11:35
You don't see like, you know, like uh if you look at like the biggest enterprises, they're not going to do that.
3:11:40
Um, so you can learn really fast by with startups.
3:11:42
That being said, on the flip side, I think that the problem we're solving is actually most painful for enterprises, right?
3:11:48
It's like the the the the most critical highstakes environments are where like failures cost the most in every single sense. >> Yeah.
3:11:58
>> How much categories of agents that you're excited about that are maybe underhyped today? Coding agents.
3:12:03
Coding agents are like sufficiently hyped.
3:12:05
I think >> coding agents are and for good reason for good reason.
3:12:11
But like uh and may maybe they're deserving of more hype.
3:12:13
maybe they're deserving of more hype. uh who knows but uh but uh what what other category you know I think I think people have been sold on the >> AI BDR >> uh haven't exactly may maybe companies are getting a ton of value from it and they're getting so much value they don't
3:12:28
want to come on TVPN and talk about it because they don't want their competitors to know >> um but uh uh and then obviously like CX feels sufficiently >> sure uh hyped but uh what else are you seeing >> man it's it there's so many different things like I think um you know, speak for example, language learning. I think
3:12:45
I think the better like >> as models get better, that experience just actually starts to become really really really viable.
3:12:52
So like that's an example of something where it's like yeah it existed a year ago, it existed two years ago, but like as voice models get better, as like the models themselves get better, it can it's actually not just like >> you know if you try to use chatbt for example to learn a language, you sort of can.
3:13:05
But if you ask it to like critique you for example, um it just never will.
3:13:10
Like if you say something wrong, it just isn't going to stop and be like, "Hey, look, actually >> it's still glazing." Absolutely right. >> Yeah, exactly.
3:13:18
>> Esta bibloteka is the most complicated Spanish sentence. >> It will. It will. Right. Um >> you're fluent. Exactly. You're fluent. You're It's like Yeah.
3:13:24
You're pretty much good to go.
3:13:25
And um even if you can get it to the point where like if you can really really like prompt it into critiquing you, it'll just like start critiquing everything, you know, which is also not what you want as like you're learning a language.
3:13:34
So like it turns out I think we see this with a lot of products that like getting something right is actually a lot of details and really really understanding that domain.
3:13:43
So I think we're seeing that in literally every domain like whether it's like marketing, whether it's like even just like the idea of having a personal assistant like notably we don't have that yet which is crazy, right?
3:13:54
models, but then none of us have chat and be like, "Hey, send this email." Right.
3:13:58
I don't think we're actually nailer mostly.
3:14:04
But >> how are you thinking about um just I I don't know if I don't know if like if you're Century for AI agents, does Sentry actually handle this?
3:14:12
But just types of AI failures that happen for more infrastructural reasons.
3:14:17
So just the GPUs are on fire or like there's just not enough GPUs in this particular cloud and you just see a spike in demand and so you just can't provision more like those types of more more tactical errors. Do you help with that?
3:14:30
>> Sort of would be the answer.
3:14:30
>> Sort of would be the answer. So um I think it's actually really interesting is that what one thing we realized about eval is that they don't catch those sort of issues like you know you're kind of testing just like the model what is the model responding but then there's all of these things that happen in between like
3:14:44
I remember really really early on when we launched one of the issues that a customer caught was like their file upload was broken so a bunch of users all started complaining about like oh like the file uploads taking too long like okay well it's not like an AI problem but it is um and so we see that with like tool calls um we saw uh one of
3:15:00
our customers had an issue sort of what you're saying which is that like they started having like they they have their own GPUs they started having like an infrastructure error and um it was mixing up responses between users and so users all started complaining like hey that's not what I like what are you talking about that's not my was like an increase in that in like
3:15:17
>> I don't know if you're talking about meta but I think that happened in meta >> it wasn't meta they're not one of our customers yet but there was a situation right people could share it was it was not that bad but it was something like I could share my chat with you but if I shared it with you and I didn't know that I was sharing ing it. It would go
3:15:30
It would go out everywhere and so yeah, stuff like that happens. >> Totally.
3:15:33
There's all these sorts of things.
3:15:34
So, you can actually catch those sort of problems.
3:15:35
It's actually one of the one of the things is like >> uh that ground truth is actually really really important because if you just see like a few errors like let's say you have tool call like your agent calls tools like yeah it's going to error once in a while, right?
3:15:48
Like that's might not be the biggest deal but especially once you if you can see when it actually starts to affect users like that's really that's really powerful. >> Yeah. Yeah, that makes sense.
3:15:54
Um, what about uh degradation of models under the hood?
3:16:00
I feel like people I don't know if it's just a meme.
3:16:02
I've noticed it here and there.
3:16:03
I'm not I'm not benchmarking everything every night like some big companies.
3:16:08
But it does feel like that sometimes, right?
3:16:09
It feels like it feels like sometimes I'm like, "Wait a minute.
3:16:13
They it used to respond in this many tokens.
3:16:15
Now it responds this many. It used to look HD.
3:16:16
Now it looks standard definition."
3:16:18
Like >> I know I agree with you.
3:16:19
I think I think it's real.
3:16:22
I know that I I can't say too much.
3:16:24
I know that at least on one occasion that I I think people were led to believe that there wasn't a thing. I I know that there was.
3:16:31
So that's you know what I mean? I can't say who. It's a big company.
3:16:34
Um and because I I noticed this and I thought it was >> I can't say whose hands were >> I can't say which one of the caught some red hands. >> Yeah, exactly.
3:16:42
And and like it it was like I thought it was a cursor problem.
3:16:44
It was like some really absurd behavior.
3:16:46
And then I went into chat and it was doing the same.
3:16:47
Oh, I just said but um but anyway, yeah, like I I I think that the reality is that like every single one of these uh providers are like having these sort of problems and they're trying to optimize costs.
3:16:58
They're trying to like make changes and like so I think it's natural >> and some of them I understand where I'm like okay well yeah realistically I haven't used that in a long time.
3:17:03
I came back I kind of I don't really mind that you put me on the lower tier.
3:17:07
I just hope that for the people that actually like went and built businesses around this that are using at the API level that are hopefully paying for the service at a high gross margin to you.
3:17:18
You're not degrading the service behind their backs like >> 100%. >> Right.
3:17:21
So anyway, uh who did the deal? Anybody we know?
3:17:26
>> You want to hit the gong?
3:17:27
>> You want to hit the gong? >> Oh, let's do it. Yeah. Yeah. >> Hit hit the gong.
3:17:29
Uh tell us how much you raised.
3:17:33
>> How much did you raise? How much did you raise?
3:17:36
Uh, we raised So, we raised $15 million in total. >> Um, uh, Light Speed. >> Who did the deal? >> Bucky. >> Yeah, let's go.
3:17:43
>> Let's hit it again for Bucky.
3:17:44
>> Let's hit it again for Bucky. >> Bucky.
3:17:45
We love >> This one's for you, Bucky.
3:17:50
>> Uh, yeah, we're big fans of Bucky over here.
3:17:51
So, it just wants to get him a shout out. >> Us, too. Us, too. Us, too.
3:17:53
I think when the moment we met him, we're like, "Okay."
3:17:56
Like, he matched our energy. Like, great vibe. >> Yeah. Yeah.
3:17:59
He's doing >> How's building the team going? >> Uh, it's going. It's going.
3:18:02
Uh, I think we're really, really picky. We've realized.
3:18:06
Um, and so it's really hard.
3:18:08
Um, and I think hiring in San Francisco is really hard.
3:18:10
Um, we have a great team.
3:18:12
Uh, it's honestly really, really small still.
3:18:14
Um, we're >> well, if you want to get out of San Francisco, you could book a wanderer with inspiring views, hotel, great many, dreamy beds, top cleaning, 20% customer service.
3:18:23
>> It's a vacation home, but better.
3:18:23
You could do an off-site there.
3:18:25
>> We could do our offsite. That's beautiful. Do your offsite.
3:18:27
>> I once used a team offsite as a recruiting tactic.
3:18:30
Like I said, we are going on an offsite in two weeks and posted a picture. >> Oh, yeah.
3:18:36
We got an amazing creative urgency. >> We're doing it.
3:18:40
So, if you're watching right now, uh we'll I'll post the picture soon of of the house. >> Fantastic. >> Fantastic.
3:18:46
>> But we have an amazing team. >> Yeah. Know. Yeah.
3:18:47
I figure if you're if you're picky and you're in San Francisco, it's like the most ruthless like talent war constant.
3:18:55
>> You know, the other thing is that that I think when you hire amazing people, they have zero tolerance for working with people that are not amazing.
3:19:03
fool yourself as a founder if you like whether you're just feels that and >> have you had to bring anyone's soup.
3:19:12
Are you familiar with this?
3:19:14
>> I'm not familiar with this. >> Okay.
3:19:15
So, apparently the AI this is from Ashley Vance.
3:19:17
This is a scoop just dropped on uh core memory on the podcast.
3:19:21
So he had Mark Chen, OpenAI's research chief on the show as part of a postGemini 3 sitdown to get the update from OpenAI.
3:19:27
And he said, "I knew the AI talent wars were rough, but not this rough.
3:19:32
Zuck is out there apparently delivering handmade soup." >> Wow.
3:19:38
>> And Open AI has soup counters and so I guess uh >> Wait, they count how much soup is soup counter.
3:19:46
>> I don't even know what this means, but >> Oh, I see. They count how many?
3:19:50
>> No, no, I think it's like a count like just like you have soup quite >> aggressive.
3:19:55
What exactly does this tit for?
3:19:58
>> We can we we we can play this on the show later, but uh but yes, I mean >> no no my my my partner has some meals for someone like you know full meals like that sort of thing works.
3:20:06
Um you know we we do typewritten I'll write a note on a typewriter you know when we do our offer letter.
3:20:12
So that that adds something a little Are you messing with us?
3:20:17
>> I love No, I'm serious. I love typewriters. >> No, I I I like that.
3:20:19
It's just a way actually value this message.
3:20:23
>> All the text is AI generated. I'm sure. >> Of course. Yeah.
3:20:25
I'm just copying from >> I think it's like a little bit of a newer. You're a revelation. >> This is a statement. >> Yeah. Yeah. >> Having fun. Uh well, that's great.
3:20:38
Congratulations on all the progress. Uh very excited.
3:20:40
I'm sure you'll be back on the show soon giving us plenty more updates.
3:20:44
And it's been fun because uh I mean I believe that we started tracking your journey via your viral joke post about uh doing the war or something.
3:20:51
But we've always had fun uh uh featuring great live in person live in person.
3:21:00
>> One one year ago today I remember just roughly one year ago I was sitting in a parking lot and I I was listening to it was the first time I ever heard of you guys.
3:21:07
You were reading like one of my tweets and it was just so surreal that like people from the internet are reading my tweets.
3:21:11
Like I think one of our customers sent it to us actually. You print it out. Yeah.
3:21:15
So, I called my mom today.
3:21:17
I was like telling her I was like, "Hey, I'm going to be on the I" I was like, "You're not you're not going to know what it is."
3:21:19
But remember that those guys that were talking about that tweet.
3:21:22
>> This was the whole This was the whole shtick was like little love letters to Silicon Valley folks.
3:21:25
Just like little messages of just, hey, we we found something that you did.
3:21:29
Fun cuz anyone can like, anyone can repost.
3:21:31
You know, it's it's easy to send a small thing.
3:21:35
It's very hard to actually print it out, sit down, talk about it.
3:21:37
Uh but we appreciate your post and we appreciate you coming on the show and hanging out today. So, thanks so much.
3:21:42
We're gonna close out the show and we'll talk to you in just a second.
3:21:46
Um, >> while he's walking off, let me tell you about get bezel.
3:21:50
com shop over 25 26,500 luxury watch authenticated in-house by bezel's team of experts.
3:21:59
I also need to tell you about eightleep. com.
3:22:01
Exceptional sleep without exception.
3:22:02
Fall asleep faster, sleep deeper, wake up energized. I had a rough night.
3:22:07
Kids have been all over the place, but I still got an 82. >> John 98 >> 98 98. >> That is remarkable.
3:22:16
Um well, is there >> are a bunch of Yeah, we'll see if >> you want to go through some breaking news.
3:22:23
>> Um Buco Capital Bloke is on the >> timeline.
3:22:27
You can feel the panic behind the urgency and intensity with which people are defending Nvidia.
3:22:31
It feels visceral and quite intense.
3:22:33
You can tell how much is riding on this.
3:22:34
Uh it makes a lot of sense.
3:22:37
Um what what else did you want to >> uh I thought it was notable.
3:22:41
Pager duty has fallen to a $ 1.
3:22:42
1 billion market cap at 500 million of ARR.
3:22:45
So trading they're not growing anymore. They're trading at 2. 1x ARR.
3:22:53
>> It's profitable >> uh according to Jason Lumpkin over at Saster.
3:22:57
Um so yeah, rough time out there if you're not growing regardless of the revenue scale.
3:23:02
Uh, two days ago we we shared that Enron back November 29th, 2001.
3:23:09
Nvidia replaced Enron in the S&P 500.
3:23:14
I saw this post go out from our incredible team and I immediately Googled to fact back fact back fact back fact back fact back fact back fact back fact back fact back fact back fact >> I was like there's no way someone has made a terrible mistake on our team and we are doing fake news unironically now.
3:23:27
We used to have some fun >> but apparently this is real. >> It's real. >> It's real.
3:23:32
was like, I'll take that spot. >> November 29th.
3:23:33
Obviously, uh that's not how it works.
3:23:36
It is uh it is much more mathematical than that.
3:23:38
I believe Standard and Pores picks the largest companies and uh after certain es and flows of the market, uh they swap folks in and out.
3:23:47
Uh but uh this went uh pretty viral, 5,000 likes.
3:23:51
Uh but what is really interesting is of course the the the Nvidia Enrod uh like comparisons are just so silly to me.
3:24:00
Obviously it's like you know the discussion is like is like will it go from being the best business in the entire history of the world to being like you know somewhat competitive and have to deal with like minor competition from other people.
3:24:10
It does not seem like it's some ridiculous Enron situation that's like so so insane.
3:24:18
Uh people are just having fun with that headline.
3:24:20
But what is incredible is this this branded shirt he's wearing. Look at this thing. >> Fantastic. >> So awesome. I love it.
3:24:29
>> Not enough people trying to go snipe vintage Nvidia merch. >> It's a great shirt. It's a great look.
3:24:36
And I feel like it's got to make a comeback. The button-down.
3:24:37
This is the pre Silicon Valley.
3:24:39
I'm just in a t-shirt era, but it's post suits, you know?
3:24:44
It's like we're not suits.
3:24:44
We're working in technology.
3:24:46
We're still going to throw on a collar, but we're going to dress it down a little bit. No tie, >> guys. Scroll up.
3:24:51
Scroll up on this for a second. >> Yeah. >> Oh, keep going. Keep going.
3:24:54
Oh, who's not who's not following Tyler?
3:24:57
You got to follow the account.
3:25:00
>> No, this is not my account.
3:25:01
>> I think this is a this is more of like a burner account situation.
3:25:04
>> Oh, it's a it's a scraper that we use to for it is it is >> got you got to correct that, Tyler. Come on.
3:25:10
Um, >> uh, uh, Gorkam over at Fall had an absolute banger.
3:25:15
Uh this was a chart showing ASML sells fewer than 500 units per year and generates 37 billion in revenue.
3:25:22
Is there any company in the world with a wider moat?
3:25:23
Uh and Gorkham says series A pitch meeting.
3:25:27
Sorry to cut you off, but what happened in December 2024 since there's like a slight dip in the chart. >> Yeah. What what did happen?
3:25:36
Why did their revenue drop in 2024? I I actually don't know.
3:25:43
Or is it just so much pull forward from 2023 or something? Um, I don't know.
3:25:49
>> Maybe they they uh were developing some hubris.
3:25:52
They decided to get complacent. >> Yes.
3:25:54
I mean, I certainly understand the the the the concept. Okay.
3:25:56
According to the CEO, customers in Taiwan had delays and weren't ready to take delivery yet, and orders got pushed back at the same time.
3:26:02
China raised to get as many machines as possible before export controls tightened. Okay, that makes sense.
3:26:07
Uh Sash Zatz says uh Oxford dictionary didn't get the memo.
3:26:13
Apparently ragebait named word of the year.
3:26:18
>> I think it I think it I >> noate that it would be the word of the year.
3:26:22
But it is so funny that you you posted this and then >> and then Oxford dictionary. >> Yeah. So this is true.
3:26:28
Rage according to the BBC rage bait named Oxford word of the year 2025.
3:26:32
It certainly feels that way on the timeline.
3:26:35
post 1 million views on this. 3. 6,000 likes.
3:26:38
People really This really set the agenda for a little bit. Wow. Congratulations.
3:26:44
What What a banger essay.
3:26:46
Uh should TVPN do a word of the year? I like that. Or uh motion. >> Mot motion.
3:26:55
>> Motion might be our word of the year. >> Word of the year.
3:26:58
>> Motion's a pretty good word of the year.
3:27:01
Motion named named Word of the Year 2025 by TBPF.
3:27:04
If you have it, you'll know. >> You'll know. We'll call you. >> Tyler has motion.
3:27:10
>> In other in other breaking news, uh uh Keith Reo is taking shots at Airwall X.
3:27:16
Uh Airwall X is now on the other side of a billion dollars in ARR.
3:27:18
What I love about this chart is that uh isn't that we hit a big >> founder Jack.
3:27:23
It's how fast the business is accelerating.
3:27:26
It took more than six years to a hundred million AR.
3:27:28
What does AirWall do exactly?
3:27:29
Can we >> I think they provide uh payment rails for a bunch of American fintexs who handle international. >> Okay. Okay.
3:27:38
And uh and so Keith Ra boy been on the show multiple times says cool growth chart.
3:27:44
Have you disclosed to US customers like Ripling, Bill.
3:27:46
com, uh Bra Non that you're quietly sending their customers to data to China.
3:27:53
Airwalk has become a Chinese backdoor into sensitive American data like uh from AI labs and defense contractors.
3:28:00
You must already know this, but your China based ops infrastructure and investors create legal obligations to assist with CCP espionage upon request.
3:28:09
Through airwalls, Beijing can assess supplier payments for AI labs so they could know who's who's using what models.
3:28:16
Uh payroll data for defense contractors, uh personal data for employees abroad.
3:28:22
Um that's obviously not good.
3:28:24
Obviously, many companies do business in China and that's not inherently a bad thing.
3:28:27
But your company has become a guaranteed vector for data transfer to the Chinese government.
3:28:31
And that's a different thing entirely.
3:28:34
You have multiple points of vulnerability, people, legal structure, cap table. Uh what's happening?
3:28:40
Uh you route global payments for US companies and critical sectors without disclosing that you're under a Chinese jurisdiction.
3:28:45
You moved your HQ to Singapore.
3:28:47
Well, that seems like a step in the right direction, maybe.
3:28:50
Uh but your largest operational footprint is in China. Okay. No. So good.
3:28:54
Maybe one step back and one step back >> and hundreds of your engineers in mainland China touch production payment systems.
3:29:02
You are subject to Chinese law that requires airwall employees to support CCP intelligence request and quietly hand over data when asked.
3:29:07
You hid this from your customers, but you are well aware of your obligations to China and that's why you insist on protection of Chinese data access to your contract.
3:29:19
Thanks to you, the Chinese government now has direct, covert, legally enforcable access to sensitive financial information. Uh, this is a big story.
3:29:26
This is a this is a crazy scoop from Keith Roui.
3:29:28
Um, and uh, I I will be interested to see where this goes, how how how quickly they can uh they can um, you know, remedy this.
3:29:37
Th this this popped up uh, a couple years ago with during the Clubhouse era.
3:29:44
the Clubhouse backend, I believe, was was at one point, you know, was working with a Chinese company.
3:29:53
Or maybe maybe it was that there was a company that did uh like peer-to-peer peer-to-peer audio streaming that was based in China.
3:30:00
And so if you were building a competitor, you might use that company.
3:30:03
And so >> I I became familiar with Airwall through uh the 20VC episode that that Harry did with uh Jack, the founder. >> Is it ripping?
3:30:13
is I mean it seems like the business is doing really well. >> Yeah. Yeah. Yeah. Yeah.
3:30:19
>> Uh >> anyways, what else?
3:30:21
>> I'm sure we'll hear more about it to say.
3:30:23
>> Um we got to get on uh with uh with Menllo Park. >> Okay.
3:30:27
Well, thank you so much for listening, >> hanging out with us today, >> tuning in.
3:30:30
We will see you tomorrow.
3:30:31
Please leave us five stars on podcast and Spotify.
3:30:34
>> The break uh the Thanksgiving uh break was absolutely brutal for us.
3:30:36
I will say every single day, >> but hopefully you had a great Thanksgiving.
3:30:41
>> Wake up and just twiddle my thumbs.
3:30:44
wishing we are podcasting. It's great to be back.
3:30:46
Hope you had an amazing break uh or or a little holiday and we will see you tomorrow. >> See you tomorrow. >> Cheers. >> Goodbye.