0:00
Right.
Search this video
0:14
>> There's misinformation clearing order inbound. >> Let's just roll.
0:29
We are surrounded by journalists. Hold your position. Come, get up. Trust the experts here.
0:49
We are actually founder mode. Five good.
1:03
>> I see multiple journalists on the horizon. Stand by. >> UAV online. >> Blaze. Double blaze. Triple blades. Double kill. Fight. Wrong. Cook is team deathmatch. We are experts. Triple blades. That's just wrong. Right.
2:27
>> Marky clearing order inbound. >> Get out.
2:41
We are surrounded by journalists. Hold to position. strike one. >> Strike two.
3:03
>> Activate golden retriever mode.
3:18
Marky clearing order inbound. Five.
3:39
>> I see multiple journalists on the horizon in Denmark. founder. >> You're watching TVPN.
3:55
Today is Wednesday, June 10th, 2026.
3:57
We are live from the TVP Ultradome, the temple of technology, the fortress of finance, the capital of capital.
4:03
Let me tell you about ramp. com. Baby, time is money. Save both. Easy.
4:08
Use corporate cards, bill pay, accounting, and a whole lot more. ramp. Ramp. You go to ramp.
4:16
>> You want some accounting help?
4:16
They're going to say yes and thank you for using us. >> And what else?
4:21
>> Um, two major stories.
4:21
Bunch of major stories, but the big one, Facebook got a movie, another movie, actually a third movie if you count the documentary.
4:31
There's been the social network, the social dilemma, and now the social reckoning.
4:35
Only two of those are uh are movies.
4:38
The social dilemma is a documentary, but uh it's very interesting and the uh anthropic got a fable and there's been a ton of reaction.
4:47
Ben Thompson was writing about it uh to the roll out there.
4:49
The Wall Street Journal uh like the announcement should have been like the really good model launches and people are amazed with what it can do.
4:58
But they they put the title anthropic puts curbs on AI models.
5:03
So the well I'm sure you've seen this on the timeline but we'll take you through some of the some of the uh some of the debates over how restricted the model is in certain areas. Bio is a big one. Cyber is another one.
5:15
There's a bunch of funny examples.
5:17
There's a bunch of reasonable arguments for doing this type of stuff.
5:21
So we'll go through all of it.
5:22
But >> can we talk about who's joining the show today? >> Absolutely.
5:25
We got a bunch of great guests.
5:28
>> We got >> What a lineup. >> What a lineup.
5:29
Hey, clicky founder Farza Majjid is joining.
5:32
You've probably seen his viral demo last week.
5:34
We're very excited to talk to him about computer use uh potential consumer AI applications, what he's doing to build hands-free AI voice assistants.
5:43
Uh the CEO of Snowflake's coming on.
5:46
CEO of Cloudflare is coming on.
5:48
Venode Kosla from Kosla Ventures is coming on.
5:50
And then uh Brett Taylor.
5:53
Uh Brett Taylor didn't even make the cut off on today's lineup. What's going on here? >> Brutal.
5:59
And of course, we have Marky Wagner.
6:01
>> Marky Wagner >> coming out announcing Emerging from Stealth with her new company. >> She's fantastic.
6:07
>> Uh so I want to kick it off with the uh the Social Network sequel.
6:09
I want to watch the trailer for The Social Reckoning because Jeremy Strong is playing Mark Zuckerberg.
6:18
Uh I've heard people say, "Hey, he looks he looks like Zuck."
6:23
Like the styling is there.
6:25
>> Henry says, "No Doug from semi analysis."
6:27
Oh, we got to get him on.
6:29
>> We should We should get Doug on. >> Let's hit him up. Um, >> hit up Doug.
6:32
>> Let's get him on >> and see if he's available.
6:34
>> Um, so, um, >> not that we have any time, but we'll make >> Well, we'll make time for Doug anytime.
6:39
Um, so, uh, the one critique is Jeremy Strong looks a little bit older than Mark Zuckerberg did in 2011 because he is >> calling him or it just it just it just doesn't it it I think it'll color the movie a little bit in the sense that the uh the film is telling the story of this like very difficult whistleblower situation.
7:00
And uh and it's different when it looks like a you know an a mature adult versus someone who was young and running the running into these problems for the first time.
7:11
Uh anyway, >> no Matthew Prince. >> What? >> He is joining. >> Yeah, he's joining.
7:15
We Yeah, we got the Cloudflare >> 1230 Pacific 12:30. Get ready.
7:17
We're going to go through Cloudflare.
7:19
They have a exciting news.
7:21
I think it's public already, but I don't want to blow up if it's going to be announced later today.
7:26
Uh but uh anyway, I want to uh I want to re react to the uh the trailer.
7:33
So, let's pull up the trailer.
7:34
>> Listen, before I go on, I want to make something clear.
7:37
>> This is how it starts. >> Okay.
7:38
>> I have a hunch you're not a fan of Facebook, >> but I am.
7:41
I am here to help Facebook, not hurt it. Okay.
7:46
>> You send me a message.
7:46
What would you like to talk about?
7:53
>> The chairman gave a session to order.
7:57
You'll read your opening statement, which we'll skip past for now.
7:58
That's a separate session.
8:00
And we'll move to witness questioning.
8:06
>> Spell your name and state your current occupation for the record.
8:08
>> So, is it not this time? >> C E R P E R G. >> Let's see. >> And your occupation? >> Zuckerberg.
8:16
>> I'm a professional defender.
8:17
>> Co-produced and directed by groups where 30% of content hits multiple risk factors. >> Hang on.
8:22
I don't piece to the social. Are you a tech reporter?
8:26
>> Is >> this my guys are counting on the next round of professional testimony to make you mark?
8:32
>> I'm happy to lend a hand, but I think you do.
8:34
>> You wrote broken code.
8:34
He was a journalist at the Wall Street Journal.
8:37
He is a journalist at the Wall Street Journal and broken a number of global news stories in our lives including >> the Facebook files.
8:44
>> The fire hose of bad information you are injecting into the air supply jet power.
8:49
>> I'm a free speech absolutist.
8:49
I'm not the one who's lying and I'm not stopping them from seeing someone who is.
8:52
It's >> not a bad Zuck impression.
8:54
He's got the voice tone pretty down.
8:57
>> Got worse as a result of time spent on the platform.
9:00
>> Senior leadership knows and is doing nothing.
9:03
>> I know there are easier enemies to make.
9:05
The mafia would be an easier enemy to make.
9:08
>> So what would you need >> to stand up the story? The internal documents.
9:11
>> This is a material violation of my NDA.
9:14
>> We're twice as big as the biggest country on earth.
9:15
We're not frightened of Congress.
9:16
We're post government around here. Please.
9:18
twice as big as the biggest country on earth.
9:22
>> We have 102 hours to get >> from a from a user standpoint. >> Oh, okay. Use it.
9:26
>> I don't want to be made an example of by a guy with unlimited resources.
9:30
>> H, I promise you, is imminent. >> Enough.
9:33
>> People around here understand that when I say no, that's the end of the debate.
9:38
I'm not 2 years out of a dorm room anymore. Charlie, look around.
9:43
Man, if they were trying to go as inspirational as the social network, they really missed the mark there. What happened? >> Yeah.
9:49
So that their intention was to try to fire up the next generation of entrepreneurs that to seems like they missed the mark inspire them to inspire them to uplift them.
9:59
>> You can build any business out of your dorm room. >> Yep.
10:02
And also it feels like if they if the goal of this film is to really really take the viewer inside uh TBD and meta super intelligence and what's going on with Nat Friedman, Daniel Gross, Alex Wang, like it feels like they just >> dropped the ball potentially. It's sort of crazy. >> Yeah.
10:22
Like I didn't see like anyone that even resembled Yan Lun.
10:25
>> That is crazy to not to not highlight that and what's going on over there. The launch of Llama.
10:30
The llama four moment that could have been the crescendo.
10:36
Behemoth and all the drama around that. You want drama.
10:38
You want Hollywood drama. You want Oscar bait. Llama for Behemoth.
10:43
That story is going to get put butts in seats. >> That's right.
10:46
>> Uh no, more seriously, it's hard for me to it's going to be really hard to not look at Jeremy Strong and and uh >> OC Kendall.
10:54
Yeah, cuz you're a succession guy.
10:57
Um, so, uh, this is the sequel to The Social Network and instead of chronicling the birth of Facebook, it's the story based on the 2021 Facebook leak by whistleblower Francis Hogan, it's going to be dramatic.
11:07
It's not going to make people like Facebook more, and it's probably going to make Americans even more distrustful of tech.
11:14
What's interesting here is that if you stop a random tech person on the street and ask them like, "What is the social reckoning about?"
11:21
They might say Cambridge Analytica because that was another drama moment.
11:24
And I I was sort of like, is it Cambridge Analytica?
11:27
Is the Is it the Hogan thing?
11:28
Um, and a lot of people don't even remember what the whistleblower story was. Little refresher there.
11:36
It was an internal document leak.
11:38
The Facebook files um showed that it that Facebook internal employees were aware of harmful societal effects from its platforms, yet persisted in prioritizing profit over addressing these harms.
11:50
Now, Nikita Beer chimed in, who now works at X, a rival platform to threads, and said, "Zuck makes a lot of mistakes, but this isn't one of them.
11:59
Meta literally had multiple teams of of $1 million per year engineers working on teen mental health, and they had the agency to override big product decisions.
12:09
They're probably a thorn in Nikita's side because he's trying to maximize click-through rates, maximize attention, and he's getting push back from, hey, there's guardrails here."
12:18
and he says, uh, the story this movie is act is about is actually a product manager who didn't get a promotion.
12:24
And so, um, that's the that's the push back.
12:27
>> The only thing the only thing >> was I >> and and uh the only thing with that argument is like you could make the same argument is like the cigarette company has milliondoll like doctors and researchers focused on making sure cigarettes are as healthy as possible, right? >> Yeah. Yeah. Yeah.
12:44
and and especially with the new fertility data, I think that there's there's a new cycle brewing of like exactly how bad is social media.
12:53
I still don't really buy the addiction thing just because I I'm familiar with like addiction defined in nicotine, which is like extremely addictive.
13:00
It's chemical and like you have cravings.
13:04
Like if you forget your phone, like do you have the same cravings?
13:06
It's it's it's sort of different.
13:07
And uh but there ser but there is like a very serious uh discussion going on around social media and its effects and this fits right.
13:17
>> The phone addiction thing is is like very real is I was at I was at an event last night and they required everybody to put their phones in these sort of like locked >> bags so you could go outside at any point sick.
13:30
>> Uh I didn't feel sick but I noticed probably like 20 times throughout the night >> Yeah.
13:35
>> I was thinking like I was thinking of going for my phone. Yeah.
13:39
>> I didn't start having like physical withdrawal symptoms. >> Yeah.
13:41
But the craving was there. >> Yeah.
13:43
I just kept thinking about it. It was like phone.
13:45
It was like phone noise, right? Phone noise. >> Oh, this is real. This is real.
13:49
>> I was having some phone noise. >> Yeah. Yeah. Yeah. Yeah. Okay.
13:50
So, yeah, that's legitimate.
13:52
Um, so the question that I had really quickly is like what is the impact on Meta?
13:56
Like uh we can pull up the stock chart.
13:59
It's a one and a half trillion dollar company.
14:00
They're trying to raise more equity for the AI buildout, hire people. Like is this good, bad?
14:07
What does it look like over the short, medium, long term?
14:08
So short term, I think this was amazing timing.
14:12
They really got lucky because uh as we will talk about like uh the fact that this trailer dropped the same time as Anthropics Fable 5 was incredibly fortuitous because Fable just took over the timeline.
14:24
Um and no one was really talking about this in tech.
14:27
Tech insiders won't really see or talk about the social reckoning this week.
14:30
Uh case in point, I put it second in the newsletter when I wrote it up.
14:33
Um, medium-term, I think Jeremy Strong is going to drag Mark into like the bad group of AI leaders while he's on his Oscar tour.
14:41
Like, there was a version of of of Facebook strategy, a meta strategy that's just like, hey, we're like Amazon, we're like Microsoft.
14:49
Like, yeah, we have bets in AI, but we're not like we're not deciding the frontier.
14:55
Obviously Mark has made the decision to hire some of the greatest researchers, invest very aggressively and this puts him sort of at the center of the conversation around well what will the effect of AI be on our society on kids?
15:07
Should kids be talking to LLM's psychosis?
15:09
all these different things will come up just as we saw 40 psychosis and then clogged code psychosis a little bit pop up and people are talking about that like you're you're you're priming the pumps for the next wave of like oh a bunch of people went into the meta meta app and like they had a bad time and now we got to talk about that.
15:25
So, um I wouldn't be surprised to see >> what do you think, uh Mark is going to be more annoyed about this week, the fact that SpaceX >> is going to be valued at meaningfully higher >> Yeah.
15:38
>> than Meta or this trailer?
15:41
>> It's actually a tough one. >> I don't know.
15:44
I mean, I think the real one is the is uh is Fable 5 because like if if if uh if Meta was using a ton of enthrophic models and then the new model comes out and says you specifically can't use it in MSL, the most important initiative in the company like that's like a pretty rough thing where you're like, oh, I've been working with this company to develop my models for my family of apps and now I can't use them.
16:09
They do, you know, you know, Meta is spending billions of dollars a year with Anthropic.
16:14
They have been for a while. >> Um, >> are Yeah. Yeah.
16:20
>> They will spend billions of dollars this year.
16:21
Um, and the question is like if Anthropic does allow Meta to use their models for Meta's AI research, what does it say about what Anthropic thinks about Meta's potential in AI?
16:34
>> Secretly love ads this whole time. I don't know. >> No.
16:38
To me, to me it it would it would be them not not feeling like they had a >> ability to get to the French. >> Totally, totally. Totally.
16:46
And then and that's an ace that's up their sleeve at any time.
16:49
They can always just take the guardrails off and be like, "Hey, more more more AI, more intelligence on demand for everyone."
16:54
Um, so, uh, I think the fallout of the medium-term, Jeremy Strong's going to go on this press tour, uh, for the Oscar and he's going to have a bunch of like really emotional, pathy sound bites and then also be, uh, like just viral because it's funny to see him doing this impression, and he's very good at like drawing.
17:10
He's already gone viral for his Mark Zuckerberg impression in the past.
17:14
Uh so I wouldn't be surprised to see the next version of a Bernie Sanders press release about AI uh highlight Daario, Sam and Zuck instead of what has historically been the order which is Sam, Daario and Demis.
17:26
So it's always based on like perceived industry power and negativity around personal brand.
17:31
So they have to have a scary quote and then they also have to have uh a lot of power in the industry.
17:37
in the industry. Zuck's power in the industry is rising and also with this movie like there's more ways to sh to take shots at him like oh look at what he did with the Facebook files remind yourself of that because a lot of people don't really remember the details they're going to after they see this movie then you can uh then you can say
17:54
well this guy's also building crazy data centers and look at how he handled this if they if he handles the AI thing like that it's going to be bad right so that so so that's like a little bit of a risk in the medium term long term I don't think the social reckoning is going to matter all that much for meta people will complain about meta data centers AI plans. >> Why did they not get Jeremy Strong like
18:10
>> Why did they not get Jeremy Strong like a wig? >> They did.
18:14
>> Oh, they they because also he is playing an older remember when he had the Caesar haircut. Yeah. So, it's different time.
18:21
So, people will complain about all this stuff, but they'll do it on Meta's family of apps.
18:24
I don't think they'll be churn.
18:25
Uh and certainly advertisers won't pull out.
18:28
won't pull out. it's impossible to pull out of meta because uh it just drives it just drives up the RO the ROAS the the return on ad spend for smaller companies and a bunch of small businesses will just jump in and say oh great like you know some big company pulled out and is boycotting Meta that's great I'm Ridge Wallet's going to jump in or somebody else is going to jump in um >> yeah people have advertisers have tried to boycott it
18:51
>> yeah you can't on Facebook it's impossible so the business is just too strong um and so and I also don't think Mark will be singled out by a regulatory hammer should it come down hard like if there's a data center ban I don't think
19:01
it's going to be uniquely focused on meta um only Daro or only Dario and Sam have like true scapegoat risk because they run pure play labs and have been so noisy about AI and they have the biggest revenues in the category uh there might
19:14
be a there might be data center regulation but it won't unfairly target meta so anyway that's my social reckoning uh take let's move on to Fable which launched yesterday and the model seems like incredibly impressive I I've
19:27
been seeing these vibecoded games that look really really good and for some of those I haven't seen those those like like the vibecoded games hit really well on social media because you take a video of them and you show the example of what
19:41
it built you share the time and people just sort of believe it it could be embellished but in general these feel like pretty solid of course like making a great game is a great mechanic and there's needs to be multiple levels and
19:54
like a single demo of like a forest putting in is not quite there, but really really useful and I'm sure we'll see a bunch more examples of like games as memes, simulators, and these things are going to get easier to build. That's That's really exciting.
20:08
Of course, there was a bunch of debate on the timeline yesterday uh and it's bleeding into today around the latest model, Fable 5.
20:16
Uh the first mythos class model uh that both seems remarkably good at long horizon tasks like software development and knowledge work, but rejects requests related to biology, cyber security, and frontier LLM development.
20:25
Interestingly, I haven't seen anyone share uh anyone share rejections around anything else.
20:34
Like, did did they remember to reject like build me a nuke?
20:37
Like, because I haven't seen anyone try that.
20:39
And it'd be very funny if it was like, oh yeah, just we we we didn't get around to that.
20:44
Or there are so many other things, but I think a lot of those other other uh like things that you should reject, queries that you should reject have been ironed out in previous uh in previous iterations.
20:54
So going back, you know, the say a slur was a big one for a while or say something rude or different pausing pausing a chat basically just like shutting down a conversation versus like switching you to a less performant model. >> Yeah.
21:11
And it creates this like screenshot that went viral pretty much continuously yesterday.
21:14
Um and so this aligns with anthropics focus on safety, but as many people have pointed out, it's also just good business.
21:19
uh you don't want competitors using your products to create directly create competitors and you also don't want financial liability or negative headlines from bad actors using your models to for nefarious purposes.
21:29
Ben Thompson called it true alignment.
21:32
Uh the take safety seriously culture aligns with business value creation which is very very rare.
21:38
Um, oftentimes the like be good or your culture uh limits what you can do and actually hurts your business, but it's something that you do in favor of brand.
21:48
Like Apple would probably be would probably be more profitable if they were using like diesel generators for all of their for all of their uh data centers.
21:55
They went clean energy because they wanted to have an environmental brand and over the long term it's helped them, but in the short term uh it's been rough.
22:03
Um, of course inexpensive.
22:05
So Ben Thompson writes, "What is so fascinating about Anthropic, however, is that while I'm sure some executives of the company are thinking this way, I also totally believe that the employee base broadly also happen to believe that they are doing the right thing."
22:18
It's fascinating to observe.
22:18
Me, the rational business analyst, sees a hard-nosed but understandable decision to cut off wouldbe competitors.
22:26
Anthropic employees and advocates, the true believers, see a regrettable but understandable safety decision that ensures that responsible and thoughtful people themselves of course will be the ones guiding our AI agi future.
22:36
This is true alignment and it's an incredible accomplishment.
22:41
Um, Facebook has tussled with this a bunch.
22:43
Uh, and and we already talked about that.
22:45
But, uh, to be clear, uh, the Fable 5 rejection threshold really does feel way too low from what people are saying.
22:51
Uh, tons of examples on the timeline of a biologist just saying hi to the model and getting kicked down to opus.
22:57
I saw you shared someone just said cyber with the devil horns, the purple devil horn emoji, and it's like we can't we're not going further.
23:06
Um, of course, like, uh, bringing down like a broadhammer, you can always dial it back over time, but every rejection is this implicit invitation to hop on the phone with an anthropic sales rep and get on the mythos enterprise plan.
23:18
And that's where the real dollars are, too.
23:20
The timeline is unhappy because the idea of democratizing science, technology, all of this is very alluring, but the pool of dollars available from all the biohackers in the world probably isn't close to the budgets available from big pharma.
23:35
And so I you you again you're in this like rational business analyst uh situation and uh you and and you fail to see how this uh is that damaging.
23:43
Um except to like the hacker community.
23:47
Uh the real tricky part is how AI a frontier AI research is handled.
23:51
Instead of outright rejecting the query and bumping the user down to opus, the model appears to answer but quietly gives a degraded answer.
23:58
And this was disclosed in the model card which is interesting.
24:02
So this is again reasonable >> business not disclosed >> to the user in the product while they're paying for which is a different which is a different path than bio and cyber.
24:09
So if you if you go LLM frontier research devil horns it will actually give you an answer apparently.
24:18
It won't bump you down immediately.
24:20
Um but so it doesn't disclose that which is odd.
24:22
Uh outright rejecting requests for AI research and just saying hey user sorry this model doesn't work for that type of project.
24:30
please share, please use another model or contact sales if you want help with this would have been much more in line with the bio and security in cyber security strategies.
24:36
Uh and they and then they also could not have disclosed and it's also possible that they just didn't need to disclose this at all.
24:45
Like they could have just released a model that was intentionally nerfed on AI research.
24:48
It would have shown up in the benchmarks because people would have benchmarked it on some sort of AI research bench and been like, "Oh, weird.
24:55
It's really good at all these other things, but it's bad at LLM research."
24:58
And um and maybe that would have been a bit of a brand hit maybe, but uh users might never know that the model was intentionally degraded around this category of work.
25:07
Uh so that leaves this third more worrying position.
25:11
Intentional degradation without disclosure in the model card.
25:13
Uh there's no evidence of this, but it's possible that other workflows might be nerfed and there's no law or even convention around disclosure.
25:20
Again, maybe good business, but a weird situation to be in.
25:22
So probably bullish for eval if you're building a business on top of a big lab.
25:27
You can imagine like a legal AI company will want to be really sure that the models they're using aren't degrading unexpectedly and not telling them.
25:33
It's different if you're like, "Hey, I've been a bio researcher for a while.
25:36
I'm using this and I know that this model was never intended for me or it or I've been using it but what you don't want is like I'm using it and now it's giving now it's leading me astray in my work and it's also not telling me that it's going to lead me astray which would be like sort of an odd outcome that I I I think they'll probably address in the near future anyway.
26:00
>> Yeah, that aspect uh triggered Dean Ball.
26:05
said, "My last observation reanthropic secret sabotage safety policy is that it undermines actually good safety policy." How?
26:13
First, it is very plausible to describe this as anti-competitive behavior.
26:17
Uh even if you are maximally sympathetic to anthropic here, you must admit this.
26:22
And it is behavior being justified in the name of AI safety.
26:25
If you believe, as I and many anthropic staff do, that it may end up being critically important to relax antitrust enforcement so that the frontier labs can cooperate and collaborate on some areas of AI safety.
26:38
Anthropic just undermined the case for that in a large way.
26:40
Overall, this is ma this massively and profoundly raises the status of the argument that AI safety has been hyped to justify monopolistic behavior by labs.
26:49
I continue to believe that AI safety is a real and serious issue that is growing in importance rather than diminishing.
26:54
if you agree with me, this incident is a setback, maybe a serious one.
26:58
And third, he says, as I have observed elsewhere, Anthropic's official corporate policy is structurally identical to the fact pattern alleged against them by the Department of War.
27:08
I still think DO acted both falsely and wrongly in that fight, but it is no longer possible to defend Anthropic with a full throat after this incident.
27:17
This raises the case for heavier-handed regulations.
27:21
Xanthropic is making awfully good case here that their product ought to be treated as utilities and thus their alignment practices should be a matter >> carrier thing >> public policy rather than private property.
27:31
I am starkly opposed to this sort of state power grab but anthropic is doing more to justify it than anyone else.
27:37
Thus, significant damage has been done to a community and entire approach to AI governance.
27:40
It was done unilaterally by anthropic likely motivated largely by self-interest and justified within the internal psychology of the firm through the lens of safety.
27:51
I suspect this is fixable in the economic and legal senses philanthropic but I fear that trust has just been broken and the goodwill extinguished will take very much time to repair.
27:58
And just to level, he wrote the AI action plan, but then also came out very publicly in support of anthropic during the conflict with the department of war, saying that the department of war is completely overstepping by uh pushing towards uh uh supply chain risk designation, putting pressure on anthropic for not wanting to work with the government in that particular way.
28:23
Um, obviously there's a whole bunch of new data that's been released and stuff and that uh that conversation has evolved, but uh he's not uh he's he I he doesn't strike me as some like crazy hater.
28:34
Uh anyway, um let me tell you about Railway.
28:38
Railway is the all-in-one intelligent cloud provider.
28:40
Use your favorite agent to deploy web apps, servers, databases, and more while Railway automatically takes care of scaling, monitoring, and security.
28:47
Uh Doug Olaflin was also uh sort of mixed results on fable.
28:53
He says when it works it's brilliant but the unilateral guard rails make me frustrated behi beyond all belief.
28:59
He has a folder of health information with like a 100 days of of aura health data.
29:04
This is a very logical thing that people would do with codeex or cloud code.
29:08
Gather up all your health data.
29:10
I actually talked to a very very prominent person in tech about how they got into vibe coding and and CLIs and one of the first things they did was reanalyzeed some Whoop data, reanalyzed some health data and he correctly detected sleep apnea and went and got it treated.
29:26
Like that is it's not the cure for cancer, but that's like incredibly awesome and like very very good and exactly what you want to see.
29:32
Who knows, maybe the thing that you detect early could lead to an increased risk of risk of cancer in the future and that would be a really really big win for our society.
29:44
And so, um, little bit tricky there.
29:46
He says there's a private investment in the life sciences tool space. Guess what? It's not safe.
29:49
Doing some code scanning for vulnerabilities, not safe.
29:53
I get the safety, but it feels incredibly out of touch for a group of a few hund a few thousand people making all making total comp in the millions, telling me what is and isn't safe.
30:01
Dario worried about inequality.
30:03
I think he has to realize that he is the inequality and the unilateral gatekeeping feels whack as hell. I don't like it.
30:12
>> People are people are very >> and like Doug Olaflin is extremely optimistic about about anthropic like he he he recognizes the power of the models in the business and has been uh very very strong supporter I I feel like a strong user.
30:25
strong user. uh he was one of the first people to admit to claude code psychosis uh and he's unhappy with the current situation but these things are very tunable I'm optimistic uh with more disclosures and more fine-tuning and uh
30:39
and a and a smoother smoother path uh I think we can get to the good outcome here >> last last one says tell me about mitochondria it's the powerhouse of the cell right >> chat pause the chat pause is rough yeah it's a It's a screenshot. >> But maybe maybe the harness is just
30:57
>> But maybe maybe the harness is just saying like sorry if you don't know that the mitochondria is a powerhouse of the cell like >> it's not worth my time.
31:06
>> I don't even want your >> I mean that's the thing with like the you keep going back and forth good. Yeah. >> Yeah.
31:12
I mean, there are expensive questions and all of these screenshots are clearly from uh like $100, $200 a month plans and uh I'm sure the GPUs are on fire as they always are with any new model launch from a Frontier company and uh it can make rational business sense and so when everything aligns, when you're like, "Yeah, we're maybe being a little bit too safe, but it's actually good for our business."
31:33
It's very easy to say yes to that stuff.
31:34
The big question is like where do you get tested where something there's a hard decision and it doesn't align with your business interest?
31:42
That's always tough, right? >> What what's going on?
31:44
Was there the the new data retention policies as well?
31:50
That also has to do with like does it trying to >> This one was funny because I assume that every tech company stores everything forever basically.
31:58
I didn't realize that they don't.
32:01
not not for these like enterprise use cases.
32:05
>> I know that the enterprises said that they wouldn't train on your data, but like when I go into any chat app, I expect to have my data from a year ago still there.
32:16
Like I want in fact one of my main critiques and frustrations with these apps is that when I go to them I want to be able to instead of hit the hitting the search bar just go into the chat box and say hey remember a couple months ago we were talking about uh Crowd Strike and I had you pull up some data.
32:34
Can you just go refresh that get the original thing?
32:36
And a lot of times these apps are like uh like I don't really have access to all your chats, but like the chats are clearly saved, but so I I understand people are upset about this, but it didn't fully clock for me.
32:48
Um why I I understand the training thing like the but >> yeah, and they're very they're very explicit.
32:53
They're not using the data to train. >> They're not. Okay.
32:55
Oh, well then that's good.
32:57
So the the steel man I believe was that uh you need to hold it for 30 days because companies probably international companies competitors will set up a ton of shell corporations and send out like pseudo random queries over at random times over random accounts through VPNs and they will triangulate something that is useful to distill or learn from the model.
33:23
And so by keeping it all, you can now run analyses and look at, wait, is there a pattern where where 25 different accounts that seem completely unrelated all seem to be triangulating the same question?
33:36
That seems reasonable to me.
33:39
Um, but stay out of my data. I'm built different.
33:43
I'm not distilling anything. I don't know.
33:45
Anyway, >> I think it's time.
33:46
>> Let me tell you about Shopify.
33:46
Shopify is the commerce platform that lets you grow with your business.
33:50
that grows with your business and lets you sell in seconds online, in store, on mobile, on social, on marketplaces, and now with AI agents.
33:57
And speaking of AI agents, we have a founder, a company.
33:59
I don't know if we should call it an AI agent. What are we calling it? FZA.
34:02
Is this Are you creating a new category? Introduce yourself.
34:07
Tell us what you're building.
34:09
>> What's going on, guys?
34:10
>> Thank you so much for >> Nana. Of course, of course.
34:11
Farza here working on Hey Clickie. Yeah. Uh, man, what is it?
34:15
Uh, it started as an AI teacher. Yeah.
34:18
And I guess now it's an AI that where you can essentially talk to your computer and it does whatever you want it to do from like >> it's personal super intelligence and you got there before the big man. >> I love it.
34:28
>> Somehow I feel like the big man doesn't even know what they got sometimes, you know, and it takes the little guy to figure it out. >> I love it. I love it.
34:34
I'm always rooting for the little guy.
34:35
So, uh, like like re reintroduce the product because you went viral last week, but it feels like you've been working on this for a while.
34:43
What's the nature of the team, the structure, the funding?
34:44
like what's actually building this?
34:46
What's adoption been like?
34:48
Like where is the product today?
34:50
>> Yeah, I started eight weeks ago.
34:50
I just wanted to build a uh an AI that could teach me >> that could teach me Da Vinci basically while I'm like actually using Da Vinci.
35:00
And I thought it was kind of >> This is Da Vinci Resolve, the color grading and video editing suite that's free from Blackmagic Design. Got it. >> Exactly. Yes.
35:07
So like I Da Vinci Resolve complicated program. Yep.
35:09
One hour YouTube videos weren't in it.
35:10
let me just build an AI that can help teach it to me while I could talk to my screen and I could learn.
35:15
>> I've been there and uh put it out and uh I thought it was really a really bad idea.
35:19
One of my friends just decided to post it.
35:22
>> So I posted it and it goes really big uh about eight weeks ago.
35:25
>> Uh after that I didn't keep working on it.
35:27
But then the cool thing was I saw all these people, you know, you know how when you put something new out there, you have all this emergent behavior kind of start happening.
35:33
So when you let a human being talk to their screen, what did they start doing?
35:37
Turns out they start doing a lot of crazy things all the way from watching anime with Clickie all the way down to like learning like Blender with Clickie.
35:44
Um because it's so easy to just talk to your screen. >> Yeah.
35:47
>> And that's kind of where it started. Yeah. About eight weeks ago. >> Okay.
35:49
So it started >> it was Yeah. Go ahead. Go ahead.
35:51
>> it was Yeah. Go ahead. Go ahead. I I I just want to hear about like the unhobling path because uh I imagine that there's demand for uh you know like these chat interfaces have existed for years where you open an app and you uh you know do the voice mode and then you
36:07
have it read it back to you and there's some that are more polished than others but uh when I think about using a computer I think rearranging windows I want anime over here and Da Vinci Resolve over here and the YouTube video over here and I want to be able to puppeteer the mouse and the keyboard have full context. So I need to be
36:21
So I need to be taking screenshots every minute, every second, 30 times a second, 144 hertz, like what are we doing there?
36:31
How are you getting the data in?
36:32
And then uh what's actually under the hood?
36:34
Because you need to process the voice actually produce an output and it's surprising to see a harness like break out of this so quickly. >> Yeah. Yeah.
36:45
It's actually uh so simple.
36:48
We use GBT realtime up front to give you like the first layer to give you like the really quick answer. Okay.
36:52
But then uh if you want like a deeper kind of like uh uh thought process over the image, you know, if you're in Dinci Resolve or if you're in some complicated software, we now use Fable 5 actually by default. Yeah.
37:03
So now Fable 5 is like absolutely mind-blowing in terms of how precise it is on screen understanding.
37:08
So we use that now by default when you want to like uh understand something on your screen.
37:12
But yeah, we only actually screenshot when you press a button. Okay. Press the button. >> One screenshot. But it's >> Yeah.
37:17
But it's really crazy because uh we can still detect the program you're on and stuff and that that's already really helpful. Sure.
37:21
Like if I if I know you're on notion for example Yeah.
37:25
and I you're there for like 10 minutes, I can just ask you and be like, "Hey, like >> what are you doing? Like can I help you?"
37:30
And I think it's that sort of new modality that's like uh been really exciting because that's something that's just not possible today with any chat interface. >> Yeah.
37:37
It's more like a like a co-orker walking up like like a co-orker just like being like present and like kind of like like um sometimes you don't need help until you're asked, right? Or you don't know.
37:49
>> I kind of describe it. >> Yeah. >> Yeah.
37:50
I kind of describe it as like having this 23-year-old intern with like a decent like like a new grad that's always watching over my shoulder and and it's pretty much like seeing patterns in what I'm doing and it's and then just tapping me and saying, "Can I do that for you?"
38:02
Um, if I had that, of course, that would be awesome, but uh it'd be cool if everybody had that.
38:08
>> A lot of these frontier models are expensive.
38:09
How are you thinking about the business model?
38:11
Because it it seems like you've built something useful.
38:16
People should just pay 10% on top of whatever the metered rate is.
38:19
But at the same time, in consumer proumer uh subscriptions, flat pricing is very popular.
38:27
But that in that introduces financial risk to you.
38:29
you have to understand your costs and how they change.
38:33
Uh what do you think the business model will be as you grow the business?
38:38
>> Yeah, I mean believe it or not like uh people are very open to paying large amounts of money like normal consumers like large amounts of money for uh for like more uh for having better access to these models. Sure.
38:50
>> Uh different interfaces.
38:50
So right now we just charge a straight up $20 a month. Okay.
38:53
Uh but even then there's a limit.
38:55
You know you get like for 20 bucks a month you get 150 agents on Clickie. Sure.
38:58
But even that's very specific because once you pass that number that's when we start losing money.
39:02
So everything's done in such a way where uh it's pretty cost effective.
39:05
Also like just so you guys know like just calling like sonnet or opus or fable it's kind of cheap overall.
39:11
Uh the expensive part is like agentic work that's really expensive.
39:16
Like just to give you an idea >> on GBD 5.
39:18
5 if you tell CEX right now to like uh go and click add to cart in Amazon that cost 25 cents on by just API. Yeah.
39:26
Um, and I know because I'm seeing the cost myself. Yeah.
39:28
And so there's a lot of tricks to kind of reduce that.
39:31
Um, but no, overall there's so many tricks to making agents cheaper, more efficient, uh, changing the model based on their quest.
39:38
>> And I didn't catch it.
39:38
Are did you haven't you have raised for this or you haven't? >> In the process. >> In the process. There we go. There we go. If you're out there. >> Yeah.
39:48
But but if you're in the process, that means like you didn't launch this and be like, "I'm down to be losing like $5,000 a day, you know, on, you know, like you you you've been focused on the economics probably earlier than you would have been if you were had raised money out the gates." >> I mean, yeah.
40:04
I mean, I'm I'm never the type of guy that wants to be losing thousands of dollars a day.
40:07
So, uh um we can stop it early, but no, I mean, that's that's the thing about AI.
40:12
that's that's the thing about AI. like you know if you I've built a lot of social apps in my life and for those costs technically don't matter based on the number of users coming in but in my case for every like 10 new people I add I'm kind of losing or losing money or making money depending on who what they're doing so I kind of gota kind of got to think about it so yeah it's kind
40:29
of AI kind of accidentally making everybody better at business uh because it's kind of required >> so is uh I'm I'm thinking about the hilarious uh the hilarious like like tradeoff in having an agent that applies a coupon code but winds up spending more money applying the coupon code than you save because it's like oh yeah we put in this it saved you 24 cents but the API call was 25 cents. Um but h how how do
40:48
Um but h how how do you how how important to you uh will like creating an ensemble of models that sort of maximize the parade of frontier be as you build this business?
41:01
be as you build this business? because I'm imagining that any day now we're going to get somewhat useful ondevice models on Mac from the new Siri models the ondevice models those will probably be very limited in what they can do but
41:20
useful for some things and that's free for you then you'll have open-source you know legacy models GPT4 class stuff and then you might only call out to a fable 5 on demand so do you need to build your own router or Is there a tool that you'll be using? Like how important will
41:35
Like how important will that be from the interactions that you're seeing?
41:41
>> That's a great question.
41:41
You're pretty much asking like what's the harness?
41:42
And uh we just built our own.
41:43
Uh because technically like this is all so early where there's nothing to use off the shelf. Yeah.
41:49
Uh you have to think deeply about it yourself and then build it yourself.
41:53
So, for example, let's just say, you know, you're on clickie right now and you're saying, "Hey, Clickie, uh, I want you to actually research, uh, the latest in the Iran war and put it in my in an ocean doc for me when you're done." Um, >> what now?
42:06
Like, should it hit Opus? Should it hit Codeex?
42:07
Should it hit like it's actually depends today and just route them ourselves.
42:12
So, yeah, we we use like >> we use like four models right now underneath the hood.
42:16
Um, >> and and do you have a model that's doing the routing?
42:19
So GBT real time actually does the routing. Oh, interesting.
42:24
>> Yeah, that's something that we that's something that we figured out that I think not even the OpenAI team really knew and they hit us up about which is it's a great router.
42:30
Um it's so good at like because it's so good at tool calling. Yeah.
42:34
Um that means it's actually really good at routing requests.
42:37
So for example, we have this tool called called Fable 5.
42:38
Uh so if a user asks something that like is a heavier pixel task, we call Fable 5.
42:43
if it's like more of like agentic work we call um uh we call GP 5.
42:47
call um uh we call GP 5.5 and so it's kind of all being done in the background but that's the kind of magic of the product right now where it's like most people who use our product have never used an agent before they don't even know what codec is and so for them it's magical where it's like oh I'm just talking to my computer and it's doing
43:02
the thing that's that's awesome that's what they want >> uh >> how do you how do you think about the tension between I think you being a generational media talent and then like startups because I actually feel like uh this is our first time meeting, but I've seen your videos over many years at this point and like you're just really really really really good at it. So much so
43:24
So much so that as like somebody who's in media, I'm like, "Well, I kind of want you to just do that podcast." >> And I don't say that.
43:33
Normally, normally it's exact opposite where it's like, "Hey, like you shouldn't do this podcast.
43:38
You should just work on your company, you know, or whatever whatever the thing is."
43:41
And and they obviously feed into each other really well.
43:44
Like you have an edge, you can get attention >> for free, you know, just investing a few hours making a video, whatever.
43:52
It's very, very powerful skill set.
43:53
But uh how do you think about that tension?
43:58
>> You know, it's funny like I've been making videos for like 15 plus years and uh I uh >> overnight success >> second time. >> Yeah.
44:06
I always just see this as a uh as like a a thing I get to do for fun on the side while I get to do the main thing which is like actually build stuff for people.
44:14
At the end of the day like I am painfully familiar with how bad of a business like making videos is and making movies is and making music is like I know how bad of a business is like firsthand.
44:24
And so I know but I know more than anybody the power of getting in front of a billion people every single month.
44:30
Um when there's a good engine underneath a good business engine underneath it um that you that that's that's powering something else.
44:36
And so uh I fully intend to do that.
44:38
So if I have this media ability I'm glad I got it.
44:41
Uh I think uh >> it's like sales. >> Yeah.
44:46
Sales really scalable sales.
44:48
>> You got to be good at sales.
44:48
Um WWC uh some movement on interoperability in iOS.
44:55
Still feels like the the the clicky iOS app is probably on the longer term and will be uh you know stuck in the walled garden for a while.
45:04
walled garden for a while. Uh but what I'm interested in is that do you see just a really really solid market because there's no Da Vinci Resolve iOS app that I'm aware of and certainly people don't there are there are like proumer sort of professional tools that really only exist on a desktop or a
45:24
laptop and so are you are you interested in like going and hacking on crazy workarounds to build something that like maybe makes people a little upset in Certino but gets you into that more casual mobile space or do you think that you just want to focus on desktop proumer uh you know uh desktop apps? >> Uh no I want it all. I I want I want >> Uh no I want it all.
45:47
I I want I want when you talk to your computer whether this is this is your computer or your computer is this computer.
45:55
>> Uh I want to be an interaction layer on top of it.
45:57
That being said though, I have no I have no interest in necessarily being like a OS level uh controller for example like uh >> kind of like what Apple is doing right now. No real interest.
46:07
Uh for us most of our customers and most of our users are essentially connecting like 15 integrations to clickie from G Suite to not to Dropbox and doing work with it.
46:16
So, I'm a lot more interested in like that reality where talk to what if I can talk to my phone in the morning and just say, "Hey, like look through all my emails and send me a uh send me a brief, you know, when you're done."
46:25
Um, and just I can just talk to my phone and do that.
46:28
I'm interested in that stuff, which is stuff I don't think Apple is going to do.
46:31
Personally, they're not going to make like connect your connect these 15 places to my Apple account and do work with it.
46:37
I just don't think they're going to do it.
46:39
>> How do you think the health of like open ecosystems in uh desktop software is broadly?
46:44
because there's two ways that you could like update an Excel file or there's actually probably a bunch, but uh you can you can you know puppeteer the mouse, move over, click, type the cell, save, you know, or you can just go edit the underlying CSV file in like raw text and then just refresh the front end.
47:03
And I imagine one's way cheaper for you.
47:06
So, you want to like lean into getting the MCP servers, the APIs, the file, the file writing, like the CLI, uh, you know, interaction down.
47:13
Uh, but do you think that there will be companies that lean into that or companies that fight that because they want you to stay, you know, mouse and keyboard at all times?
47:27
>> That's a good question.
47:27
I don't think so because at the end of the day, what Clickie is doing in the background, just so you guys know how it works, we literally took the Rust binary that's in codeex and we package it with our with our app. >> Okay.
47:39
>> So that when you actually call the clicking agent, you're just calling a subspawn of codeex. Yeah.
47:43
And so this is on purpose like I just want to use the best uh the best model and the best kind of thing possible. Yeah.
47:48
And so >> uh no, I don't think that's going to end up happening. In fact, it's better.
47:52
Uh for example, if you ask me clickie right now, >> Yeah.
47:56
Hey, hey, like how do I actually like uh add a formula to this Google sheet here?
48:00
Uh there's two answers there.
48:02
One answer is let me show you.
48:03
The second answer is I can show you, but do you want me to do that for you? Sure.
48:06
And I think that's where we're going to start going with computers where you're just going to start your computer's just going to say can I just do this for you? Like I see you doing it. Like I can just do this.
48:13
Um so that's where we're kind of >> Well, good luck with the fund raise.
48:16
I'm sure you'll be back on the show soon.
48:19
Let us know when it closes.
48:19
Incredibly excited for Yeah, I love watching you you win and and have fun in the process.
48:26
>> Yeah, it's great to meet you. >> Congratulations. We'll talk to you soon. >> Have a good one. >> Later.
48:31
>> Let me tell you about MongoDB.
48:31
What's the only thing faster than the AI market, your business on MongoDB? Don't just build AI.
48:37
Own the data platform that powers it.
48:40
And up next, we have Trent from Side Talk.
48:42
He's a creator and co-founder of Side Talk.
48:45
He's in the waiting room.
48:47
We'll bring him into the team in Ultradome.
48:48
He'll turn Nick's fandom into one of the internet's most recognizable sports media brands. >> How you doing? Take us through. How is New York? >> How is it going?
49:00
>> Yo, it's crazy out here. I'm not going to lie.
49:02
>> I was about to do this interview from my apartment and then I realized, hold on. >> Yeah.
49:06
>> Has anyone done a TVPN interview live from the streets of New York?
49:09
>> You got to go to the sidewalk.
49:09
The founder of Side Talking >> Amazing.
49:13
For those who don't know you, take us through your media empire.
49:16
take us through your strategy. Break it down for us. >> Okay.
49:22
So, I have a little company called Side Talk.
49:24
You might see the logo and pardon the potential loss of voice from this Nick crowd, but uh you know, we kind of pretty much run around New York City and let people do and say whatever they want into our microphone.
49:34
We created the bing bong trend years ago.
49:38
We created these NYX videos.
49:38
Uh unfortunately, there's a little bit of chaos outside of Madison Square Garden due to us, but yeah, we're pretty much let people say what they want in New York.
49:46
and uh it's it's pretty fun and entertaining.
49:49
>> What's the secret to a good street interview?
49:53
>> You got to keep it sharable and engaging.
49:55
It sounds obvious, but like why would you want to watch an interview of someone talking about something boring or or low energy or anything like that?
50:02
What would you want to send to your friends?
50:03
And that's kind of the the princip.
50:05
>> So you ask people about like language model diffusion and like >> always always always that's what we love talking about >> token pricing and all private credit. Yep.
50:18
>> That's what the people want to see.
50:18
We're going to start talking about that outside the Nick games I think. >> Yeah.
50:21
Where where uh so I I understand that a lot of the uh you know man on the street interview format these days.
50:26
It it can be very prepped.
50:29
can be uh there's a PR person that's pitching uh some successful business person for the what do you do for a living?
50:36
It looks candid, but in fact it's stage.
50:38
Uh do you ever participate in that or is everything candid?
50:41
Is that part of the brand?
50:44
>> We're 100% natural on the streets.
50:44
You don't know what you're going to expect. >> Organic. Okay. >> Yes, sir. Organic.
50:49
But it's really cool, too. >> Yeah. Yeah. Then uh >> Yeah, please. Go ahead.
50:54
Then my my followup is like then if you aren't prepping and understand who you're going to be interacting with, like how do you vet like where is the bar?
51:02
If you see somebody walking and they have a baseball hat on and they probably don't want to be bothered, is there some sort of social contract where you shouldn't like the the is there a doom scenario where everyone in New York is getting asked what do you do for a living 15 times when they walk to go get their morning coffee?
51:19
>> Yeah, honestly it's it's a bit of an epidemic going on.
51:20
I actually am kind of scared that I'm going to get pulled up on right now and ask what I do for a living or what song I'm listening to. Yeah. But >> Oh, listen. That's another one.
51:30
>> The the people the people of New York I think they distinctify side talk a little bit from that.
51:34
>> They know we're not there to to ask them something like I don't know so so cookie cutter like that. >> Sure.
51:40
>> Uh and I don't know now people come to us when they see the microphone which is great.
51:43
And uh >> yeah we we we just go with the flow.
51:46
>> How do you think about monetization?
51:48
short form monetization notoriously hard, but you have some merch.
51:50
Break break break down like the business side of the business.
51:55
>> Yeah, it's actually very interesting.
51:56
You would think that you would that you know I would have made a dollar off of Tik Tok by now.
52:00
I don't think I've ever gotten paid from Tik Tok or Instagram.
52:04
Literally not a dime, which is which is fine.
52:06
It's cool on Tik Tok through. But uh no, it's cool.
52:08
We we do a lot of branded work for companies.
52:10
So we've worked with everyone from the NFL to Netflix, Nike, Google, Amazon, creating a lot of content for them.
52:16
So, they'll come to us with a product or an event that they want to highlight and they say, "Hey, you know, you know how to get clicks and views and we kind of pretty much apply our formula to that and then create a really good video for them." >> Okay.
52:29
I I have uh like a hot take and I want you to walk me through whether you agree or disagree.
52:34
Um I think you should just do a midroll ad in a threeminut Instagram reel.
52:40
Uh because I see the I see the branded integrations.
52:44
It's like, oh, uh, there was this example in on subway takes where someone's take was Android's better than iPhone.
52:50
It went massively viral on r/android.
52:52
And then the second and then Google was like, "Oh, we'd love to work with you on another take that's the same thing, but it's branded."
52:59
And those sponsored takes, they often don't go as viral.
53:03
It's hard to get branded content to go viral.
53:05
But in some of these longer threeinut videos, if I'm locked in after 2 minutes, I would sit through a 10-second mid roll of like, "Hey, this was brought to you by this.
53:16
Thanks for sponsoring this video."
53:16
And then it's just all organic content around it. Does that work?
53:19
Is that already happening?
53:20
Are you thinking about that? What's the deal? >> Definitely.
53:24
We One thing we really like doing is kind of like a natural integration.
53:27
So, we just did something with Nike, which is really cool, where they hit us up for these NYX colored shoes. Okay.
53:33
Who better to get word out about these colors and side talk.
53:36
So we had people wear them in the episode.
53:38
We were talking about them. We had react to them.
53:41
I think that feels authentic because that's kind of something we would do anyway.
53:46
>> And it did really well. >> That's good.
53:50
>> Uh predictions for tonight.
53:51
>> Let me know if you're losing me by the way. >> A little bit.
53:53
We are losing you, but the energy is still there.
53:58
>> Um predictions for tonight are >> Nicks and five. >> Knicks and five.
54:02
What's the plan for covering the game?
54:04
Like, is this performance?
54:06
But but isn't it in your isn't it kind of in your best interest for it to be all se all all seven games?
54:12
Like, you know, more more attention like you kind of want, you know, you we we obviously want your Knicks to win, >> but you don't want them to win too fast.
54:22
>> It's got If we can guarantee a win in seven, I would like seven.
54:25
If we can guarantee seven, that would be great.
54:29
But listen, we need to get the win more than anything. >> Okay.
54:31
What what what is the process of a shoot day?
54:34
How early are you getting there?
54:36
Who are you bringing with you? How big is the team? >> Definitely.
54:41
It depends on what we're doing.
54:42
We do such a wide variety of stuff.
54:44
We can randomly go to a hot dog eating contest in Coney Island.
54:46
We can go highlight a random character in Brooklyn.
54:49
We can go obviously film these Knicks games.
54:51
So, it's pretty dependent on what's going on.
54:53
But with the Knicks games, we kind of show up in like the third quarter as it's ending.
54:58
Unfortunately, we can't bring our equipment into the stadium.
54:59
So, we lowkey just have to stand there and watch the game on our ESPN app and hope that the Knicks win or we hear, you know, someone discord, something like that.
55:07
And then the chaos erupts. We go into the chaos.
55:10
We're there for like hours and hours at a time.
55:11
And at the end of the day, people see, you know, 55 seconds, but we're out shooting for like eight hours throughout these playoff games. >> Got it. >> Wild.
55:22
>> Well, thank you so much for coming on. We appreciate you. Good luck.
55:24
Yeah, we will uh we'll feature some of your videos in uh in tomorrow's episode from tonight. Good luck out there. Have fun. Be safe.
55:33
>> Yeah, we'll talk to you soon.
55:34
>> I'm hoping for I'm hoping for Nixon 7, personally, but can't guarantee it yet.
55:40
So, have have fun out there.
55:40
Thank you for Thank you for this walking talk. >> Thank you guys. Appreciate you.
55:46
>> Yeah, great to hang, dude. >> Appreciate it. Thank you.
55:47
Let me tell you about Crowd Strike. Your business is AI.
55:51
Their business is securing it.
55:51
Crowd Strike secures AI and stops breaches.
55:53
We will be joined by the CEO of Snowflake in just a few minutes.
55:59
Uh, thank you for tuning in.
56:02
Um, we need to move on to the next story.
56:07
>> Canon says, "The cashier at Home Depot just asked if I want to round up to support the SpaceX IPO."
56:14
>> SpaceX IPO is going to be big Friday.
56:14
We might have a surprise guest. It's going to be fun.
56:18
Also, there's this crazy meta story going on.
56:21
not met of the company, but like story about a story behind Tai Morse.
56:26
He's going all out trying to get an Elon interview and it's been fascinating watching him build the craziest set in podcasting history.
56:36
We're rooting for you, Ty. Good luck. Hopefully, you land it.
56:38
Uh it is the quiet period, so it might be difficult, but I think that there's a plan to extend the effort and uh everyone's rooting for it to happen.
56:47
It's been viral on X many, many times.
56:48
Also, uh, Bloomberg's reporting that the SpaceX IPO is 4x over subscribed.
56:54
Do they know that already?
56:57
Does do you get that information at this point?
56:59
That's really good news for the market, for the for the overall.
57:03
There was an interesting article.
57:05
I mean, we can go through some of this.
57:07
Uh, I think this was in the economist SpaceX uh talking about um where is it?
57:16
Can the market swallow SpaceX anthropic and open AI?
57:19
Watch out for indigestion.
57:22
Um but uh they talk about the fro the float the free float.
57:25
This was very interesting.
57:27
So uh obviously the the company might be worth a trillion dollars.
57:32
How much of that trillion dollars is actually actively being traded?
57:37
Like sure, Fidelity might at any point in time have a price at which they're willing to buy more and a price at which they're willing to buy to sell their stake.
57:46
Um, but founders are often locked up.
57:48
Founders often want to maintain control.
57:50
Employees are locked up for a certain amount of time.
57:52
Certain investors might be locked up.
57:54
Unless you're Bill Gurley, don't try and lock him up.
57:57
He nobody can lock him up. He's a wild man.
57:59
But uh the float is important because if there's if it's a trillion dollar company and it's the CEO passed away generation ago and there's all financial managers in it's only owned by hedge funds.
58:12
This was the story of Take Two before Strauss Zelnick came in.
58:17
Take 2, the makers of Grand Theft Auto.
58:19
Uh it was entirely held by uh by shareholders by financial investors and they were unhappy with the management team and the management team didn't have any equity.
58:27
And so he was able to raise his hand and say like, "I'll run this thing."
58:30
And they were like, "Absolutely, thank you."
58:32
And they let him take over the company without really putting anything up.
58:36
He didn't need to do a hostile takeover like bring a bunch of capital to bear to get control of that company and become the CEO.
58:42
And it's been a fantastic success for him and fantastic success for Take Two shareholders who stuck along stuck around for the Strauss Zelnik era.
58:48
But uh the the the float matters and at Microsoft it's 100% floated.
58:54
The free float is 100% because the founders have moved on and divested and they're not locked up.
59:01
At Apple, it's like 99%, Broadcom's also at 98%, Nvidia at 96%, Amazon's at 91%, only Jeff Bezos is considered sort of off the table.
59:13
uh Alphabet at 90%, Tesla at 89% and Meta is notably of the big tech companies the lowest free float at something like 88% 86%.
59:28
And the reason for that is because Mark Zuckerberg has a lot of control.
59:31
If he sells his stake, he loses control.
59:33
So no one's expecting him to sell even if the stock goes way up or whatever.
59:36
Like he's going to maintain his position because he wants to run that company.
59:42
Now SpaceX is in an interesting interesting place.
59:44
So 13% of Meta shares are owned by Mark Zuckerberg.
59:46
SpaceX plans to release locked up shares in a series of tranches.
59:50
Uh if its IPO issues $75 billion of shares valuing the firm at its hoped for $1.
59:56
8 trillion valuation, the initial free float because if you buy the IPO, you're not locked up.
1:00:04
You can sell the next day if you want.
1:00:06
Uh but this means >> technically it's it's not uh like the whether whether somebody's like buying through you know investing through like JPM, Morgan Stanley or Goldman all these places like >> technically you can go and sell they just might restrict you from other IPOs.
1:00:27
So you would expect like serious and and this happens like on all the different apps you know Robin Hood public etc.
1:00:31
If you're buying into an IPO, they they basically are asking you nicely, do not sell. Uh don't don't flip. Don't flip. Uh here. >> Yeah. >> But people can't.
1:00:43
And so with how many retail dollars are flowing into this, I expect a lot of people that are buying >> um buying, you know, you know, basically buying the IPO. Yeah.
1:00:57
>> To start trading almost immediately. >> Exactly.
1:00:59
But importantly, it's only 4% of the free float. So only 4% of that 1.
1:01:04
8 trillion, it's a lot of money, but it's only 4% will be really free trading.
1:01:08
And then of course during the IPO process, you're vetting investors and you're trying to get the people that will hold forever.
1:01:13
And Elon has done a fantastic job of that with the indices.
1:01:18
So he's gotten uh the NASDAQ, the uh S&P, and a few others to uh to uh like commit.
1:01:27
So uh although NASDAQ has already shortened the seasoning period before index inclusion to 15 trading days, uh the Russell slashed its waiting time to 5 days, most share indices weight firms in proportion to the value only of the shares released for public trading.
1:01:43
And this is important because people look at the S&P 500 and say, "Wait, S&P 500 uh a$ two trillion dollar company coming into that.
1:01:51
your weight if you're just buying the S&P 500 is going to be more than 1%.
1:01:55
It might be in the singledigit percent. That's a lot.
1:01:58
And we've been talking about the S&P 499 for the Bears, right?
1:02:02
Uh in fact, the initial weight in the S&P 500 will be around 0. 1%.
1:02:08
Because it's just that 4% free float and then it increases over time as the shares get locked up.
1:02:12
We actually have a chart here of how the lockups work.
1:02:13
And it takes basically >> the full year almost two for everything to get unlocked.
1:02:21
And it's also triggered based on share price appreciation.
1:02:25
So if the shares trade up 30% or more then more shares become unlocked and the road to 100% uh lock lock up ending uh happens slowly.
1:02:37
And so uh just a little bit of an interesting deep dive into how the SpaceX IPO will fare.
1:02:42
Anything else to talk about on that story before we move on? >> No.
1:02:47
Apparently, Senator Warren, >> yes, >> has urged the SEC to halt SpaceX's IPO, citing governance risks, Elon Musk's control, and potential foreign, especially Chinese investment concerns.
1:03:00
>> She also highlighted SpaceX role as a US defense contractor.
1:03:03
Uh, >> Senator Warren has never met a business that she liked, >> I think, except maybe >> large financial institutions. >> Who knows? >> She she likes those.
1:03:17
>> Well, we have our next guest in the waiting room.
1:03:19
Uh, Sidhar Ramaswami from Snowflake.
1:03:22
He is the CEO and we're very pleased to be joined by him. Welcome to the show.
1:03:26
Thank you so much for taking the time. How are you doing? >> I'm well. Thank you for having me.
1:03:32
>> I can imagine you're doing well.
1:03:32
the the company is doing fantastically.
1:03:34
I am interested to hear I mean obviously there's been an emotional roller coaster but I'm more interested in the data that you're seeing that shows acceleration so much promise so much opportunity in this particular business.
1:03:53
>> Yeah, data is the foundation um for insights.
1:03:56
I worked in a data company, work in the search ads team at Google where great data was the flywheel that made it into a great business.
1:04:02
That's what we aspired to do with uh each and every one of our customers.
1:04:07
AI has done a couple of things.
1:04:10
It's made the act of bringing data into Snowflake doing >> just a whole lot easier.
1:04:14
Pipelines uh that used to take weeks to set up or migrations that used to take years um now have dropped to days and months um which is pretty remarkable.
1:04:24
But people also realize how much more power they can get by having great conversational access to their data.
1:04:31
And now with things like Snowflake Co-work, how you can actually solve pretty meaningful business problems in an interactive way.
1:04:39
It's that it's that loop that is compounding for us that drives a lot of our growth.
1:04:44
>> Is has there been any movement downstream for Snowflake?
1:04:46
Whenever I hear the story of a a data lake, a properly managed data warehouse as an entrepreneur even in a small business, I think that sounds amazing.
1:04:58
That sounds like the dream.
1:05:01
And then I see the companies that you know actually have the resources to onboard uh be you know immensely large and obviously that's fantastic for your business.
1:05:10
Uh but I'm interested in in what the future looks like.
1:05:14
Is there a world where a 10person startup with AI agents and and these pipelines that can be built much faster can actually set up their business and then not have to go through a cumbersome migration in the future.
1:05:30
>> Yeah, we are seeing that happen in front of our eyes with Snowflake's own data team.
1:05:34
>> You know, they like with every other company used to >> years long.
1:05:38
You ask them for a feature, they'd be like here take a ticket come back, you know, in a few quarters.
1:05:41
um what they've been able to do with agentic AI is grind through those kinds of backlogs very quickly. Absolutely.
1:05:50
Data and agents are going to make it um much easier to set up these things and have them evolve as they go along.
1:05:56
I'll give you uh examples from we rewrote our both our support teams and our S sur teams uh software.
1:06:04
teams uh software. These are the folks that take care of different kinds of problems that we get into an um and the net result of that by agents investigating the problem and if you see enough instances of a problem that you have not seen you go figure it out go write a tool for it it's now in a
1:06:23
self-healing loop I increasingly imagine that more and more operations get into that kind of a mindset where there are very competent agents uh sets of recipes you can call them skills there'll be some other name tomorrow uh that take care of problems, but it it is a self-learning loop. That's what AI makes
1:06:37
That's what AI makes possible. >> Got it.
1:06:40
Where are we in the path to like fully automated migrations, fully automated new integrations?
1:06:49
Uh it feels like every new model release, we're like, "This is the one. This is the one.
1:06:54
The non-technical CEO is going to be doing the job now.
1:06:57
He's just going to say, hey, we we have two systems. Go fix it. Go integrate it.
1:07:01
Uh get it into Snowflake.
1:07:04
it's going to happen and then reality sets in and there's not just token costs and cost of that to to delegating to AI but also just the real world of of other decisions that need to be made that might be outside of the context.
1:07:17
So what are the the the remaining uh challenges that uh come up when you're building a new integration or working on a migration?
1:07:30
Yeah, I think the timelines are definitely shrinking.
1:07:32
Um, when it comes to legac migrating off of legacy uh platforms, the on-prem ones.
1:07:37
Yeah, >> we used to think of them the biggest ones as running into say like three odd years and these strike terror into every company's heart because it's three years of like really anxious development work and waiting for results and hoping things don't go wrong.
1:07:52
To give you very uh concrete numbers, we now feel a lot more confident that we can tackle really tough migrations and have them finish in a couple quarters.
1:08:00
That's a big deal is going from 12 down to one or two.
1:08:02
And um the migration team which work on top of agent harnesses that we create confident that the technical parts of these problems uh will be solved by the end of the year.
1:08:16
But there is the change management.
1:08:18
There is all of the other applications that are running on the old system that you have to validate on the new system and ensure that there is business continuity.
1:08:25
That kind of change management is still going to be non-trivial.
1:08:29
But the technical parts, the act of uh doing the migration or writing new software, these are the things that AI is absolutely driving down. >> What is the biggest?
1:08:40
>> And by the way, I actually go check out every feature that my uh team puts out um using our coding agents.
1:08:44
That's already a thing and honestly it's not that hard if you have a modestly technical background. >> Yeah. Yeah. No, that makes sense.
1:08:51
Uh what is the biggest lesson that you or the company have taken forward from Frank Slutman's era?
1:09:02
>> Frank was always about amazing go to market execution. >> Yeah. >> And we remember that.
1:09:05
And uh in fact our sales teams >> uh are incredibly wellversed now in our products and in AI as a whole because we created products that could work for them >> and having an effective go to market motion is the real strength of Snowflake.
1:09:24
Snowflake. Again to make this super concrete uh the number of uh use cases these are new projects basically within customers um that are account executives one went up by some 75% uh year on year um that's the power of
1:09:39
sort of using AI to get work done faster and uh that is a tradition that will that that will continue what we have added on focus that comes from a product first person like me and AI I um in sort of a generic way or about worry about things
1:09:57
like what you know what can it do um but I think sometimes what we forget is if you have a wellbuilt system that uses like the latest um it's actually an incredible delight all kinds of information that you would want I was at a conference recently
1:10:15
every piece of information that I wanted to know about a customer that I was having a conversation with I would look up even live if I didn't know I would do it in front of them and show them the power of Snowflake's AI. Uh, a big part
1:10:24
Uh, a big part of our success over the past couple of years has been internalizing this.
1:10:29
AI is a different way to think about product and product delivery and the companies that are going to succeed are the ones that truly internalize that that there is a big chasm between how we did it with web pages and clicking on things to these agentic flows that can truly solve problems.
1:10:46
Then how is the sales the role of the sales rep changing the sales associate changing?
1:10:53
Are you looking for any different skills or are you doubling down on what worked in the pre-agentic AI era?
1:11:05
>> First of all, a lot more familiarity with Snowflake is a must. >> Sure.
1:11:10
>> Because these folks now have access to products that show off what Snowflake is about.
1:11:15
I expect every sales rep to have on their phone and be able to show off the actual impact of having a product like that with access to information to each and every one of our customers.
1:11:28
When it comes to our solution engineers, these are the technical pre-sales people.
1:11:32
Used to be that we'd have six demos total for all of Snowflake and maybe they do a little bit of customization and say, "Hey, here's what Snowflake can do."
1:11:40
Now, pretty much all of them are capable of creating a custom demo tailor made for a specific customer in a vertical in a way that makes sense to them that will really illustrate the power of Snowflake.
1:11:53
It's that kind of empathy that they can use the technology that we create to drive.
1:11:59
I would say it's changed in a really really big way because a lot more possible, but the bar is also a lot higher in terms of what you're expected to bring to the table.
1:12:09
Can you walk me through the partnership with Amazon Web Services?
1:12:12
I I I I'm sure you compete in some categories, but uh how did this come together?
1:12:19
What does it mean for the future of your business?
1:12:24
>> Yeah, we work with the hyperscalers.
1:12:24
AWS is the biggest uh partner. >> Yeah.
1:12:29
>> And uh you know, we work with them on a number of different levels.
1:12:32
We work on product integration. >> Sure.
1:12:36
you know buy a lot of capacity from them a lot of GPUs that our inference team runs uh runs models on uh but where we really shine is in joint go to market uh AWS is an incredibly customer first company and if a customer says hey I want AWS plus snowflake they lean into
1:12:55
it they bring us and we really focus on how do we solve problems giantly uh for customers and there's a deep deep relationship both at the account level going all the way up to the CEO level that ensures that problems that come up in practice get solved very quickly. Um
1:13:09
Um it's it's the kind of partnership that uh one only dreams of.
1:13:15
It's incredibly effective in practice.
1:13:19
What's been your what's been your strategy for trying to predict uh AI progress because uh you know we have people on the show every single day somebody from a lab somebody that's an investor you know uh customers of the labs things like that everyone has a different opinion you have to weight
1:13:38
them differently based on their uh incentives but you're running a you know massive company that's leveraging these tools but at the same time you want to understand what the capability will one year out, two years out and it feels harder than ever to actually be able to predict that. >> I think this is a is a really good
1:13:55
>> I think this is a is a really good question and one that to deal with.
1:13:57
I don't think the software industry as a whole has internalized that this precious commodity that we all loved creating by hand which is software um the cost of creating new software is going to keep going down.
1:14:15
I tell people that the best analogy that I can give for them is to go back to let's say the 2000s when the cost of creating and distributing new content truly dropped precipitously and you have to work through what is the implication of something like this.
1:14:31
uh for example, I think simply getting deployed into an enterprise and being part of uh a workflow there is uh no longer uh a guarantee that things will going to things are going to keep continuing because switching um is also easier.
1:14:48
I I would say that's the first thing every every company that's in the world of creating software as a big part of their business needs to think through what is the durable value that they are creating.
1:15:03
I think it's a really hard question to internalize and so we spend a lot of time just thinking about it.
1:15:08
The second one and this is uh this gets to the heart of your question um and I honestly just have a pract um which is that this is a bad time to be making two and threeear plans.
1:15:20
anytime things are getting better or worse by 20 30% every month uh the human ability to predict these things which are exponential in nature um is uh you know it just doesn't it doesn't really work uh I joke to people that at this uh
1:15:37
point in time we should treat coding agents the same way we treat traffic any person is always going to look at uh look at Google maps first before they go somewhere because it's a allseeing guy agents are sort of becoming like that when it comes comes to when it comes to
1:15:51
software and the more you have that open mentality that more things might be possible with a new model release than possible before um the more you will be pleasantly surprised and you'll be ready to take advantage of it and so I stress that childlike quality in figuring out
1:16:08
what is possible obviously you know there are nuances to how do you put in work in such a way that even if the model got better you'll still be able to take advantage of what you have done before there are practical problems s like that. Um, but the biggest thing
1:16:20
Um, but the biggest thing that I ask for is that open-mindedness to what is possible today because a lot different from what was possible say a month ago.
1:16:30
>> Are you positioning Snowflake as a company that's in the token path or in the token stream?
1:16:36
I don't know if you've heard this phrase that's uh sort of new coinage for uh a position in the AI value chain.
1:16:44
Um, and it feels like you're firmly in the in the token stream in the token path, but I'm wondering about what you think about that category.
1:16:54
Is that is that narrowing you?
1:16:57
Is that of is that valuable to communicate to customers and investors?
1:17:03
Uh, what do you think of the phrase like in the token path?
1:17:08
>> It's a useful way to think about how value is going to get created. Mhm.
1:17:13
>> Remember back in the web era we talked about you know number of visits, number of unique visitors.
1:17:18
Um these were the things are weekly active versus monthly active.
1:17:23
These were the abstractions that we used uh to ground the numbers that various people would report in a reality.
1:17:32
reality. Um you can think of the token path as existing along this path that is using the power and knowledge that these models have um and creating value for customers and uh both in the data where our coding agent called KCO comes into play where we want everyone that wants to get value from data to be using it to all of the artifacts in that path but also in the consumption path u where our product is snowflake co-work where I
1:18:03
want every employee in every company to be able to access the valuable data that is in Snowflake but also in other applications via things like MCP um to be using co-work um and so we very much think about can we be in this path of
1:18:18
coding up the things that are going to drive consumption and creating value uh but also in the path of how does an enduser query enterprise data take an action on that data uh and also creating value. So, we very much think about
1:18:31
So, we very much think about this.
1:18:34
>> So, my my my last question is um you said it's a bad time to be making two and threeyear plans.
1:18:38
Uh how far are you looking into the future?
1:18:41
Do you have a one-year plan, a threemonth plan?
1:18:43
Uh what >> well, as a public company, you you have >> Yeah, you have to have forecast, but but more qualitatively, like what are you gunning for this year or over whatever time horizon you think is appropriate at this point in time?
1:18:58
Our overall strategy is very clear.
1:19:02
It's just what I described.
1:19:04
Um, anytime anyone wants to do something with data, I want them to use uh agent tools that are created by Snowflake uh Coco for example to be able to do that.
1:19:14
We want to have the best data platform in the world so that when whenever someone thinks, hey, I need a historical view of this or I need an OOLTP database to do this.
1:19:22
Um, they think of Snowflake.
1:19:25
um they can use any coding agent not just our own to be able to uh to be able to do that.
1:19:29
Similarly, we think a lot about how can we make sure that we sell co-work to CRO's and demonstrate what we have done internally within Snowflake just transform our own sales organization.
1:19:42
So we think a lot about how do we take our customers through the same journey that we have been through in terms of how can AI make a difference to the business.
1:19:51
There'll be a lot of details along the way, but that's the path that we are headed in.
1:19:56
AI leveraging the full power of data.
1:19:59
>> Well, congratulations.
1:19:59
Thank you so much for taking the time to get >> Yeah, great to meet you and uh congrats to the team on all the momentum. >> Yes.
1:20:05
Stay strong in the token path. >> We love to see it. We'll talk to you soon. Have a good day. >> Goodbye.
1:20:10
Uh let me tell you about >> agents meet the canvas.
1:20:14
Your AI agents can now create and modify your Figma files with design system context. Uh, yes.
1:20:21
Apologies for the technical issues cut out.
1:20:23
A lot of people are saying, "Is this AI? It's not AI. Is this AI?"
1:20:25
There was some type of clip.
1:20:29
>> I I've never seen that exact technical failure before where it cuts out entirely.
1:20:34
We've heard scratchy audio.
1:20:36
We've heard degraded video.
1:20:36
Uh, it the video was coming through crystal clear, but the audio was cutting in and out a little bit, but I think we got a lot of good stuff.
1:20:43
Uh, Coco, they should have called the AI agent Frosty Snowflake.
1:20:48
Frosty >> probably intellectual property.
1:20:52
>> Frosty the agent >> is a jolly happy guy. >> Coder. Yeah. Something.
1:20:57
>> Uh anyway, uh Citadel, Citadel Securities, former former employer of mine when I was an intern, uh wrote a post on tokconomics.
1:21:05
Uh they're one of the most significant hedge funds and they just dropped tokconomics. I have it here.
1:21:12
I have the second print out.
1:21:13
Tyler, you want to bring that over to me?
1:21:15
It's not what you expected is what they say.
1:21:17
So said uh we we we've argued for some time that agentic and complex workflows delivered by frontier models would be expensive to run constrained by physical bottlenecks and vulnerable to unrealistic expectations of frictionless deployment costs.
1:21:32
That judgment now looks less contrarian than it did when we set it out in February.
1:21:36
They're taking a victory lap over at Citadel Securities.
1:21:42
Amazon has now removed its token leaderboard.
1:21:44
Microsoft has canled clawed code subscriptions and there have been multiple reports of unexpectedly large token bills.
1:21:50
The salient point is that even the most powerful technologies must pass through the prosaic discipline of cost curves, capacity restraint constraints and marginal returns.
1:22:04
Adoption is therefore becoming less about what frontier models can do in principle and more about the price and scarcity of the inputs required to make AI operational at scale.
1:22:12
Compute, power, cooling, memory, bandwidth, and inference budgets are real and binding constraints.
1:22:19
They go into economic theory talking about the the the three functions that prices provide.
1:22:24
They signal scarcity, create incentives for substitution, and ration scarce resources toward their highest value uses.
1:22:33
This is the stuff that goes viral on Tik Tok. We're we're deep in it.
1:22:35
You heard of the keeping the viewer watching.
1:22:41
This is the most important thing.
1:22:42
You read the Citadel Security macro strategy report.
1:22:43
But they do say uh they ration scarce capacity toward the areas where marginal productivity and AI justifies the marginal cost of using it.
1:22:53
They're talking about ROI maxing for the economy at large.
1:22:54
Simpler models may be more cost-effective productivity augmenting pathway uh until physical constraints are eased.
1:23:02
We we hence see growing signs of a bifurcation in frontier versus everyday AI usage.
1:23:10
They actually shared some data here uh around token expenditure index dropping the log growth rate declining and so folks are pulling back a little bit and shifting towards cheaper models as they figure out a new workflow something gets unlocked and then they say hey we got to actually make this ROI positive.
1:23:29
So that's the latest from Citadel.
1:23:31
Let's see what Amazon Web Services is saying because they're blacking on the timeline.
1:23:37
They said more AI generated code doesn't make your team faster.
1:23:39
It might actually slow you down. Crazy post from AWS.
1:23:47
>> Would have expected it from like some AI thought leader, podcaster, Twitter person, >> AI bearer.
1:23:54
>> But yeah, uh it's funny coming from the AWS cloud gold check.
1:23:57
It's funny because it's also like if you were to write this, >> I think a lot of people would write this as more AI generated code doesn't necessarily make your team faster. >> No, just perfect.
1:24:12
>> No, more AI generated code doesn't make your team faster, which is just like factually incorrect for like a lot of teams that we know, right, that are like we're actually moving faster than ever.
1:24:21
Also, okay, uh read the next post in this thread.
1:24:24
It's like very clearly AI generated.
1:24:26
Okay, >> the real bottleneck was never writing code.
1:24:30
>> It's releasing it, debugging it, keeping it running well.
1:24:33
So when Honeycomb C Honeycomb io >> CTO >> CTO charity majors set a productivity target, she didn't chase 10x.
1:24:40
She chose 2x and built from there.
1:24:45
>> You think this is AI cuz the choose 2x, and amperand built from there.
1:24:51
>> I think it was human edited, but like that that uh >> that syntax feels a little AI. I agree. I agree.
1:24:55
Her team also skipped the mandates and built a set of AI values instead.
1:24:59
Every AI output has to have human owner.
1:25:01
>> If you don't want your name on it, it's probably not good work.
1:25:04
>> Yeah, that's it's reasonable.
1:25:04
Uh quality first, quantity second.
1:25:07
Uh so they had a podcast about it.
1:25:10
Uh coming from AWS, that's a wild statement.
1:25:12
This is this is very funny.
1:25:14
Um but yes, also funny with the backdrop of uh of a $500 million budget. >> Well, yeah.
1:25:22
And it's funny because uh Amazon is of course one of the largest investors in both OpenAI and philanthropic to >> leading coding model provider.
1:25:34
>> Hey, they didn't say more AI generated posts don't make your social media team faster.
1:25:39
Entirely possible it sped them up.
1:25:41
They certainly got 14,000 likes on this.
1:25:44
This is banger post from a corporate account.
1:25:45
Rare to see those numbers uh put up on the timeline from a corporate account.
1:25:50
So, congrats to the team over there putting up some fun numbers.
1:25:54
Um, Shriram Krishnan, uh, the I don't know how to differentiate them. He's a kerniejackson.
1:26:01
com investor advisor to bending spoons.
1:26:02
Uh, not the former A16Z A16Z GP who worked for the government for a little bit.
1:26:08
Uh, Shriram Krishnan says, "Not many Silicon Valley C VCs took our bending spoons intro in 2022.
1:26:16
These VCs thought it was beneath them to be associated with the company.
1:26:20
From 2024 onward, many clamorred for an introduction to sell their portfolio companies.
1:26:25
Bending Spoons uh is going public.
1:26:28
So, Bending Spoons will be a public company and they own a number of uh of you know historical companies.
1:26:37
They own AOL and have done a bunch of other stuff.
1:26:39
Eric Seuford broke down the >> They literally own America Online.
1:26:44
>> They own America Online. And it must be electric.
1:26:45
Uh I can't get into my mobile dev memo email.
1:26:48
I'm just going to skip this >> and move on.
1:26:52
But you can uh uh Eric Suford over at mobile mobile dev memo wrote a breakdown of the souvenirator wrote a breakdown of bending spoons filing to go public.
1:27:00
The Italian rollup operator that owns among many other properties AOL.
1:27:04
They own Eventbrite and Vimeo.
1:27:07
Don't they also own Evernote or am I getting that mixed up?
1:27:11
am I getting that mixed up? someone owned Evernote and the Evernote employee growth chart looks like a fast takeoff like post2020 uh COVID they went from like a thousand employees to like almost 6,000 employees or they triple or doubled the workforce and Bending Spoon's you know known as
1:27:30
sort of like the the operational excellent almost like a private equity firm private like they're you know they're going to do cuts they have done a few cuts but like the the actual headcount has like head held fairly stable And I think the business has gotten to a much better place. So interesting to see
1:27:45
So interesting to see there's a time and a place for uh being aggressive expanding when there's a big moment and the market's up and then there's also a time to consolidate and focus on operational efficiency. Be your partner.
1:28:00
>> Why is this guy so obsessed with uh GTA?
1:28:04
John, >> have you ever played GTA? >> Any GTA?
1:28:07
>> Maybe like uh >> GTA 3, GTA 4.
1:28:08
Maybe like for like two minutes. >> Two minutes.
1:28:12
>> And I'd be And I was like, "Wow, you just >> loser. What a loser. Never gamed in his life.
1:28:19
>> I couldn't get into it. I couldn't get into it." >> Yeah. You're rust guy.
1:28:21
You're You're >> everything.
1:28:24
Like I I just >> Even as a child, I was like, "Okay, everything that you can do in this game, you can just go do real life minus." >> No, you can't. It's all crimes. Wait.
1:28:34
>> Uh Jordy turned it down. >> He turned it down.
1:28:37
>> There was something like 10.
1:28:37
Maybe I'm pure of heart and I don't enjoy I was like okay like crap.
1:28:42
>> They were going to let me play for like 15 hours, 10 hours, maybe five hours, but I turned it down.
1:28:45
They actually wanted to pay me.
1:28:47
They wanted to pay me to play.
1:28:49
>> Yeah, they wanted to pay you to play.
1:28:50
Strauss said actually I I would love for you to play to play.
1:28:54
>> Um no I one of my first jobs I was a video game tester.
1:28:56
I got paid to play video games. >> Best job.
1:29:01
>> Play video games all day. >> They got you hooked.
1:29:04
>> Yeah, >> they they they got you hooked.
1:29:06
>> No, I was already hooked. I was elite.
1:29:08
Uh, so yeah, you go play the games while it's in development. You file bug reports.
1:29:11
You're quality assurance tester, but you're basically your job is to play video games all day.
1:29:15
Worked with a guy who was super chill and like he was just like, "This is my career.
1:29:19
Like I'm this is I'm happy with this.
1:29:21
Like I'm I'm I'm a video game tester."
1:29:23
And then I I I checked in with him a couple years later and I was like, "Hey man, like how's the video game testing business going?"
1:29:30
And he was like, "You know, I actually had a shift in my career. I'm a DVD tester now. it's even easier.
1:29:37
So, you just sit there, watch the full movie, and then tell the tell the people like, "Yep, like there were no audio bugs or there was no there there were no drop frames in the whole movie. I sat there.
1:29:48
I watched the whole movie and it looked fine. Thumbs up."
1:29:52
And then every once in a while he'd be like, "Oh, I was in the menus of the DVD and when I went to behind the scenes footage, it wouldn't play the correct video."
1:30:00
It it had the title for, you know, interview with the director and it actually played interview with a cinematographer. So, you got to fix that.
1:30:06
Like, these things happen like UI bugs pop up in DVD menus.
1:30:08
But what a wild job, DVD tester. It's a good one. Good one.
1:30:15
Anyway, this fellow is obsessed with GTA 6 and he's doing anything and everything to understand whether or not GTA 6 will be released on time.
1:30:26
There's two methods here.
1:30:26
If you want to know GTA 6 is going to get released on time, you can just shake down Strauss Zelnik for the truth.
1:30:31
Or you can do what he did, which is he set up acoustics around the Rockstar headquarters to measure audio levels near a meeting room.
1:30:40
And he discovered that his equipment had disappeared.
1:30:41
And when he returned to check on the latest data, he didn't stop there.
1:30:45
He started camping outside monitoring traffic and tracking the number of cars and their estimated value.
1:30:52
Oh, there's an executive in town.
1:30:53
that means something for the GTA 6 release date.
1:30:55
Uh oh, there's more cars coming in.
1:30:58
Probably a lot of people working late nights getting it out.
1:31:02
Uh burning the midnight oil.
1:31:04
And so he believes such data can be used to speculate on how many executives may be spending time at the HQ as opposed to regular employees and that an increase in both traffic and high-v value vehicles could indicate trailer 3 is coming.
1:31:17
He's also considering an even more insane idea, which is to measure the amount of oxygen in the area to monitor how many people are around. And he did it.
1:31:26
He started monitoring the oxygen around Rockstar North.
1:31:27
And he believes trailer 3 is coming soon after noticing that oxygen levels dropped to about 20 20. 3%.
1:31:33
We actually need to use basis points this time from 20. 3 to 20. 04 after 5:00 p. m.
1:31:39
suggesting increased oxygen consumption in the area after normal working hours.
1:31:44
He's also monitoring the number of cigarette butts.
1:31:47
I thought it was just a GTA 6 fan. I'm excited. Everyone's excited.
1:31:52
Maybe it's a Rockstar or a Take 2 shareholder.
1:31:54
You had a different theory.
1:31:56
What was your theory on his motivations?
1:31:59
>> I don't believe that people are going to go to these lengths unless they're making money on it. >> Okay.
1:32:04
And how would he make money?
1:32:06
>> There's a lot of like prediction markets with, you know, plenty of volume. >> Yeah.
1:32:10
could be trading the stock directly, but could also be betting on a prediction market. That's >> Oh, yeah.
1:32:15
If there is theoretically, if there were to be another delay, I'm sure the stock would probably react to that. >> Yeah, absolutely.
1:32:23
>> It is very very interesting. Fun story.
1:32:26
>> A hedge fund should hire this guy to do true >> other types of detective work.
1:32:31
>> Yeah, it's I it almost feels like it's going over the line. Maybe illegal. I don't know.
1:32:35
Like if I I I would be very frustrated if someone was standing outside of our Ultra Dome with a measuring everything to see if we're going to interview Matthew Prince yet next.
1:32:44
And we are because he's in the waiting room and we're bringing him in to the TBPN Ultra Dome. Welcome to the show.
1:32:50
Matthew, thank you so much for taking the time.
1:32:52
How are you doing, >> gentlemen? Good to see you. I'm doing well. >> Great to see you. >> Great to see you.
1:32:57
>> What's new in your world?
1:32:57
You made some acquisitions. You made an acquisition. Take us through it. >> Yeah.
1:33:01
Uh so we we bought a a company called Void Zero which um uh makes Vit which is one of the most popular developer uh platforms that's there.
1:33:07
Uh just an incredible team.
1:33:10
Uh Evan Yu who's the founder is just a just a first class human being.
1:33:15
Um someone who who uh our team is super excited to work with.
1:33:20
The team that he's assembled uh is is just great.
1:33:22
And I think that this is increasingly becoming the platform that uh is being used to power a lot of the uh agents that are running around around the internet and uh and a lot of those agents are running on Cloudflare and so we think it's just a really natural uh combination.
1:33:36
>> How simple is the synergy?
1:33:36
Is it uh you'll funnel those 130 million users who download uh vit every month?
1:33:43
I think it's 130 million weekly downloads uh weekly. Wow. Yeah, that's a lot.
1:33:47
uh into you know preferring Cloudflare, defaulting to Cloudflare or is there more synergy under the hood around uh developer integration, company integration like how are you thinking about this playing out?
1:34:03
>> Yeah, we continue we we plan to continue to uh leave it as an open source project uh and support it and invest in it uh that way.
1:34:10
We want to integrate it uh closely with Cloudflare's developer platform and make sure that Cloudflare is the best place to run any sort of VIT project uh that that you that you have.
1:34:20
Uh but it'll work in any any of the different platforms um as well.
1:34:22
And so, uh, we just really wanted to make sure that Evan and his team had the support to make sure they could continue, uh, to to really invest, uh, in in what was, uh, in what was just an incredible platform.
1:34:34
And, and we think that that is going to drive more developers to Cloudflare, uh, Cloudflare's workers platform as well.
1:34:40
>> What are the headaches for developers these days?
1:34:42
I I know everyone's concerned about token costs and token budgets.
1:34:46
Sometimes that doesn't show up for the developer.
1:34:47
It's more like the CEO that's worried about it.
1:34:49
uh is uptime more difficult to maintain?
1:34:51
We've been seeing screenshots of uh different uptime tracker status pages that have more and more red and yellow on them.
1:35:01
Like what's at the top of the stack and and how are you helping?
1:35:05
>> Yeah, you know, I think that um that you need a different architecture uh than than we've than we've had uh to build sort of the last generation of applications for what's coming with with agents.
1:35:16
If you imagine, you know, there are about 100 million knowledge workers uh in the United States.
1:35:19
If all of those knowledge workers uh had one agent which was which was working on their behalf, uh that would that would and and each of those was each of those agents was running in a container, a traditional container like something you'd get at a AWS or a Google cloud.
1:35:34
AWS or a Google cloud. um the amount of just CPU resources uh that would be needed to to run just those those agents uh assuming they've just had one agent per person uh is about 50% of the total CPU capacity that's generated uh you
1:35:51
know by all the different CPU manufacturers that are out there uh today and that's just the United States knowledge workers if you take that to the world then it's several times um you know 30 or 40 times the capacity uh of of GPUs uh and and CPUs that that are existing today. And so what we really
1:36:06
And so what we really think is that as these agents are creating code, you need a different platform for it.
1:36:13
And and Cloudflare was built, Cloudflare workers was built not on on containers as much, but on something um called isolates, which is a much more efficient uh technology.
1:36:20
And so what we're seeing is that as people are building these agents, as they're using them, uh it's just a much more natural place uh to be running them.
1:36:28
And it's I think why you're seeing more and more of the big AI labs have Cloudflare as the preferred target uh for where where their code gets run.
1:36:36
You know, OpenAI uh released a project um uh for their enterprise users a little while ago that again is is targeting us and and and we want to make sure that with with things like Vit uh as as first class citizens on Cloudflare that we can help power that future because again it's not going to be kind of the same uh system that we that we built with the hyperscalers.
1:36:56
it's going to be something different and and again I think that we have a really good shot of of building that different thing.
1:37:01
>> So can you help me understand more about the problems of like CPU bottlenecks and then maybe some of the solutions?
1:37:06
I'm just thinking about like >> I I I would think bandwidth would be an issue obviously CPU load but is there a world where we get to some sort of convention around maybe it's the robots.
1:37:18
txt txt or just the user agent.
1:37:21
And when Google bot shows up or any other AI system, AI agent shows up, you're just delivering a it's almost an MCP server, but you're just delivering something that looks more like a JSON package, something that's a little lighter, little less CPU intensive.
1:37:34
Is there a path there to optimization?
1:37:37
It feels like we're in the era of like squeeze every ounce of performance out of everything.
1:37:43
Um, but what's actually going to happen here?
1:37:46
where where like are we just screwed or is there an opportunity?
1:37:52
>> No, you know, I think so I think two two different different problems.
1:37:54
So the first is um if you if you're as as as you ask um these these different AI systems to perform tasks on your behalf, um what what has to happen behind the scenes, especially as those tasks get complicated, is that they need to be coordinated in in some way.
1:38:09
So if you say plan a vacation for me or something like that, what what goes on behind the scenes is that there there's coordination there and the best way to be make that as efficient as possible.
1:38:20
What what agents are really good and and and uh and these uh uh various LLMs are really good at is actually writing code.
1:38:27
And so that code needs somewhere to live.
1:38:28
And the problem is that if that lives in a container, then you've got to bring in an operating system.
1:38:32
You've got to bring in all all the tooling and everything else.
1:38:35
And that's actually extremely heavy weight.
1:38:36
And so that's the first place that you've got both a CPU and a GPU bottleneck that's there.
1:38:42
And so with with something like um Cloudflare workers and isolates, you it's just a much lighter weight uh system.
1:38:48
And so that means that you can have more agents running on the same CPU infrastructure and and be able to provide that.
1:38:53
And again, I think that's why a lot of these next generation uh uh tools are actually built using our our platform.
1:39:00
platform. you mentioned something else which is you know as as these agents go out and interact with the rest of the web um that you want to make sure that that is done in the most most efficient way and so if they're pulling down you know all of the HTML uh from a from a web page if and it's and those web pages
1:39:18
are designed kind of to be consumed by by humans there's just a lot of croft on that that isn't isn't necessarily uh as important and so some of the things that we're doing are you know for those those customers of ours that want to make sure their content content is consumed by by agents. We want to make sure it's as
1:39:32
We want to make sure it's as accessible as possible.
1:39:34
We are automatically converting things into markdown which is a much simpler uh uh system that that saves you a ton of tokens.
1:39:42
It saves you a ton of processing.
1:39:44
It means that your context window uh can be can you can fit more useful information into a into a context window.
1:39:51
So, I think there's a ton of optimizations and we're helping both on the developer side as well as on the content side making sure that we can uh have have these agents be as powerful as possible and and get as much done as possible.
1:40:04
>> John Gruber total victory.
1:40:04
You know, he invented Markdown. >> I did not. >> Rubinator. Yeah. >> No way. >> Yeah. Isn't that amazing? >> Yeah.
1:40:11
And I mean it's it's one of these things that just you know it it turned out it was it was ahead of its time but it's uh it's such a key for making sure that we can take information and make it you know as compact as possible.
1:40:22
>> John Gruber created the the the format for God.
1:40:27
It's a funny funny world we live in.
1:40:29
>> Talk about uh it was last week you guys announced that the threshold had been passed around agent versus human traffic. Talk about that moment.
1:40:36
Did it happen sooner or later than you expected? Uh >> much sooner.
1:40:44
>> Yeah, I mean it was I I at um the end of last year uh so the end of of um 2025 I said that I I thought that we would pass um uh bot traffic um and that's that's you know AC across the board.
1:40:58
So that's like Google's crawler but also you know the new agents which are which are coming out I thought would would pass human traffic by the end of 2027.
1:41:04
uh about 3 months ago I revised that um based on the traffic that we were seeing to say that it would actually be in the first half of 2027.
1:41:13
And so the fact that it actually happened in the first half of 2026 >> uh is just is just has just been extraordinary.
1:41:20
And it just shows how quickly uh this this is is growing.
1:41:22
And and the and the real key here is that I think that if you know you or I as humans were researching to go buy a digital camera or something, we might visit five websites and do you know a little bit of research and some price comparison.
1:41:36
You know, you just watch as you use these agents.
1:41:38
They have boundless uh attention to be able to just go to maybe 5,000 websites to find exactly what what you're you're looking for, the best price, the best delivery, the best service, and everything that's that's there.
1:41:52
And so that's just driving an enormous amount of consumption of of of the internet.
1:41:57
At the same time, the other thing that's happening is that for a long time since about 2015, um the the web has kind of plateaued.
1:42:04
there were not there were there were more websites that were um being shut down than were being created uh during that time.
1:42:10
In the last 18 months though um we're we're back to the the web growing uh at a rate which is is exponential um and and what and and it looks um sort of similar to what what was happening back in the early 2000s in terms of in terms of growth of the web.
1:42:25
And so I think you're seeing both more um sort of consumption of what's online, but you're also seeing more things online.
1:42:33
uh as it as it goes forward.
1:42:36
And so in both of those directions, you know, that's that's that's going to continue to just drive, you know, more and more use of of the uh of the internet.
1:42:43
And I wouldn't be surprised if, you know, going going, you know, forward, say, 5 years, that that bot traffic will be a thousand times uh human traffic on online.
1:42:52
And we've and we've got to make sure that we make the internet work for for that new future. >> Yeah.
1:42:57
like every year we're just going to add another nine to the 99 99.
1:42:59
999% of internet traffic is bots.
1:43:04
Uh do do you know what the baseline was like pre-I were we at like 1% bot traffic?
1:43:08
I mean I've said >> it was it was more it was more than that.
1:43:14
It was about 20% um for for Yes.
1:43:18
And it's, you know, Google was the was the largest and then and then obviously there's a lot of of malicious things that that run around online, but that was that was around what what it was um for and it was pretty stable um over you know the history of Cloudflare at least.
1:43:31
So where where we could measure it so since 2010 um you know it was always kind of around that 20% range and then uh and then it's and then it's grown you know it's it's in the last few years it started growing steadily and then it really accelerated uh in the first half of of uh 2026.
1:43:45
first half of of uh 2026. Um I want to talk about the inference stack I guess like we're seeing two things play out sort of a bifurcation like a WWDC uh Apple's launching ondevice inference that's uh it's not going to be frontier but it's going to be usable for sure on your phone on a computer uh they can obviously go to their private cloud as well and then you have like the new
1:44:11
NVL72 models there's stuff in between on cerebrus chip fast but expensive Then there's everything from uh oh it runs on commodity hardware but it's still pretty big you need a real desktop for it and I'm wondering about uh how
1:44:26
you see Cloudflare fitting into that you it would be a logical extension to Cloudflare workers to have inference on the edge inference in different places like where do you see yourself offering inference if at all? Yeah. Well, we we Yeah.
1:44:40
Well, we we we it's it's actually kind of funny.
1:44:43
Um back in uh 2022, uh we we issued a a press release that there was a um a graphics chips company that we were going to partner with uh in order to put GPUs at the edge edge of our network. >> Yeah.
1:45:00
>> Um that graphics chip company turned out to be Nvidia.
1:45:03
>> And uh and um and and what was what was amazing was at the time it was just crickets. Like nobody cared. >> Sure.
1:45:09
And uh but but we we had we had and and so we we sort of did a little of it, but we hadn't really um rolled it out broadly.
1:45:15
Uh and and what was funny was and then you fast forward two years later and all of a sudden everyone cared.
1:45:21
And so we we basically just reissued the exact same press release and um and and now you know that's become that's become a pretty pretty important part of our our our business.
1:45:30
Um and so today you can run inference at the edge of Cloudflare.
1:45:33
Um and and because of the fact that we're, you know, in over 350 cities worldwide, you know, we're within within milliseconds of the vast majority of the world's population, we become a very natural place for inference to happen.
1:45:43
I I my my working assumption has always been that about 50% of the inference that happens will be on device, whether that's your your phone or your or your um laptop or or whatever.
1:45:55
Um, but that there needs to be some standard protocol where your phone or your laptop or whatever that local thing is can hand those either longunning tasks or larger tasks off next to the network.
1:46:06
And so that would be to us.
1:46:07
And then if if for some reason you need something that's more than that that you could hand it back to, you know, some centralized data center with um, you know, with with more more capacity than than we may have.
1:46:15
And I think that that that's what um that's what we're increasingly seeing.
1:46:19
Um I think my my assumption was a little bit um probably off.
1:46:24
Um I think actually think less is going to happen on device today because I think more and more of the tasks are going to be these longunning uh tasks where it's not going to be just you know what what's you know what's the you know what's the temperature in in New York today.
1:46:39
Uh instead it's going to be something like help me plan a a vacation go you know take into account all of these different things shop for uh you know different hotels in different places.
1:46:48
plan all the um plan all the uh the the travel between between the different locations.
1:46:55
Here are the criteria I have.
1:46:55
And that might be something that takes, you know, maybe, you know, not not going to be seconds.
1:47:00
It might be, you know, minutes or hours or days in some cases to run.
1:47:05
You're going to have I feel like you're going to have longunning agents just for Park City snow forecasting and reporting tracking like individual individual individual runs, you know, maybe a network of maybe a network of drones that are identifying what's tracked out >> optimizing your routes on the mountain. Lot >> of stuff.
1:47:26
>> That's exactly that's that's exactly right.
1:47:27
Hopefully hopefully it snows this year unlike last year. So >> yeah.
1:47:30
So, uh, what are the decisions when you're doing inference on the edge?
1:47:35
Because it's sort of a hard pitch to say, I'm going to shave off 300 milliseconds, 600 milliseconds, when you're talking about a 20 minute workflow or even a even a 10-second workflow, you're getting me a 3% boost.
1:47:52
Is that going to make the difference?
1:47:53
But are you optimistic that there will be sort of like a in between state where there will be inference that happens and it needs to happen within the request loop under a second and then the the latency matters. >> Yeah.
1:48:07
Well, I think I think uh there there are sort of three different um buckets why why people are choosing to run inference tasks uh on Cloudflare.
1:48:13
I think I think the first bucket is exactly what you said. It's just it's latency.
1:48:19
And so especially for those things that have human and computer interaction where where you know any delay feels like it's it's it it hurts you that people are trying to make sure that they can get the best best performance possible. >> Yeah.
1:48:31
For voice models I mean it feels like that has to Yeah.
1:48:33
voice models have got to be local. Yeah. I love that. >> Yeah.
1:48:37
And so and so that's and so you know again there's a case for that.
1:48:39
I think the second case is that a lot of times for either privacy or regulatory reasons, people want to keep things as close to where they physically are as possible.
1:48:49
And so, you know, I think that a lot of especially in Europe and otherwise, uh, of places think that, you know, they made the mistake with the internet kind of 1. 0 and 2.
1:48:56
0 of having everything go back to Ashurn, Virginia, everything go back to the United States.
1:49:03
And so, I think that there's there's a real kind of sovereign uh desire to keep things local.
1:49:08
Uh, and that's important.
1:49:09
And then I think the third thing is we we actually can be oftent times much more cost effective uh because of the fact that you know we we are this we are a network and and because of how we're interconnected with networks around the world bandwidth for us is effectively
1:49:23
free and so it's very easy for us to get things onto our network and because of where we deploy uh systems we often are in places where we don't have to pay for the space or the power which allows us to then pass those savings on to customers. So it can be significantly
1:49:36
So it can be significantly more cost effective uh to use inference with us than it can be in some other other places.
1:49:42
That's somewhat counterintuitive, but but um but it's sort of the nature of what we've what we've built and kind of the secret to what's allowed Cloudflare to um to to to to deliver as much as we have over time.
1:49:54
There >> there's a ton of uh ton of debate over data centers, how they get built, how big they should be, what we need, the push back, regulation, etc.
1:50:05
uh how are you is any of this affecting your business?
1:50:07
Are you thinking about how you position your footprint in the real world?
1:50:12
Uh how are you processing the the evolving discussion around data center construction?
1:50:20
>> Yeah, I mean I so so first of all I mean we we're going to need more data centers.
1:50:23
Um I think a lot of a lot of the discussion is is somewhat silliness.
1:50:27
I mean the the water um consumption I mean these are closed loop systems.
1:50:29
I mean, a golf course uses more water than than um probably all the data centers in the United States over over the course of of a year.
1:50:37
So, I think that I think, you know, there there's there's there's a lot of kind of silly u concerns that are there.
1:50:43
Um and but but there are but there are real concerns as well.
1:50:45
We've got to make sure that there are efficient ways to get power to these things that it's not taking power from from other other systems that the grid can support support that.
1:50:54
And so, I think those are all all um all good considerations for Cloudflare ourselves.
1:50:59
you know, it it we're we're um we're a little bit different where in in the case of you know, if you're a AWS or a Google or um you know, a SpaceX, you're building one big facility or a handful of very big facilities and putting a ton of machines in that that one facility. Cloudflare is different.
1:51:16
We have a ton of machines, but we spread them massively all around the world.
1:51:20
And oftent times we we want to go into the places where networks come together.
1:51:23
And so those those can be you know some of the oldest um data centers in the world and in any of those facilities we may not have you may we may have only hundreds of machines um but over collectively around the world we've got you know what is you know effectively much larger than any individual data center which is which is out there.
1:51:43
So I think we're less less impacted from you know the new builds of data centers.
1:51:47
I think we're much more impacted by how do we find our way into the existing data centers and then how do we make sure that the equipment that we're deploying is as power efficient as possible because we're often given sort of here's the envelope that you have to fit your fit your uh your equipment in.
1:52:02
Uh and we're often guests of you know whoever the local ISP is or or um or or or whoever you know partner is that we're working with in in the you know the the thousands of buildings that we're in all around the world.
1:52:15
>> What's new with Italy? Oh yeah.
1:52:19
Um yeah, Italy and Spain like it's it's uh it's it's so your summer plans your summer world peace here.
1:52:26
Come on are limited dyed. Yeah.
1:52:29
No no no visa for me or or the Maui coast.
1:52:34
Um no it's uh it's you know I think it's it's interesting um it's been actually kind of puzzling um for us.
1:52:41
So all of this stems back so so for those who are who don't know in in both Italy and in Spain um there there have been pretty aggressive uh tactics to come after either me personally or or Cloudflare uh and it's all been driven largely by um the football leagues uh the the soccer league in those places.
1:53:00
And what they're concerned about is is piracy.
1:53:02
Um the the thing that's been ironic about it is like we don't like piracy on our network either.
1:53:09
Uh we don't make any money off of it. It it takes bandwidth. It steals resources.
1:53:13
And so we have a whole team that's working all the time to shut shut the pirates down.
1:53:17
And uh and yet, you know, the pirates are clever.
1:53:20
They find ways to to use our our system.
1:53:22
And because, you know, huge percentage of the internet uh sits behind Cloudflare.
1:53:27
You know, there's going to be stuff from time to time that we don't we don't capture.
1:53:29
It takes us some time to catch.
1:53:31
with leagues everywhere else in the world.
1:53:32
So, the NBA and NFL and and MLB and and the Premier League in in the UK and and others, we work really closely with them to if they identify a pirate stream for us to, you know, pull that down because again, it costs us money. We don't like it.
1:53:47
But for whatever reason, uh, La Liga in in, uh, in in Spain and, uh, and the and the and the League in Italy, uh, have have decided that instead of instead of partnering with us to take this down, they they would sue us instead.
1:54:00
Uh, and so so again, yes, it it it does cramp some of my summer plans.
1:54:06
>> Uh, I want to talk about IPOs.
1:54:06
Life is a public company.
1:54:09
There's a few entrepreneurs that are trying to go public this year.
1:54:12
Uh, you had a very successful IPO.
1:54:14
The stock is up more than a 100,000 basis points since you went public. Fantastic run.
1:54:19
Um, but advice for >> that's we're putting everything in uh basis points now sounds even better.
1:54:26
I mean the stock of a thousand% it's fantastic but 100,000 basis points sounds even better.
1:54:33
>> Bas more 110,000 basis points. Who's counting?
1:54:36
Uh but anyway uh what was what was your process like?
1:54:41
We were talking about the level of float lockups.
1:54:44
Like what were the hard decisions?
1:54:46
Uh were there things that we're talking about now with these big IPOs that weren't even on the top 10 of your to-do list?
1:54:54
Um what is important to a successful uh IPO and then running a public company?
1:55:01
>> Yeah, I think you know there are a handful of things that um that that that that are you know interesting lessons for us.
1:55:08
So first of all like I love being a public company.
1:55:10
It's um you know I think that um it's interesting you know um the the the in in jurisdictions where you don't have no fault divorce um spousal homicides are much higher >> um which is I think kind of an analogy to the difference between um private market investors and public market investor people.
1:55:34
>> It's really hard to fire your VCs.
1:55:34
>> It's really hard to fire your VCs. it's hard for your VCs to fire you and so as a result like actually it's kind of dysfunctional whereas public market like I love our investors because if I do something stupid they call me up and they say that was stupid I sold your stock >> and then we and we have a conversation
1:55:49
and sometimes they say oh actually that makes sense now and I bought it back right but I think you can have actually a much more honest much more real conversation and um and and that's and that that's felt actually just a lot healthier >> uh than than what you see in in in the market
1:56:03
>> it's it's so yeah it's so true I mean Uh there's every every everybody that's angel invested in any number of meaningful companies is going to have portfolio companies where you've fully given up on the company and you just try to be try to be nice and yeah help if you can but at at some point you're just totally disassociated with the investment and
1:56:23
you're just like uh and and that's that's that's a reality but um yeah so it's very healthy to be like yep I'm disassociated like moving on put my energy elsewhere >> in both and in both directions right I I mean, it's just a better it's a it's a better relationship and and so I I like I loved the process of going public. Um
1:56:40
Um I think the thing that we did um like it was it was an opportunity for us to really kind of retell our story.
1:56:47
Like you don't get to reinvent yourself very often, but the process of writing the S1 was incredibly um just an opportunity to sit down and really do it.
1:56:55
and and we you know really dug in and and told the story in a in a way that to this day we still refer to parts of that uh that that document to sort of explain what is Cloudflare and and how did we work.
1:57:06
Cloudflare and and how did we work. I think that the most clever thing uh that we that we did um and this was advice that I got from uh from Ryan Smith uh from Qualrex was he said um you know he's like what are you doing about friends and family and I'm like we're not going to do it because you can take 5% of a of a IPO and allocate it to friends and family and I didn't want to
1:57:26
like get my aunt like over if it if the stock went down or something I didn't want to have to explain that over Thanksgiving Thanksgiving dinner and he said no no no you're thinking about it wrong >> think about the people who if they, you know, if you, they owed you a favor, um, that they could make a meaningful difference in the future of of Cloudflare and then offer them to the ability to invest in in the IPO. >> Yeah. >> Yeah.
1:57:50
>> And I was like, is that like is that legal?
1:57:52
He's like, check with your lawyers, check with the bankers, but the answer is yes.
1:57:55
And I was like, some people are going to have like conflicts.
1:57:58
He said it doesn't matter like even it just the fact you offered it even if they can't do it will will mean that you know um that that that that they'll they'll always remember that super fondly.
1:58:09
And it was incredibly good advice.
1:58:10
And we made this crazy list of all of these you know people who you know who were like you know at some point that might be an important relationship uh for us.
1:58:19
And like 75% of them said yes.
1:58:23
and and they all made, you know, a ton of money as a as a result of it.
1:58:27
And it was um it was it was one of those kind of >> Now your aunt doesn't talk to you anymore.
1:58:33
>> My well, yeah, my aunt doesn't talk.
1:58:35
>> Like I missed out on $100,000 base basis point move. How could you?
1:58:40
>> I I I am always comforted that we actually so we we priced it >> uh $15 a share and it went up to and that's actually the other thing that's interesting.
1:58:48
everyone like Bill, Bill Gurley and I have fights about this um from from time to time about IPOs like you have a lot of control over how much the pop is going to be and we we sat with bankers and we were like listen we want it to go up about 20%. >> Yeah.
1:59:03
>> And and and they were like okay if you want it to go up to 20% you price it right here.
1:59:06
So we priced it right at that point and it closed the first day at 18 which is exactly up 20%. >> Okay. Okay.
1:59:12
But just to steal man a little bit is that is that a function of of your business?
1:59:15
Not to like be rude or anything, but there are some businesses that are just like meme stocks out of the gate and like they're in some weird thing and there's some froth and like they have less control or is it always controllable?
1:59:27
Well, and and part of that was like, did you think the fair value of your business was 18 or 20% higher than where you priced it?
1:59:35
And so you wanted like you basically wanted the market to to sort of repric and you didn't care about not >> Yeah.
1:59:42
I mean, but and they if we had priced it at like 13, it would have gone up more. >> Yeah.
1:59:47
>> But but the problem again like you you want to you want to play for the long term here.
1:59:52
And so my again my fight with Bill is you know he says you know that you should you should do um uh you know some sort of system where you where you actually you know price it at whatever the top price is.
2:00:03
The problem with that is if you look at all the companies uh that that have done that.
2:00:07
So the Airbnbs and and uh and others that they basically, you know, that's that's great for your private market investors, that's great for the VCs, that's great for any of the the the founders that are taking money off the table, but you got to keep running the company.
2:00:21
Like the IPO isn't the end.
2:00:22
It might be the end for the the VCs, but it's not the end for the operators of the company.
2:00:26
And you want to be able to kind of build that relationship and build that over time.
2:00:31
And so I think, you know, we worked incredibly hard to be very sort of choosy on who our kind of who we gave allocations to.
2:00:38
And we we we went through every single name on the allocation list.
2:00:42
Um and and and I think as a result of doing that, um you know, if you look at like who our top 10 shareholders are, it's been incredibly stable, uh over over time.
2:00:51
Uh and they're and they're they're folks who give us again great advice and and um and and have been really just really really great to work with.
2:00:59
So, I I I I'm sure that there are times you don't have it, but I was I was I was surprised at how the banks have a really good sense of if you price it here, here's what's going to happen.
2:01:10
And and in our case, it was it was, you know, to the to the penny.
2:01:15
>> Uh last question, how are you thinking about long-term planning as a CEO when we're on these exponentials?
2:01:22
we're on these exponentials? you know the the example of like agent versus human web traffic like you have you had like more data than almost anyone on the planet and your prediction was still off your original prediction was still off by you know 20 18 months 24 months some something in that uh range so how are you thinking about overall business uh planning when you know model capability
2:01:44
is increasing and and you're just seeing sort of exponentials everywhere >> yeah I mean I think um a couple of different things so one I think we're we're trying to make bets on on people who who can who can really function in in really dynamic uh environments and and that for us has been um somewhat different than I I think a lot of other people a lot of other people have cut
2:02:08
back for example on kind of early career hiring um we we you know Cleer's about 5,000 employees um a little bit less than that now um uh but we hired 1,111 interns uh this this summer so straight out of straight out of and they're amazing like they're just they're just killing it because they they are so native to to these tools and and again I think we've always thought of our job is training the interns but
2:02:34
this year like the interns are helping train a lot of us in how to how to how to do these things um you know really well so I think staying super dynamic and then you know from our perspective just making sure that we've got optionality uh going forward so you know it's it's it has gotten tougher uh to get equipment uh today um memory prices are are through the roof. Um, you know,
2:02:52
Um, you know, things are harder, but we've we we just have a team that's always constantly trying to figure out how how to to refine that.
2:02:58
Um, and then generally, I think we're we're trying to be, you know, big adopters of AI, not just in developer, uh, land, but also across the entire the entire business.
2:03:06
So, we built something called Cloudflare OS, uh, which allows, you know, anyone on our finance team, legal team, you know, accounting team, finance, uh, anyone that's out there at the company to be able to have access to tools and and and and really um, figure out how how their job uh, gets done.
2:03:24
The clever thing that we did to seed it was we actually set up this email address.
2:03:29
Uh, and initially we told the team that it was this magic AI model and if they just wrote to the email address and and ask it, you know, I I have this part of my job that I need to get done.
2:03:39
How do I get that that done?
2:03:41
Uh, it would, you know, sometimes it would ask more questions, but it would it would actually give you kind of a response back.
2:03:46
a response back. what what we didn't tell people uh was actually it was a team team of humans behind the scenes that was initially receiving these these emails and then they were using AI systems to kind of flesh things out but what they were really doing was recording all of the kind of key jobs to be done within the organization cuz you know you have to if you're going to have AI systems that can help facilitate this
2:04:08
you actually have to record what are the steps that people are are doing as as uh as a part of that and so as a result of that we've been able to seed this this Cloudflare OS now with the ability to do things like if you're on the sales team
2:04:19
and you have to do like an account plan, you know, you can literally just type slashac account plan and then describe what it is that you want to do and it will output that and that's made that's made our team so much more productive. Um, and again, I think it's giving us
2:04:31
Um, and again, I think it's giving us the flexibility to be able to respond to, you know, whatever comes next. >> Very cool. Final final question.
2:04:37
Are we going to are we going to see an exponential increase in the number of interns this summer?
2:04:43
Yeah, we see 10 10,000 interns. >> I don't know.
2:04:48
I mean, it's I I I I we've got to find a place to put them all.
2:04:50
And so it's I mean our office in Austin like we it's a big space in Austin.
2:04:55
We've run out of space there because we have so many interns there.
2:04:59
It's um I just I just think that there it's I think a lot of companies are doing this wrong where they're saying like we're we're not going to bet on on on the new people coming out of school because, you know, AI is going to replace them.
2:05:11
I mean, these these kids know how to use AI better than anyone else.
2:05:15
So, we're going to bet incredibly heavily on that.
2:05:16
And it's um and it's so far it's paying off.
2:05:20
I mean, these that like I think the person who eventually replaces me might might literally be one of the interns to today because they're just in so so so smart. >> It's happened. It's happened before.
2:05:28
I mean, I don't know if Satan Nadell was an intern, but he has that famous video of him as a product manager demoing Excel and Yeah. >> Oh, yeah. company man. >> I remember.
2:05:38
I mean, I I was on a bunch of of calls with Satia um before he was before he was CEO.
2:05:42
Um and he was he was just a product manager, you know, at at that.
2:05:47
So, I think just if you you can find really great people and uh and be able to goat out there.
2:05:57
>> You're the goat as well.
2:05:59
>> Uh yeah, you know, whenever you put up a 100,000 basis point move in your stock over just a couple years, you're in you're in contention for goat.
2:06:06
I I say uh thank you so much. Always a fantastic time.
2:06:12
>> Thanks for jumping on. Good to see you guys.
2:06:13
And by the way, I haven't talked to you since the the acquisition. >> Thank you. I I appreciate it. We'll talk to you soon. Have a good one.
2:06:20
>> Uh let me tell you about console. com.
2:06:23
Console automates I builds AI agents that automate 70% of IT, HR, and finance support, giving employees instant resolution for access requests and password resets.
2:06:32
We will be joined by Venode Kosla in just a few minutes.
2:06:38
>> While we wait, Mad's summer vacation is looking good.
2:06:41
Let's see what's on the schedule. >> What is on >> 6 a. m.
2:06:44
He's waking up on the wrong side of the bed. 7 a. m.
2:06:46
He's split between judging a book by its cover and beating a dead horse. Okay. >> 8:00 a. m.
2:06:52
This is where it gets spicy.
2:06:54
He's going to poke the bear.
2:06:55
>> Going to poke the bear.
2:06:56
>> And shortly after that, he's skating on thin ice right into walking on eggshells. >> Wow. And uh at 10:00 a. m.
2:07:01
he decides, "All right, it's time to address the elephant in the room." Okay.
2:07:06
>> At 11 gets a bit dark.
2:07:06
He's going to cry over spilled milk.
2:07:10
>> But at noon, he's letting the cat out of the bag. That's exciting.
2:07:13
>> And that's where it really ramps up.
2:07:14
He's starting at 1 to bark up the wrong tree.
2:07:17
And again, but at the same time holding his horses, >> but then at 2 p. m.
2:07:21
he's going to add fuel to the fire. >> That's fantastic.
2:07:24
>> And uh I can't really think of a more perfect summer day.
2:07:26
perfect summer day. John this >> you know you know there's uh you can share your calendar with someone else and you can share it privately so you can just see that I'm booked but you don't actually know if I'm having a meeting with someone maybe I'm interviewing Venode Kosa you won't see
2:07:38
that but you could build a plugin that changes the book to just these things that would be delightful maybe the next Google April Fool I would love that >> yeah if somebody just can only see that you're letting the cat out of the bag >> that's that's pretty exciting kind of mysterious >> okay don't interrupt him. It's an
2:07:55
It's an important meeting.
2:07:57
He's letting the cat out of the bag.
2:07:58
Uh but hey, he's just beaten a dead horse.
2:08:00
Maybe I should go grab coffee with him.
2:08:02
See if he wants to spin out on that. >> There's a website.
2:08:07
>> This is big turning Joe Weisenthal's tweets into full articles.
2:08:13
>> This is crazy because there's a there's a live show that turns his tweets into like minutes of just back and forth conversation and banter.
2:08:19
Uh but somebody made a version of our show that's just web pages.
2:08:23
So, they scrape every Joe Weisenthal tweet and they turn it into a full SEO optimized article apparently. Uh, it it is great.
2:08:33
And Joe looks great in this photo. >> He's looking good. He's looking quizzical. He's having fun.
2:08:38
Well, uh, let's bring in our next guest, Venode Kosla, the founder and managing director of Kosla Ventures.
2:08:45
Venode, thank you so much for taking the time. How are you doing? >> I'm doing great. Great to be here.
2:08:52
>> Fantastic to have you.
2:08:52
Uh, I want to kick it off with some of the recent deal making that you've been doing.
2:08:57
Your outlook on artificial intelligence is obviously front and center, but what has been the most recent fund raise that you've done?
2:09:04
Why did it get you excited?
2:09:10
>> Well, um, obviously there's a lot upon yesterday or today.
2:09:14
uh we are working a lot and doing a lot of new things uh that I'm pretty excited about is a company like Pmana >> but more importantly the broad category of autoformmalization which isn't talked about much but humans aren't very good at talking to AI incribing what they precisely want and that's where autoformmalization becomes important and companies like Gana become very important.
2:09:51
>> How did you get into auto formalization?
2:09:53
And I guess the big meta question is as an investor you have to consider is auto formalization on the direct path for the other labs.
2:10:02
How do you think about competition in this area?
2:10:03
Investors are going back and forth and it's been fascinating to see things that appear to be directly in the path of the big labs.
2:10:12
They're still growing revenues 234 billion.
2:10:15
We're seeing $60 billion acquisitions.
2:10:17
Uh everything seems to be going well up and down the stack.
2:10:19
But why is there an in why is there a particularly big opportunity right now?
2:10:29
>> Well, the labs they will do more and more over time, but I still see the opposite being very broad for venture capital.
2:10:40
notes uh in in clever are coming up with great ideas. >> Mhm.
2:10:48
>> One of example looks at the attack code and says it's not very precise because it was written by humans and you can make it into a machine checkable language. Mhm.
2:11:00
>> Mathematicians turn all their problems in this specific language which is hard for humans to understand.
2:11:08
Uh autoformmalization allows that kind of formalization.
2:11:12
So computers can operate on it precisely not probabilistically.
2:11:18
And that's a really important thing.
2:11:18
If you look at today's world, AI systems have been awesome in the rate of capability development be awesome, but they have some pretty big gaping holes. They still hallucinate.
2:11:31
I don't think hallucination is going away anytime soon. >> Yeah.
2:11:37
>> So, the reliability is low.
2:11:37
Humans aren't great at specifying what they want.
2:11:41
the specification problem. >> Mhm.
2:11:44
>> And then verification of what are you getting the right answers or are they hallucinating?
2:11:50
Those are all large interesting problems which are open for entrepreneurs.
2:11:58
Things can be built in certain domains. >> Mhm.
2:12:01
>> Uh where you build a lot of value and a a lot of capability that the general models don't do.
2:12:08
How are you thinking about applications of this technology in particular?
2:12:14
Uh it's been so hard to predict.
2:12:17
Uh the original GPT, no one was really expecting a knowledge retrieval service.
2:12:23
It just sort of went viral when chat GPT launched.
2:12:26
Then coding agents, uh some people had predicted those, but it was sort of an unexpected emerging capability.
2:12:32
Do you expect that auto formalization as the next critical frontier of AI just makes all the existing applications better or is it actually going to unlock new applications?
2:12:48
>> I think it will allow humans to use AI where it wasn't previously possible.
2:12:54
>> If you want to know what your bank account is, you can't have a hallucination. >> Mhm.
2:13:00
um if you have a medical problem you can't have hallucination.
2:13:04
So autoformmalization will significantly enhance existing models in domains like tax law.
2:13:12
You know you can have lots of tax uh startups but they can't formalize.
2:13:20
You take tax law, it's a very specific thing and that's an area pmana is uh initially to say can you formalize the tax code so computers can work on it precisely and not be subject to probabilistic answers or likely answers.
2:13:40
So it's very very important in certain domains uh to have precision, reliability, verification um and and that's where autoformmalization can significantly enhance the applications and still leverage all the power of LLAPS. >> Mhm.
2:13:58
Well, we're excited to talk to the founder of PMANA in just a minute.
2:14:01
Uh we've been having a few technical errors.
2:14:06
So uh we will end this here and hopefully we can have you back on the show soon to go deeper.
2:14:09
Uh thank you so much for taking the time to chat with us today, Venode. Uh it's been a pleasure.
2:14:16
>> Thanks for jumping on. It's an honor.
2:14:17
>> Have a great rest of your day. >> Great to be here. >> Thank you. Goodbye.
2:14:21
>> Um let me tell you about public. com.
2:14:24
Investing for those who take it seriously.
2:14:25
Stocks, options, bonds, crypto, treasuries, and more with great customer service.
2:14:30
Uh unfortunate technical difficulties.
2:14:33
I couldn't really hear anything personally, but fortunately >> the founder was discussing right now.
2:14:43
So, we're going to hear it again and again. Sorry about that.
2:14:46
Uh >> Nesh Aurora is on the timeline quote tweeting.
2:14:53
I like that he's you don't think of him as like a big poster.
2:14:57
Obviously, 73,000 followers.
2:14:57
He's successful but he said congratulating Prime Minister Narendra Modi on becoming India's longest serving elected prime minister 4,399 days of leadership earned the trust earned through the trust of 1.
2:15:10
4 billion people across three democratic mandates from lifting 250 million people out of poverty to making India the world's fastest growing major economy.
2:15:18
PM Mod's tenure has been nothing short of transformational.
2:15:22
Look forward to a continued US India partnership. Love that.
2:15:27
Yo Gash says, "Bro went from stopping global ransomware attacks at Palo Alto Networks to doing math calculations on PM Mod's calendar milestones like it's his final exam." >> I like it. >> Just in quotes, too. >> It's great.
2:15:40
And >> Nash appreciates it. >> Great.
2:15:43
Uh there's been debate over beverage media. Have you seen this?
2:15:44
This is a very I saw this eight likes, but I think this is an interesting post worth discussing.
2:15:50
So, Vasa, I don't know exactly this account before, but beverage media doesn't have the good nuts to call out celebrity brands objectively because the economy is fanboy of fanboying and CPG.
2:16:01
Celebrity founders sell tickets to events, but they don't always move the industry forward.
2:16:07
I think he's probably subtly subtweeting the maybe Tom Brady launch of the of the the the nuts, which is a reference to the coconut water brand that uh Tom Brady launched.
2:16:18
But this does happen all the time.
2:16:20
There's a lot of celebrity brands.
2:16:21
We've had some celebrities on the show to talk about their brands. It's fun.
2:16:24
Sometimes it works, sometimes it doesn't.
2:16:26
That's the fun part about interviewing a celebrity about their brand.
2:16:29
You kind of get to clock it and be like, uh, you know, the this person's really on to something.
2:16:34
I think the >> I think why they they named the brand that is they ran fully ran out of other names.
2:16:40
So that's literally the last possible name that they could choose.
2:16:44
>> We're not getting an invite.
2:16:47
uh Tom Brady will not be coming to our to our event and selling tickets.
2:16:48
Um but it's interesting because yeah, BevNet uh reports on uh CPG brands.
2:16:53
They're one of the insider media, sort of the tech crunch of CPG, at least in beverages.
2:17:00
They also have Nosh for food.
2:17:00
And there's a number of other CPG media and critics.
2:17:03
But uh if you got to have one of those celebrity founders at your event to sell tickets, uh you get into this weird you get uh sort of audience captured.
2:17:12
I don't know, celeb captured.
2:17:14
just get raptured in the in the aura of the celebrities and you have to stay out till 2 am parting with them.
2:17:19
It's craziness out there.
2:17:21
Uh but beverage media uh I think it's a it's just interesting in the backdrop of more more of a more and more celebrity brands.
2:17:30
Although it does feel like we're at peak celebrity brand at this point.
2:17:32
And I feel like the more modern celebrity strategy is just invest in companies that have a number of good VCs on the cap table. Be happy with that. You can still a farm. Yeah. Okay. Disagree. Explain.
2:17:49
>> I think celebrity celebrity brands have always been a thing.
2:17:51
They're just way more in your face because now celebrities have their own channels that they can communicate.
2:17:56
>> I'd say they're less in my face.
2:17:56
I I say we're on a decline.
2:18:00
We've reached top celebrity brand.
2:18:02
>> No, >> I'm quoting John F.
2:18:03
>> No, I think it's just getting it's getting easier and easier to start a business.
2:18:06
I think you'll see more celebrity brands.
2:18:08
I don't necessarily think you'll see more breakout brand. >> Okay.
2:18:11
So, your claim is is quantity is increasing.
2:18:13
>> I think it's probably getting harder.
2:18:15
>> Quality is decreasing.
2:18:17
>> Quality >> more distribution, but >> yeah. Yeah.
2:18:18
Greater greater I mean there's going to be some breakout celebrity brands. Yep.
2:18:21
There's going to be >> big exits.
2:18:24
Y >> uh there's going to be as many failures as ever, right?
2:18:28
I mean, my favorite example was like Travis Scott a few years ago launched a uh RTD like canned cocktail thing >> uh shut down I think within like a year like uh may maybe >> maybe maybe more and so I I think you're just going to see more and more and more.
2:18:48
>> Would you put the Johnny Knoxville Neocloud in the in the bucket of celebrity brands?
2:18:55
>> That is something I would like to see less. >> Yeah. Less 2K.
2:18:59
>> Less beverage, more NeoClouds.
2:18:59
I want to see Johnny Knoxville levered up. Yeah, right.
2:19:06
I want to see >> GPUs himself. >> Yeah.
2:19:09
>> I want him like operationally involved too.
2:19:11
Not just like a pretty face endorsement. Yeah.
2:19:14
>> Not just a pretty face.
2:19:14
Um, so anyways, I think more on the horizon.
2:19:17
Uh, and it's not necess, you know, a lot of I don't know.
2:19:22
I feel like some people don't don't like this, but I think it's fine. >> It's fine. >> It's fine.
2:19:26
You heard it here first, Jordy Hayes. It's fine.
2:19:30
>> You know what else is fine?
2:19:30
The New York Stock Exchange.
2:19:32
Want to change the world?
2:19:33
Raise capital at the New York Stock Exchange.
2:19:35
Our next guest might be there soon with Pmana Labs.
2:19:37
We have Ron John who's the co-founder in the waiting room.
2:19:42
Let's bring him into the TVP and Ultradome. Welcome to the show. How are you doing?
2:19:45
Thank you so much for taking the time. >> Doing great.
2:19:48
John Chi, uh, >> since it is first time on the show, please introduce the company.
2:19:52
We heard a little bit about it from Venode, but I want to hear it from you. >> Yeah.
2:19:57
Bravo, by the way, of getting being able to get Venode to just like personally pitch your company for 15 minutes on on internet television.
2:20:04
It's a it's a good uh it's a good sales pack. Good pitch. >> Thank you. Thank you so much.
2:20:11
Uh at Pmana, we we try to make AI more provably accurate. >> Yeah.
2:20:19
The current frontier of AI focuses on recall which is more along the lines of how do you make more answers come out >> but we are trying to focus on very specific mission critical domains like tax, legal, healthcare and governance where a wrong answer could be catastrophic.
2:20:37
>> You wouldn't want uh AI diagnosing you wrong and ending up in a catastrophic situation.
2:20:42
So that's where we focus on.
2:20:44
>> Uh and the technology underneath it is formal verification.
2:20:46
Yeah, >> I would like you guys to dive a little bit deeper and I can get more technical as it >> Yeah.
2:20:53
I I I'm just wondering um about the actual like roll out of this.
2:20:57
Like people are using AI tools for healthcare now.
2:21:01
They've been using WebMD.
2:21:01
People have been misdiagnosed with cancer for by random internet posts for like decades.
2:21:08
You go to WebMD, you're like, I have a headache and it's like you got brain cancer and then you freak out.
2:21:11
Go to the doctor and they're like, what are you talking about? Take an Advil.
2:21:13
But um so like who exactly is is not adopting AI right now because of the hallucination problem.
2:21:26
>> Uh one two major examples currently are like big four firms which are tackling tax and legal cases.
2:21:33
You would have heard about the Australian government suing one of the big fours >> stating that there is a hallucination which ended up happening.
2:21:40
uh and AI can actually make a loss on the fly. >> Yeah.
2:21:45
>> Uh if it wants to suit a particular situation, it can make up loss on the fly which can lead to catastrophic situations.
2:21:50
And uh another thing which we have noticed is also that insurance providers, major insurance providers of US have also moved away from ensuring AI outputs.
2:22:00
U so there is a very specific clause they have introduced to say that uh we will not ensure if AI is behind uh whatever you have said.
2:22:09
So those are some things which have come out specific to domains like tax, legal and healthare.
2:22:16
Uh for healthare we can even see the models themselves typically stating that you need to go to a doctor to verify.
2:22:23
We will we will be absolved of the responsibility. >> Yeah. How much? >> Yeah.
2:22:28
Imagine your CPA being like great I just I just finished your your return.
2:22:33
It's 99% chance it's accurate.
2:22:37
I mean, >> there's 99% chance >> the steelman here is like that is how I feel about my lawyer and my like I love I love the people I work with, but like humans do hallucinate.
2:22:46
And so if you give me something that's doesn't even need to be superhuman, something that's going to just sit alongside that person, I'm pretty happy to buy.
2:22:52
But I understand that I would buy more if I could be more sure.
2:22:56
And so I understand that.
2:22:58
What I'm what I'm confused about is like uh how are you thinking about building new foundation model, new LLM, new AI technology that that bakes formalization into the actual workflow versus like you know we were talking to Dr.
2:23:14
Dr. card from Palunteer last week about you know the idea of like you've built this ontology you have your ground truth you have your system of record and then you can have the LLM hallucinate all over the place because at a certain
2:23:27
point they're going to run into the ontology and that's going to be the formal the formalization process like is there something is it is an extension of that two systems running side by side checking each other or is it more of an entirely new system. >> Uh so it would be a couple of systems uh
2:23:45
>> Uh so it would be a couple of systems uh interlocking with each other.
2:23:49
>> Uh we actually have four layers in the stack.
2:23:52
I can go through an example of how it happens.
2:23:54
If you end up asking >> whether a particular transaction is taxable in Illinois. >> Yeah.
2:24:01
>> You have uh the first thing which we do is actually a fair bit offline for tax specifically.
2:24:06
We are formalizing the US tax code. >> Mhm.
2:24:09
So we are taking a very formal domain but expressed in English today and we are converting that into lean which is a proving language.
2:24:16
So when we so uh when we do that we get it ratified by experts.
2:24:22
So our work is deeply involved in formalizing specified domains and ensuring that we get it ratified by the experts themselves too.
2:24:31
>> In that place we we do use a lot of LLMs because the current knowledge base is already not formalized.
2:24:35
it is present in English and once we have the whole knowledge base whenever a question pops up we convert it into a series of constraints.
2:24:45
You would have noticed Venod also talking about how humans are not great at specifications.
2:24:48
So we ensure that we ask the right set of questions so that ultimately we can give you a reliable answer.
2:24:56
>> Post that we have a solver and prover working in tandem to ultimately give you an answer along with a proof of correctness. Mhm.
2:25:04
>> So the key difference between an ontology based approach and a lean based approach is that a lean-based approach is intrinsically verifiable.
2:25:10
Your answer comes with a proof of correctness.
2:25:13
The proof is something which a mathematician trusts.
2:25:17
That's supposed to give you more uh comfort in a way.
2:25:19
And like you suggested uh you're right in saying that even like best humans might make mistakes.
2:25:27
mistakes. uh but if you actually have a mathematical proof encoded in lean you can choose to trust it and that's the frontier which we are aiming for and it involves like fair bit of work between formal verification experts who are a deep part of who we are at labs
2:25:44
>> interesting >> working along with CPAs frontier CPAs frontier lawyers and frontier doctors >> yeah uh how are are models getting better at interacting with lean I've seen a lot of progress in frontier math and the different really hard problems are getting solved. Sometimes they don't
2:26:02
Sometimes they don't use lean and that's sort of a a flex.
2:26:06
Sometimes they do use lean and they perform even better then.
2:26:08
Uh but is that getting just table stakes now?
2:26:11
Because I think uh your average white collar worker doesn't know how to interact with lean. >> Absolutely.
2:26:20
Uh so the models are getting better at proving with lean.
2:26:24
>> Uh but we are building more foundational models along the lines.
2:26:26
of formalization in the first place.
2:26:28
If you say something in English, how do you convert that into a construct in lean?
2:26:32
That's where our focus is on.
2:26:35
>> Uh if you think about it, lean is a system which has been around for the last 12 years.
2:26:40
>> It has taken humanity around 9 years to formalize a significant portion of math.
2:26:44
M >> we are taking that same technology and we are we are trying to blitz through tax, legal and healthare where we are identifying what is the core knowledge base and then we are converting it into lean uh where you can reason on top of it provably. >> That's very cool.
2:27:01
>> What were you doing before this?
2:27:03
>> Uh I was a machine learning engineer at Google primarily focusing on Google maps.
2:27:08
Interestingly, I draw parallels over there because uh I was focusing on getting addresses and phone numbers accurately in Google Maps.
2:27:15
>> If you think about it, Google Maps is like a very very messy real world domain.
2:27:21
>> Uh and we have been able to make it trustable.
2:27:22
A lot of people do trust uh Google Maps to a very large extent.
2:27:27
>> Uh I intend to bring the same level of rigor to multiple other domains along the way. >> Very cool. Well, congratulations.
2:27:32
Thank you so much for taking the time to come on the show. Great to meet you.
2:27:36
>> We'll talk to you soon. >> Have a good one. Thank you so much. >> Cheers.
2:27:39
>> Let me tell you about Codeex.
2:27:39
Codeex is a powerful work for work workspace for getting work done with AI agents.
2:27:45
Whether you're writing code, analyzing data, creating content, or automating business workflows, Codeex helps you move projects forward from start to finish. Uh, that's right.
2:27:55
>> So, there's a bitter feud going on.
2:27:57
>> There was a there was a good uh comment over on the X chat from Steve.
2:28:03
>> He said, we're on the topic of celebrity brands.
2:28:05
He said, "Flip it on its head, Jordy.
2:28:07
Brands by average Joe's like a painter or the owner of a lawn care company instead of celebrities."
2:28:13
I think there could actually be something here.
2:28:17
I think he's joking, but imagine as a brand, you're trying to differentiate.
2:28:21
You hire an actor that's just playing the role of an average Joe.
2:28:25
>> Hey, this is >> average Joe Electrolytes, right?
2:28:30
>> I'm not a I'm not a I'm not a professional athlete. I'm just a normal guy. Yeah.
2:28:33
I'm just painting houses, right?
2:28:36
I'm just painting houses.
2:28:36
I get a little dehydrated. I take my average Joe. >> It's relatable. >> I like it. I like it.
2:28:43
>> The average Joe brand meta Steve is calling.
2:28:47
>> I I enjoy it as an idea. I enjoy it as a joke.
2:28:49
I do think that there is a version of this, which is if you walk through an arrow or Whole Foods and you pick up random CPG products and you look on the back, there's often what's called like the love language or whatever.
2:28:59
So, it's like a couple paragraphs of text.
2:29:01
Uh, good example is uh uh is it dots pretzels or there's some sort of like pretzel twist that's very popular right now? It's a little spicy.
2:29:11
Uh, it has the the name of the founder and it and it tells the story of like an average Joe.
2:29:18
The Buzz Ball for example, that is an average Joe brand.
2:29:21
It was not a celebrity brand. >> Disagree. >> Explain.
2:29:25
>> There's nothing average about that founder.
2:29:26
one of the most elite entrepreneurs >> after she started the company.
2:29:31
>> I'm going to make the case that she was always elite. >> Always good.
2:29:34
>> The buzz the buzz ball >> woke it up. >> Maybe. Maybe.
2:29:37
But there are a lot of brands that their whole origin story is like, I was just a normal person.
2:29:42
I had a problem for myself.
2:29:45
Uh I was going for, you know, uh olive oil and I smashed the bottle so I made a squeezy bottle.
2:29:53
bottle. Like there's all these stories that aren't celebrity brands at least at the start and they have you know rags to riches stories that are told sort of quietly on the back of the packaging and then slowly over time eventually Unilever buys the company and takes that off and
2:30:08
>> yeah I just like thinking of it as like a from the marketing lens a brand that is like >> doesn't have this like heartwarming story but is just like yeah we just hired a guy to be like >> just be the guy just be the average Joe just be an elite snowboarder like Nema starts Salt and Stone. You know, he
2:30:24
You know, he wasn't a celebrity before, although he was a fantastic snowboarder.
2:30:30
>> Before we bring in our next guest, Marky, let's put pull up this video.
2:30:36
Somebody put up fake tech ads. What's going on here?
2:30:40
>> I don't I love ads, but I hate fraud.
2:30:43
So, I don't know how to feel about this.
2:30:45
This is going to be peculiar.
2:30:48
>> I don't think there's anything wrong here >> with fake tech ads. That's real ad space.
2:30:52
That could have been sold.
2:30:52
Jordy, you're destroying shareholder value by running fake ads. We would never do that.
2:30:59
We did do that early in the days.
2:31:01
If you go back through the archive, Dr. Squatch had that feel. Oh, Vince Vaughn.
2:31:07
Vince Vaughn running a uh >> um we have the video here. >> Okay. No, no, no.
2:31:17
I know exactly where they're going with this. We put the Q and Q1777. This is hilarious.
2:31:27
What if Texas was upside down?
2:31:30
>> I mean, this is this is actually how this is how actually how normal people view.
2:31:36
>> What if the Rizzler was purple? >> Wireflow. >> You pay it. >> We pay you. You pay us. We pay you. Wireflow. >> Wow. What does this say? Dennis can tell you.
2:31:52
>> Dennis can tell you >> brain.
2:31:54
>> I like Zip Link is now.
2:31:56
>> They put up a lot of these.
2:31:56
Wow, they really went hard. Who is behind this?
2:32:00
This has to be a tech insider or someone making >> Zip Link is now frogal.
2:32:02
The cloud-based online safety you know and love now in the palm of your hand.
2:32:09
>> They put up so many of these.
2:32:09
This is such a great stunt.
2:32:11
Is this a if this is a uh if this is a launching campaign, a launch campaign for a tech brand marketing firm, it's genius.
2:32:19
And whoever's behind this should actually be calling every tech company and saying, "What?
2:32:25
You want your ads to not look like these?
2:32:27
We know how to make good ads.
2:32:29
What if forks were spoons? What if cutlery?"
2:32:33
>> People have been asking that question recently. Anyway, very fun.
2:32:34
Uh let me tell you about Cisco. Where's Cisco?
2:32:41
Cisco critical infrastructure for the AI era.
2:32:44
Unlock seamless real-time experiences and new value with Cisco.
2:32:47
Our next guest is Marie Wagner from Poetic.
2:32:49
She's the founder and CEO future.
2:32:53
>> She's a prolific mafia player.
2:32:53
What's your favorite role in Mafia? >> Ooh, mafia.
2:33:00
I think who doesn't love to be fascist in Mafia?
2:33:02
I you know, you always kind of hope you're getting the fascist card, but uh what's the fafia?
2:33:05
I thought it was just mafia towns person. >> Oh, sorry. Secret Hitler.
2:33:12
>> Oh, you're the secret Hitler. Okay. >> Yes. Yes. Um, uh, yeah.
2:33:14
I mean, mafia, you obvious. It's always power.
2:33:19
It's a It's less empowering to be the town's person because you feel like you're just going to get killed. You have no superpowers.
2:33:25
But running or sheriff, that can be fun, but then if you get killed off early, it's the end.
2:33:30
Uh, >> I'm just thinking of it because the the latest episode of Mafia episode two from Founders Fund dropped and you're backed by Founders Fund.
2:33:38
But since this is your first time on the show, why don't you introduce yourself?
2:33:41
Tell us how you wound up in a position to raise money from Founders Fund. >> Cool. Yeah.
2:33:45
So, I'm the CEO of and wait. Okay. Um, yeah.
2:33:49
So, I'm the CEO and founder of Poetic. Yeah.
2:33:51
And Poetic's an AI system that will learn and execute super complicated processes in some of the biggest companies in the world at over 99% accuracy, which is which is quite hard. >> Yeah.
2:34:04
So what is the process for uh obtaining and documenting the workflow?
2:34:09
This feels like you could do something very database driven.
2:34:11
You could build an ontology.
2:34:14
You could uh just wait and maybe the models get better and stop hallucinating.
2:34:19
There's been debates over the various approaches like why what did you pick and why?
2:34:23
Wait, before we before we get into that, like talk more about >> uh the kinds of workflows where AC like >> you know we we caught up a few weeks ago and and you were saying like some of the customers that you're talking to even if even if somebody came through and they were like our AI system is 98.
2:34:40
5% accurate that that would actually create you know hundreds or thousands of hours of like issues in an organization and it would actually like rolling that kind system out would have sort of like negative enterprise value.
2:34:58
>> That's just a pitch problem.
2:34:58
If you're pitching a software like that, you just got to tell someone our system has 9,000 basis points of accuracy. This is the goal. >> Put in basis points.
2:35:08
>> Put it in basis points.
2:35:08
Also, >> anyways, talk talk more about the problem before we talk about the solution. >> Yeah.
2:35:13
>> Yeah. So I think the problem is you know one of the things you've seen is you know obviously AI is incredible at writing code um and has really crushed that but you know a lot of the main processes that are at the heart of these giant businesses have remained pretty untouched by AI um and the reason why is
2:35:32
that the rules that govern them you know the 10,000 secret rules they live in people's heads right um and these rules need to be followed every single time you know we work on things like anti-moneyaundering and underwriting and fraud investigations where every single step matters. And to actually get
2:35:45
And to actually get from where we are now to those running on machines, you need two things.
2:35:50
One is a system that can learn the process.
2:35:53
Not the 100 pages of written down stuff, but the 10,000 secret rules that live in people's heads that they've never written down before. So that's part one.
2:36:05
And then you have to be able to run them with many nines of accuracy.
2:36:07
Not, you know, I was talking with one CEO.
2:36:09
He's like, you know, 80% on an eval is great.
2:36:13
Um, but in underwriting, it's unusable.
2:36:15
>> And I think that's the view for a lot of these B processes that are at the heart of these like really big and old businesses.
2:36:20
So, you got to do two things.
2:36:21
You got a system that can learn all the rules and then run them with nines.
2:36:25
And if you have that, you can get from where we are now to this version of the future that I think everybody's really excited about.
2:36:30
But you need to build that bridge.
2:36:34
>> Not everyone's excited about it.
2:36:34
AI has a very low approval rate but but tech leaders are certainly uh excited about it but uh so why have you started with such big companies like I feel like you've been you've been you know at the heart of the startup ecosystem for so long if you came on here and said oh we have so many you know my friends companies on board these are uh the named customers SoFi Chime AIG like these are large institutions at this point.
2:37:05
Um why why start at the top? Feels harder.
2:37:10
>> Yeah, I think our our view was um when you're when you're building something like this, if you build something that's less powerful, it's hard to sometimes to make it more powerful.
2:37:18
Um like a lot of these drag and drop tools, they were simpler, they were easier, they could do simpler things faster, but they just couldn't handle super complex things.
2:37:24
Um and so you kind of get kneecapped in terms of what you can represent and build.
2:37:30
And the view is, hey, AI is going to be doing a lot of this like writing of the software.
2:37:35
Um you'd be in a better place optimizing for the power of the thing.
2:37:39
Can it express and run these fivehour processes um that require those nines and then everything else becomes really easy actually >> and I don't think that people are going to be having a platform for the easy stuff and the hard stuff.
2:37:53
I think they're just going to be running all of their processes in one spot.
2:37:57
>> Uh I mean you mentioned like tens of thousands of hours of information stuck in people's heads.
2:38:02
Like what is data collection?
2:38:04
What does data collection look like?
2:38:05
Conversations like it feels like yeah like is it the McKenzie model or is it the the the you know self-s serve software model like where do you want to sit? >> Yeah.
2:38:17
I mean what's interesting is um it's evolved over time but the place that it has ended up going is looking closer to data labeling.
2:38:22
So what happens is a big com company will say here's my biggest process there 100 pages of documentation and you know that that's only 20% of it.
2:38:30
So the question is like how do you get that other 80?
2:38:34
>> Well, what we'll do is we'll take that operating procedure.
2:38:35
We'll generate the AI operating procedure.
2:38:37
It's written in step-by-step English.
2:38:41
>> What our system does is turn that into code under the hood and you run it.
2:38:44
>> And when you run it and you put in front of those experts, they have a ton of opinions.
2:38:47
They have a lot of feedback to say, you forgot that the threshold's actually a million, not 10,000. Right?
2:38:52
All these little things >> which get merged back into that document.
2:38:55
And you're sort of doing that more and more automatically.
2:38:57
Um so that it looks almost like training a program or something like that um rather than sort of uh like long calls or or uh you know process mining in like the the normal sense and so more and more it looks like people giving feedback into the system directly and that updates the rules that are written in it.
2:39:15
>> And so you're using AI models to generate that code that is then deterministic which gets you the reliability that these companies need. Correct. >> Yep. Yep. Exactly.
2:39:25
So the, you know, the source of truth is that AI operating procedure. It's English. But what are those? Turn it into code.
2:39:32
So when it runs, if the if the world's the same as yesterday, it's just going to run as code. Great. Nothing to see here.
2:39:36
But if something changes, like the column name changes from month to month or the save button moves.
2:39:43
Only then will AI go in, repair it, look at what the English goal was, and then update it.
2:39:48
Um, and so if you do that, you can get the best of, hey, it's code.
2:39:54
It's precise when it runs, but if something changes, instead of breaking, it will actually just kick back and recover.
2:40:00
And that's important because a lot of this work code couldn't do alone and agents couldn't do alone.
2:40:07
So code is very static, right? It's like very brittle.
2:40:10
So even one small date change or something could break the whole thing.
2:40:14
And so that's one side of it.
2:40:16
Agents, on the other hand, are pretty improvisational and they think step by step.
2:40:20
And you know when you're figuring out what to do as you go eventually you're going to make a mistake.
2:40:24
This is something that's kind of in the middle where it's code if things are the same.
2:40:27
AI is there to test heal recover if things are different and that's how you can kind of get those nines.
2:40:33
>> So are you hiring forward deployed engineers like what is the role of engineering in your organization at this point? >> Yeah.
2:40:40
So we do we hire tons of forward deployed engineers from all the best spots.
2:40:45
spots. um you know whether it's you know places like Palunteer or you know even like retool or scale and all these other sort of new places that have had their own versions of forward deployed um and yeah I think it's interesting it it is over time looked less like extremely just only focusing on engineering like hard engineering it requires being able to change how people operate right yeah I can write the best code now but even
2:41:08
if you have the most incredible piece of software you still have to change how the business organizes around it and So the people who understand like business and can think about hey like what is the best possible fraud process going to
2:41:22
look like and you know what how should we reorganize the business around this new kind of thing >> and they understand AI are the ones that are just like totally crushing it. >> Yeah that feels like an entirely new
2:41:29
>> Yeah that feels like an entirely new skill set like much more people driven but also like forward thinking in technology.
2:41:35
technology. I don't know it's a it's a it's an interesting new >> how has enterprise sentiment around just AI changed during the course of building the company right because like every you know it is I would say it's a roller coaster in some ways it's like going up into the right right there's generally more excitement about the potential but at the same time there's these sort of like period troughs of uh of disillusionment uh and you guys are
2:42:03
coming out out of stealth at a time when again I think companies are more excited than ever but at the same time sort of like understand the overall shortcomings and and we had uh Karp on the show uh last week and you know he was just
2:42:18
saying like a lot of these deploy codes coming in they're just trying to like they're they're they're trying to deploy their their goal is to deploy AI but they don't fully understand all of this like business process under under the hood. uh and he makes some good points.
2:42:33
uh and he makes some good points.
2:42:37
>> Yeah, I think sentiment has evolved quite a bit.
2:42:39
Um you know, I think it was extremely excited earlier in the year and then you know around the board meeting times more and more CEOs would come to me and say hey like you know >> how do I get ROI here?
2:42:50
Um and you know is this what your other customers are seeing um uh in these domains?
2:42:55
And so I think now people have realized like you cannot just throw an agent at a problem and expect uh to see the result that you want, right?
2:43:04
Um for an agent to to do a useful bit of work, it needs to learn how to do that work and run it like quite accurately.
2:43:10
Um and that knowledge transfer between how's the work done today and getting it written down enough to where AI can run it. Um it's just hard.
2:43:19
Um and it requires, you know, we sort of jokingly call it the great migration internally.
2:43:23
like you have to go and migrate these like tremendous amounts of rules into something that AI can touch and improve and evolve.
2:43:28
Um, and if AI can't touch it, it's not going to be able to help it.
2:43:33
>> Um, and so, you know, I think deployment is really important because in Yeah, until that transition happens, >> it's going to be hard to just throw tokens to to see better outcomes.
2:43:45
>> Uh, give us some backstory on what you were doing before, how you wound up here. Yeah.
2:43:51
So, uh I got my start I got extremely into AI in middle school after reading too much sci-fi. You know, a lot of Dune. A lot of >> Dune. >> A lot of Dune. Mainly Dune.
2:44:03
>> Is that the message in Dune?
2:44:03
I thought Dune was the but right I guess. Yeah. Yeah.
2:44:08
You want >> Well, I just felt like it was going to be really important during my life.
2:44:12
be really important during my life. um you know and you know you read and you watch sci-fi and you know the biggest difference between the future and today it's like mostly machines thinking like that is sort of the main ad sort of BC moment um
2:44:27
>> in like hard sci-fi in soft sci-fi it's like time travel and faster than light speed and like aliens but I take moon is made of cheese >> that's the type of sci-fi I'm trying to read and that's the white pill uh so then uh walk me through the consulting work that you did and how this like evolved out of that. Was there like a
2:44:47
Was there like a clear moment where you're like I'm stopping that and starting a company or is this an evolution or a a change of structure? >> Yeah.
2:44:57
So I was initially got my start in research.
2:44:59
So I was I was at Stanford working in AI like Whimo and you know Google.
2:45:02
Um and then you know one day I sort of realized that you know all the things I had built like you know didn't didn't matter too much.
2:45:08
Um and that was Friday and I ended up dropping out on Monday.
2:45:12
Um, and the idea was like, hey, you know, we've had software for decades.
2:45:16
What are people still doing and why?
2:45:18
And I felt like I didn't really understand anything about how the world works.
2:45:21
So, to your point, started consulting and the idea was like, let's just go into >> some of the oldest companies around and understand like >> what has software still not touched and and what happened there.
2:45:31
Um, and when you do that, uh, and you start doing some of the work yourself and like going into, you know, I went to North Carolina, you know, you're watching people who have done this stuff for like decades.
2:45:40
um you realize that a huge amount of the work is really just operating procedures, documents >> and people are just following them >> and that this class of work everywhere and you know whether it's underwriting and claims and insurance or onboarding customers and fraud and banks, it's just still sort of done by by people and software hasn't been able to go there.
2:46:02
The reason why is because the second you write all this code to do that process, something will change, right?
2:46:07
a button will move, a column will change.
2:46:08
Um, or maybe even the process changes.
2:46:11
You want yearly instead of monthly transactions.
2:46:14
Um, and when you, you know, automate and see that happen enough, you you realize like, hey, there's a missing kind of material here that can flex but still have guarantees.
2:46:21
And I really just waited for the models to get better.
2:46:23
Um, I knew some of the researchers and said, hey, the models of today are not it.
2:46:26
Uh, please let me know when they get good enough to stuff. And I truly just waited. So, >> that's great.
2:46:34
Well, uh, congratulations on the round. How much did you raise? I want to hit the gong. >> Oh, yeah.
2:46:40
So, you raised $50 million. >> Congratulations. And thank you.
2:46:46
>> Who uh who besides FF is in? >> Yeah.
2:46:49
Kleiner Perkins, >> First Harmonic, uh Genius Ventures. >> Cool.
2:46:54
>> All all participated in the round.
2:46:56
>> Round of applause for Genius. >> Ben and Adam. >> Love. >> It's great news.
2:47:02
>> Well, have a great rest of your day to have you on the show.
2:47:05
Congratulations on coming out into uh into the world and uh I'm sure you'll be back on very soon.
2:47:12
>> We'll talk to you soon, Mark. >> Thank you. Appreciate it. See you all. >> Cheers. >> Goodbye. Um let's see. Canva AI.
2:47:18
We got to talk about the dog.
2:47:23
The dog walker and the dentist.
2:47:25
The bitter feud between a dog walker and a dentist over who owns the beach.
2:47:30
A lakefront owner likened his neighbor's shoreline walks to a home invasion in a dispute that could be headed for the Wisconsin Supreme Court. Who you got, Jordy?
2:47:39
Knowing nothing about this story, are you going dentist or dog walker? >> Uh, >> Pure Vibes. >> Pure Vibes. >> Flip a coin. Which one? Who you got? You don't know. It's your job to know.
2:47:56
>> Is the dentist a day trader?
2:47:56
So, it's your job to pick someone wildly and then defend it throughout this article.
2:48:00
A recent trial in Shorewood, Wisconsin had all the trappings of a minor legal dispute, a disgruntled neighbor.
2:48:06
A defendant representing himself who called his own father as a character witness. You love that?
2:48:11
Hey, Dad, tell him I'm a good person.
2:48:14
13undu $313 were at stake. $313.
2:48:24
That's 31,300 cents if you're paying attention.
2:48:26
If it uh but an if academic and devoted dog walker Paul Floreshime gets his way, the case will go all the way up to the Supreme Court, the Wisconsin Supreme Court that is, and reshape the contours of shoreline access to one of the Great Lakes.
2:48:41
It started when Flores started walking his two dogs past the Lake Michigan property of dentist Domagala, locally known for the time he spends in a tiki style boat house and deck that doubles as a surveillance post.
2:48:56
From there, Domagala monitors traffic and sets off alarms to scare walkers, swimmers, and kayakers away. He's got an air horn. >> He's got an air horn.
2:49:04
He sets it off when somebody comes by. It's crazy.
2:49:08
Floorshine repeatedly ignored signs outside of the dentist's house that said private property behind the sign only water access behind this beyond this point.
2:49:14
Domala kept calling the cops and the village eventually issued a trespassing citation.
2:49:17
Rather than pay the fine and walk away, floorime dug in at stake is what right people in Wisconsin have to take a shoreline stroll.
2:49:27
It's high stakes stuff and we'll be following this case closely.
2:49:30
Of course, >> one of the most important stories in technology, >> but more importantly, we have Brett Taylor from Sierra here in the waiting room.
2:49:39
Welcome to the show, Brett. How you doing? >> Doing great. How are you?
2:49:42
>> You going dentist or dog walker?
2:49:45
>> We won't make you take a side. That's a tough question.
2:49:47
>> That's a tough question.
2:49:48
>> We're going to ask you the real tough questions. How's business going? >> Business is great.
2:49:52
Uh we just crossed 200 million in ARR. Uh, we were >> great.
2:50:01
>> I was hoping to get a gong out of that.
2:50:03
And >> revenue revenue gongs are always the best gong. >> Yeah. Yeah.
2:50:07
I mean, fundraising is easy. >> Yeah.
2:50:10
Uh, and today we announced we're certified Fed Ramp, which really simply put means now um, federal government agencies have access to Sierra, which I think is really exciting.
2:50:19
If you look at, you know, the conundrum of the federal government, it's like we want greater services for our citizens, but we have a really big national debt.
2:50:26
I think AI is going to be a huge benefit to, you know, programs whether it's, you know, getting a passport or Medicare.
2:50:35
Um, so this is the first sort of the door opening so we can serve these agencies, which I'm really excited about. >> Interesting.
2:50:41
I don't think of like if I you're breaking my brain because I don't think of like getting a driver's license to the DMV as like a customer service interaction, but is that where we're going where more and more of these like uh form based like an address change on like an ID like something like that like you know if if Yeah, there's there's a lot of these.
2:51:05
So yeah, I mean >> there is if you think about all these government services, you've got Medicare, you've got Medicaid, you've got the Department of Veterans Affairs, you've got uh you know, immigration like you know the passport agency, all of them run large call centers, but more importantly people depend on them.
2:51:22
You know people depend on them.
2:51:24
If you have you know someone on Medicare, if you have a vet who depends on Department of Veterans Affairs for their healthcare, it's a really big both pay provider.
2:51:32
If you look at, you know, what we do for, you know, folks like Sigma and Blue Cross Blue Shield and you squint, well, it's almost that entire stack in a in a federal agency and, you know, for good reason.
2:51:44
There's just a very high bar uh for cloud technologies when you're serving the federal government and this is the main certification that kind of opens the door there.
2:51:51
We're already working with a few agencies, none that I can fortunately talk about today.
2:51:54
We work with FINRA, the regulator.
2:51:55
So, we we've always been really passionate about we want the best technology to be available even in regulated industries.
2:52:03
And so it's a big step for us.
2:52:03
We're really excited about it.
2:52:05
I hope hope we can contribute to the quality of life as US citizens, you know, and do so in a way that's a lot more cost effective. >> Yeah.
2:52:12
Um I'm I Yeah, it's interesting like would you would you is the is the biggest low low hanging fruit just taking some of these agencies and making them available 247.
2:52:21
I feel like one of the most frustrating things that I run into is like call from 9 to 5 and I'm like I have a job. Yeah.
2:52:32
Yeah. The other the other thing is >> yeah the thing like uh uh the difference between interacting with like a public utility versus like a waste management right like waste man like waste management right like >> you know maybe it's not perfect but like it's pretty fast like you can get stuff
2:52:50
done like it's efficient and then like the trying to do the same exact thing with like a public utility you know usually you I I I always personally estimate like okay if it would take me like you know 10 minutes with waste management it'll take me like 30 minutes with a public utility. And so like
2:53:03
And so like bringing bring that there's there's no reason that those two things can't be at par, but it is entirely it is almost entirely a uh a technology and like process problem, right? >> 100%.
2:53:18
And there's a couple factors I think go into why it's so hard to have great experience with some of these government programs.
2:53:25
uh you know first a lot of people do need to call things on the phone and it's only been the past couple years that we could digitize a phone call with AI agents.
2:53:32
So we've essentially digitized the last remaining analog channel and everything you said is true.
2:53:38
It could be 247 as importantly it could be multilingual.
2:53:40
If you imagine staffing up a big call center and you want to represent all the languages in this country from Spanish, Mandarin tool, you know, like go down the list really really expensive to do.
2:53:49
Now you can do it at scale.
2:53:51
it at scale. each new language the marginal cost is effectively zero and then the other part of it is for good reason as I said there's just a a really high compliance bar for technologies to be available and as a consequence if you look at you know what waste management or just pick any private company they
2:54:07
can go off and get the best of breed solution for any given technology problem they have that is not something all federal agencies can do and so what I'm really proud of is that Sierra is the best of breed solution in customer experience we power everyone from you know Singtel to rocket mortgage on our
2:54:22
platform and now because we have the certification federal agencies have access to the best of breed solution in this space as well and so I'm really hopeful uh I think AI if you look at some of the parts of the economy the governments both state and federal the healthcare industry which has gotten progressively less productive over the
2:54:41
past decade it's one of the few parts of the economy that has I think AI is incredibly important because these are the parts of the economy that need productivity you know we actually want to have great service as citizens and reduce our spending, AI is the answer. So, we're really passionate about it.
2:54:56
So, we're really passionate about it.
2:54:58
We're patriots here and we love our technology and so I'm just excited that we can apply it in a in a in a great way.
2:55:04
>> Talk about uh uh pilot programs and conversion to real customer.
2:55:08
Is that timeline a KPI for you?
2:55:12
Is that something that you monitor very closely?
2:55:16
I have to imagine it's speeding up, but uh whenever you shake the hand of a CEO and they say, "Let's do this," or someone in the federal government perhaps uh an agency lead uh and they say, "I'm ready.
2:55:26
I'm I'm bought in on the vision. The demo looked good."
2:55:28
But then there's obviously an implementation time.
2:55:33
What does that look like now?
2:55:33
What did it look like a year ago?
2:55:36
Where's it going in the future?
2:55:39
>> We track it and we measure it in days, not weeks.
2:55:41
And so, you know, some of our favorite, like Nordstrom, you know, from, you know, concept to 100% of their phone calls was 35 days. >> That's good. Said 3,000 days is good.
2:55:54
And then we're back to, oh, you're gaming the stats.
2:55:57
>> What's neat about that, too, is we started out at 1% of their phone calls and I think four weeks and went to 100% in just a week.
2:56:04
And that's because of the Nordstrom team. Yeah.
2:56:05
Um, you know, actually in Fiserve, which is an amazing, you know, financial technology company for banks, just broke the record.
2:56:13
Um, so I don't know if I'm at liberty to show the number of days, but they have an amazing president named Dia.
2:56:18
She was just top down like we're going to do this and broke the record.
2:56:22
So, we have a leaderboard internally about how quickly when a customer talks to us, can we go live and start delivering value?
2:56:29
For what it's worth, I think it's really relevant in this world.
2:56:32
You can Financial Times, Wall Street Journal, it's all, you know, token maxing.
2:56:35
What value are we getting from the tokens?
2:56:37
Like it is kind of in the headlines right now.
2:56:39
We don't really have that issue with our clients because like our thing is like let's start answering the phone as quickly as possible.
2:56:46
Start driving those operating expense savings as quickly as possible, but as importantly, show that you can actually improve sales, improve seat at the same time.
2:56:54
our, you know, as you know, we've been really pioneering this idea of outcomesbased pricing with the idea being you only pay Sierra when we successfully resolve a call or successfully make a sale.
2:57:05
It's it sounds like a small business model thing, but it's big because we don't get paid until it's live.
2:57:11
it's live. So it really aligns like all the incentives like we don't care about the sales process we care about the go live process and I think it really changes I would say almost the social dynamics between vendor and customer where we really become a partner because we have the same incentives as our
2:57:28
clients and I think it's a really big shift and you know how much I'm an advocate for it but I actually think a lot of the you know token maxing issues would go away if more companies were just focused on outcomes were >> how are you tracking the popularity of the message talk to a human >> or the state or the statement. >> No, no, no. When I pick up the phone, I >> No, no, no.
2:57:47
When I pick up the phone, I say talk to an agent.
2:57:49
Agent and then they put me on the phone with an AI agent. It's great. >> Yeah. No, thank you for that. You're on my side. No.
2:57:56
So, we actually have a name for it.
2:57:58
We call it greeting acceptance.
2:58:00
Uh, which is basically what percentage of the time do people get past the greeting and actually give the AI agent a shot? It's going up. people in general.
2:58:08
I think the problem was we had 10 15 years of bots that were really bad and so just you're just like paying down this debt because if you were talking to an AI more than 3 years ago, it sucked.
2:58:20
Like let's just be blunt about it.
2:58:20
Now you can talk to an AI and it's conversational. It's multilingual.
2:58:23
It has access to systems.
2:58:25
I mean the AI agents of today are just like a completely it's like horse and carriage versus flying car, right?
2:58:30
They're just very different technologies.
2:58:32
The problem is in the first 5 seconds of a call, you're wondering, is this one of the crappy ones or is this like one of the good ones?
2:58:39
And so we have a lot of techniques on our platform to help people with that.
2:58:42
I think the real key is making sure the first greeting is personalized, is engaging, is empathetic.
2:58:49
Uh, and I think the trend is in the right direction.
2:58:51
I think over time, I mean, I don't know if you have this experience, if I have to call a restaurant to make a reservation, I'm like, I'm going to a different place.
2:58:59
Like, if you don't have Open Table or Resi or whatever, like I'm out.
2:59:00
I think we're going to get to the point where people demand an AI because they're not going to want to wait on hold, you know.
2:59:07
And what's also interesting is these AI agents once they're properly configured are faster for me as a consumer.
2:59:11
They have access to my information.
2:59:13
They can get to things done quickly.
2:59:15
We're not 100% there and as I said, we're just paying down the debt of bad tech that we've had for the past 10 years.
2:59:21
Uh but I'm hopeful in three or four years it will be the norm.
2:59:24
When did it become popular to to get the uh like there's a 30 minute wait like press one if you want a call back because I felt like that wasn't a thing when I was a kid but now that feels like a pretty meaningful innovation at some point in the last 10 years where it actually is really convenient to not be like okay like I'm I'm just going to >> some guy has a patent on that and has >> Yeah, I'm sure.
2:59:48
But like that was like the last great innovation in uh customer service. Yeah, it was.
2:59:54
>> Well, I think it's going in a really good direction.
2:59:55
I mean, first I think, you know, not having to wait on hold.
2:59:57
We should never have to wait on hold again.
3:00:00
Like, that is going to be a thing of the past.
3:00:02
>> What if the GPUs are on fire?
3:00:02
It's like we'll call you back and the GPUs call like cool off.
3:00:08
We're load balancing right now. >> Please wait on hold. Yeah, >> maybe. Imagine that world.
3:00:12
That is a bizarre outcome.
3:00:14
I mean, on that on that note, are are there any are there any like uh how close are we to thinking about like A6 for voice models or we were talking to Matthew Prince from Cloudflare about uh edge computing for voice models that feels more relevant than yes, if it's you're going to cook for an hour on some deep research project, go do it in the on the NVL72.
3:00:38
But um how are you actually at the point where you're starting to think about that optimization or is that more a few years away?
3:00:43
I think it's probably not a few years away, but it's definitely in the future. But I'm excited for it.
3:00:47
I actually agree with your intuition that, you know, we're entering a world where first the frontier models are actually separating in my opinion.
3:00:56
GPT55, the recent releases from Anthropic.
3:00:58
recent releases from Anthropic. I mean you can just see it like the gap between the best models like GPT55 and the the others the open source models is growing not shrinking which is probably not something we would have predicted a year ago or at least like certainly not what I saw in X and and the press but we've
3:01:16
reached sufficient intelligence for a lot of different domains like you don't need you know a multi-billion dollar multi-billion parameter count model to do a simple classification or you know just do basic you know transcription so as a consequence I'm hopeful that over the next few years we'll end up with a
3:01:32
constellation of models with really purpose-built ones and as you said just given the power of the hardware in our pockets right now I think edge will be a part of it there's just no doubt and it and it just stands to reason there might be creative ways to you know maybe convert what you're doing into tokens
3:01:47
you know so you have like lower bandwidth connection if you're doing voice could there be opportunities there there also could be some privacy benefits to that I you know I have a strong intuition though no knowledge but just intuition That's got to be something Apple's thinking about just given their their posture on privacy. So
3:02:01
So I'm very bullish on the frontier models like you know OpenAI and Enthropic have you know really shown the the strength of their research groups but not at the expense of all the specialized models.
3:02:12
I think we're in a world where you know just as uh we become more sophisticated in deploying AI we're going to have a constellation of models with just different capabilities.
3:02:21
So I'm excited for it but it's not now.
3:02:23
It's it's in the the short-term future. >> Yeah. Yeah sense.
3:02:26
Uh, did you guys have to build in guard rails to prevent people from access every every like month there's a post that goes viral of somebody that's like figured out like some end point that they can hit to get like >> Chipotle >> like a frontier model >> code somebody figured out how to wire up well they took some sort of like uh open code fork of like codeex agent coding model but it's wired up to Chipotle's customer service in the back end.
3:02:52
This is a huge risk because you could run up a big token bill that way.
3:02:55
Is this is this actually tough?
3:02:59
>> We have guard rails on it and you know it's it's just like any other thing.
3:03:03
It's you know sort of you're always in a race the good guy the white hat versus the black hat to you know handle it you know in practice.
3:03:09
Um I think it's more like digital graffiti.
3:03:11
I think nowadays like in the early days you'd see a post online of a chatbot saying something goofy like oh my gosh.
3:03:16
Now people are like yeah you you fooled with it you know like come on you know.
3:03:21
Um so I think first I think it's it's probably um jailbreaking I think is a really interesting area of AI security but probably more around the areas of cyber security and bio and things like that.
3:03:32
I think this type of thing you it's why people work with places like Sierra to have good guard rails.
3:03:38
I also think like you know social media sort of become numb to it at this point you know it's like oh great you got a chatbot to say something silly again you know.
3:03:44
Um I I think the bigger issue will be just as these frontier models become more capable the guard rails around using them need to be more and more effective.
3:03:54
And I think you know as you think about kind of the mission of OpenAI ensuring AGIS benefits humanity you really need to make sure that those guardrails are effective against jailbreaking which is not you you you know it's not like possible to make perfect by the way but just because these models are sort of capable under the hood which is why there's been such a interesting you know topic around cyber security and others.
3:04:16
So that's probably the area I think more about though you know the digital graffiti is going to continue to exist as well.
3:04:21
>> Yeah that's a good good term.
3:04:21
I have you ever heard of 002 cents? >> No. Tell me.
3:04:27
>> Famous customer service nightmare call where someone called Verizon and they were getting build at 2 cents and they were and the and the actual rate was supposed to be 0. 002 cents.
3:04:39
So they were off by two orders of magnitude.
3:04:42
And they're talking to this customer service agent.
3:04:45
We'll play it on the show after we wrap, but um uh they're talking to this customer service agent and the customer service agent just actually doesn't know the difference between a decimal scent and an actual scent.
3:04:54
And it the frustration that just ensues is like turned into this massively viral video.
3:05:02
There's whole websites about it. Verizonmath. blogpot. com from 2006.
3:05:07
And it's like this whole deep dive.
3:05:10
I'm sure you've heard about all these different >> I will take a look at it.
3:05:12
The good part is a super intelligent AI probably will know >> I think it will some ground truthing some code in there.
3:05:18
There's there's opportunities here level the raw audio for this call is 27 minutes.
3:05:28
The guy went back and forth and it's hilarious.
3:05:30
There's cut downs we'll have to play but um yeah he had unlimited data plan in the US and recently crossed the border to Canada.
3:05:38
Prior to crossing the border he called customer service to find out what rates he'd be paying.
3:05:41
The data rate he was quoted was 0.
3:05:42
002 cents per kilobyte and then he got build at 0. 002 per kilobyte.
3:05:51
So it was a hundred times more and he couldn't get out of his quagmires Kofka-esque customer service interaction.
3:05:58
Anyway, thank you so much for taking the time to come on the show. It's always a pleasure. >> Always a pleasure. Incredible progress.
3:06:04
Looking forward to the next revenue gong. >> Yep. Come back anytime. >> I'll come back. See you >> accelerating. We'll talk to you soon. Goodbye. >> Oh, make a few sense.
3:06:17
>> We're gonna now play this 27 minute video.
3:06:20
>> No, we can play this one.
3:06:20
Uh, this is a this is a just a little bit of this. Yeah. Yeah. Yeah.
3:06:25
This is the one the the the three the three minute version. YouTube math fail.
3:06:29
I want to hear a little bit about this.
3:06:31
See if it holds up or if see see if I'm washed at this point being old and remembering a a 70s video. Not this one. It's the 002 cents.
3:06:42
Uh this is exciting news from Canva.
3:06:42
We can talk about it later, but uh you can turn your chat GBT images into fully editable fully editable uh Canva designs with magic layers without ever leaving the chat. Great integration there.
3:06:54
Apple intelligence has some more updates too we can run through later in the week.
3:07:02
Um, but let's play the 002 cents at least a few minutes because I want to hear if it holds usage that was done while in Canada.
3:07:10
>> Do you recognize that there there's actually >> 02? >> Yes.
3:07:14
Do you recognize there's a difference between those two numbers? >> No. And Z2? >> Yes.
3:07:24
Is there a difference between >> They're both the same if you if you look at them on paperwise.
3:07:29
>> No, they're not actually. So if you take 0.
3:07:31
002 >> cents, remember it's cents >> times 35,000 896 $71. 79. >> No, that would be 71.
3:07:44
How much should I be charged >> by by the way this is calculated? $71.
3:07:52
>> I like this guy's crashing out over 70ucks of a penny per kilobyte. >> Two10. Hold up. Hold on.
3:07:57
210 of a penny would be 0. 2 cents. You quoted me 0. 002 cents.
3:08:06
>> I think this guy also might be entirely wrong.
3:08:07
He might have just been quoted the wrong number. >> 02, sir.
3:08:11
>> 02 what >> cents per kilobyte.
3:08:14
So you just said it was 0. 2.
3:08:18
>> They quoted it back to him.
3:08:21
>> And then you also said it's 02 cents.
3:08:24
Those are two completely different numbers.
3:08:26
They're 100fold different.
3:08:28
George, hold on one second for me. Okay.
3:08:33
>> This guy so dedicated. >> This is Andrew.
3:08:35
I'm a manager on the floor. How can I help you?
3:08:38
>> Another manager involved.
3:08:39
>> Do you recognize that there's a difference between $1 and one >> coming out of the gate? >> Definitely. >> Definitely.
3:08:43
>> Do you recognize there's a difference between half a dollar and half a cent? >> Definitely.
3:08:49
>> Then do you therefore recognize there's a difference between 0. 2 and 0. 002? No.
3:09:01
>> And I I mean I'm trying to get what you're saying here, but it's just not >> And we're talking about cents, right? >> Right. 02.
3:09:06
If we multiply that by the amount of kilobyte usage that you have >> 35893 >> 35893, that comes out to what you paid. 7179 >> cents.
3:09:20
You never did the conversion from cents to dollars. I don't know. I'm not a mathematician.
3:09:27
>> I'm not a mathematician.
3:09:28
>> Your two cents, >> right? >> Times my 35,893.
3:09:34
It's is a number, but it's still in cents. >> We're not quoting 0. 2. We're quoting 02. >> Oh god. Honestly.
3:09:44
>> Well, I mean, it's obviously a difference of opinion. >> It's not opinion. Okay. Well, you know what?
3:09:48
I'm going to post this recording on my blog. Oh, >> cancel. That's fine.
3:09:55
>> That's what I'm gonna do.
3:09:55
And and and then you guys all at Verizon can learn. >> Yeah.
3:10:01
Oh, what a funny interaction. >> Incredible. >> Deep.
3:10:05
>> Well, now now that you get a PhD level mathematicians, >> somebody's got to run this on a on an AI lab.
3:10:11
I was quoted point I was coded I was coded five cents per million tokens, not five dollars per million tokens. and confuse someone.
3:10:20
Get a 100 100x discount on your on your uh bill, your token bill.
3:10:25
Uh Andre has a space something. What was it?
3:10:30
Space observation network.
3:10:32
These are groundbased satellites or groundbased telescopes that monitor space. And look at this.
3:10:36
Captured by Anderoll's network of 400 telescopes deployed around the globe.
3:10:40
The second stage of the Falcon Heavy launch of Viaat 3 F3 performing a ret a routine thrust event.
3:10:48
This produced a spiral-shaped plume effect, a nominal part of operations for a successful launch of Viosat's latest satellite.
3:10:55
I thought this is a cool image.
3:10:58
And we learned in the process that they have 400 telescopes all over the world.
3:11:05
Anyway, we can dig into all of this more later, more tomorrow.
3:11:09
We'll be back at 11:00 a. m. Pacific. Thank you for tuning in.
3:11:14
>> Leave us five stars on Apple Podcast and Spotify. >> Show everyone.
3:11:16
Have signpon or evening of your entire life >> and we'll talk to you later. >> We love you.