0:01
[Music] [Music] experts. Double blaze. infinite [Music] waves.
Search this video
0:37
I see a large IPO on the horizon.
0:42
[Music] [Music] I see multiple journalists on the horizon.
0:58
Stand [Music] [Music] by. Heat. Heat. N.
1:27
[Music] [Music] [Music] Friendly IPL in [Music] the team deathmatch.
2:08
[Music] [Music] market clearing order [Music] inbound. Hey, hey, hey.
2:41
[Music] [Music] We are surrounded by journalists. Hold your position.
3:27
We came to this [Music] world to shape our future. [Music] Triple wings. Feel the music.
4:07
[Music] [Music] [Music] music.
4:36
I see a large IPO on the horizon. Do it. Let's go. We're live.
4:44
I think I think my mic's hot already. Let's go. We're hot. You're watching TVPN.
4:52
Today is Wednesday, June 4th, 2025.
4:55
We are live from the TBPN Ultra Dome, the temple of technology, the fortress of finance, the capital of capital.
5:03
We have a great show for you today, folks.
5:05
Um, a story that we've been covering for months now is continuing to develop and is now in the pink pages of the Financial Times.
5:13
This is the rippling deal drama.
5:16
It continues and in their briefing section it says Silicon Valley spy drama, a feud between two startups over alleged corporate espionage.
5:29
Rare rare case where the the drama does not involve the media actually.
5:33
Uh a feud between two startups over alleged corporate espa corporate espionage has taken a new twist after a 12 billion dollar HR software group deal claimed that rival Ripling has had directed one of its staff to piller its assets.
5:48
So they're going back and forth and today deals firing back.
5:53
Now I heard a little bit about this story.
5:55
we kind of there were the rumor mill was was uh churning over this idea that um you know deal had something up their sleeve and they were going to fire back.
6:06
Of course that's the way these things go as the uh as as these you know battles evolve.
6:11
The interesting thing there's a couple interesting things in here is that um the the the rippling allegations resulted in a lawsuit and ultimately I think criminal charges.
6:24
Uh so uh Ripling alleged earlier this year that a staff member at Deal had been spying on behalf of Deal.
6:31
This is of course that send that watch to London moment.
6:33
Um send that watch to London.
6:35
The employee disappointed John.
6:38
That is such uh an incredible line in Silicon Valley history.
6:42
Sad that people don't use it more.
6:45
It really failed to break through. Yeah.
6:47
It's it's uh you know we were discussing this earlier about this is like the most dramatic story in of the year in Silicon Valley and yet I've talked to multiple people in Silicon Valley who have just completely missed it because it's just it's it still is just HR enterprise SAS drama.
7:03
Those people simply do not care.
7:07
They just don't they're you have to really pay them to care.
7:10
have to get some sort of like payout system going like X to be like every time you pay attention to this, you know, you're going to get a micro payment. Yeah.
7:18
I mean, they're just it just doesn't have kind of the the weight of like Elon, Doge, Tesla, humanoid robots, AI, AGI, AI Doom, right?
7:24
It's just it's just payroll after all. Forget it, Jake. It's just payroll.
7:32
Uh well anyway uh in new legal filings seen by the Financial Times deal has countered that Rippling has been actively engaged in a carefully this is a quote from that uh in a carefully co coco coordinated espionage campaign through which it infiltrated deals customer platform by fraudulent means and pilferred the company's most valuable proprietary assets.
7:55
What's interesting is that they they they're stopping short of of naming like a spy, right?
8:00
Um, and so the the the uh the actual uh approach here, Deal has sought to dismiss Ripley's initial claims of direct corporate espionage and has filed a lawsuit in Delaware, alleging that his rival is trying to impugn uh Deal's reputation and his latest filings were lodged to yesterday as an amendment to that case.
8:20
It alleges that Brett al Brett Alexander Johnson, someone I have not met, Ripley's competitive intelligence manager, po posed as a customer in access details of deals, products, and businesses business practice over the course of six months.
8:33
Now, what's what's what's interesting is that like um access details of deals, products, and businesses.
8:38
products, and businesses. So this sounds like they were asking customers for what deal was doing maybe instead of say uh instead of say going inside of deal slack like there's no allegations that they were inside of deal's corporate systems uh which is a very distinct line
8:58
to cross and so obviously the story is evolving and we've invited Alex the founder of deal on the show uh would love to get his side of it would love to get Parker on as Well, but of course our worst enemies, the legal compliance teams and enterprise startups are getting in our way. They can uh figure
9:16
They can uh figure out some type of settlement agreement that involves um uh a sort of MMA I was going to say cage match pay-per-view on TVPN.
9:26
I was going to say cage match.
9:28
That's like at a certain point like they'll reach some sort of you know end to the story, but I'm sure there will still be bad blood.
9:35
It has to be ended in the octagon in the octagon. They Yeah.
9:40
I mean, we've been teased so many times with, oh, is uh is Trey going to box Jason Calacanis or is Zuck going to fight Elon?
9:48
Like, this could be maybe third times the charm here.
9:50
You know, maybe this is the one that gets it done. Fantastic in Dubai.
9:54
Yeah, we'll we'll bring it out to the Gulf of the Gulf of course.
10:00
For some reason, it'll happen in the Gulf.
10:02
I think just the purses haven't gotten quite big enough, right?
10:06
Like you see some of these boxing pay-per-views. Yeah.
10:08
Easily into the nine figures.
10:10
If we can, you know, position this as non-dilutive funding. Yeah.
10:15
For these startups, maybe there's something to be done.
10:16
Maybe that would get them over the line of saying, "Yep, I'll I'll put on the gloves. I'll get in the cage." Yeah. Yeah.
10:21
It I mean, it's been a fascinating story.
10:24
Obviously, these things move like extremely slowly as they go through the courts.
10:27
Uh so it feels like you know uh with FTX or Theronos like we we got the bombshell accusations then it took like a year or two to get any sort of to move past like alleged wrongdoing to actually understanding the scope of what happened.
10:44
Um but I'm sure that the courts are working through it and we will be following it the entire time.
10:47
Um the lawsuit this is from Techrunch which is turning 20 years old next year uh next next week.
10:53
Um, the lawsuit is also full of insults hurled at Ripling's CEO, Parker Conrad, and mentions his troubles at his previous company's benefits.
11:02
Irrelevant in my opinion.
11:02
Uh, uh, at times the There's nothing wrong with having a couple thrown a couple back in the office.
11:09
Wait, he had uh some unfortunate things happen at Zenits. Really? Really?
11:15
I'm hearing this for the first time.
11:19
You got a sound effect now.
11:19
I missed it on the first attempt.
11:22
At times at times the complaint ventures into psychoanalysis territory to understand Conrad is to understand rippling the suit claims.
11:32
Man there everyone's chirping but you know I think I think this show we want to be independent so we're taking the side of co because co is invested in both. Yes. Yes.
11:40
Let's give it up for let's give it up for diversification in crossover allocators. Yes.
11:44
Just just heads I win heads I win tells you tells you lose. Exactly.
11:50
That is the real the real game.
11:52
Yeah, we've talked about this before.
11:54
Ultimately, there's hundreds of billions of dollars of payroll market cap and they can both be big businesses and hopefully they get over uh you know, and maybe this is all a sideshow for for for what's really going on at I mean I mean quantum payroll.
12:10
This all comes down to both companies racing to develop quantum payroll quantum payroll.
12:14
Quantum payroll for sure.
12:16
And that's kind of the real story.
12:18
And then this is just the drama floating up to the surface. Yeah.
12:23
And so yeah, they're they're they're um they're fighting back.
12:26
Interesting timing, too, because wasn't it just yesterday that deal said it had been profitable for years and is generating over 1 billion in annual revenue.
12:32
And so I wonder if there's a sequence of events here where it's like, okay, take a breather, be really silent for a while, then come out with some some promising news about the financial health of the company, then fire back in the media and in the courts with a with a counter suit.
12:50
Um, but, you know, unclear and it doesn't seem like it's quite as aggressive as what uh Zenitz found or or Ripley found.
12:59
It's certainly not as uh it doesn't take you on as much of a journey. smoking gun necessarily.
13:03
I mean the uh the the techrunch article here actually says when Y Combinator grad Code tool launched an agentic security platform last month among other things sets up honeypotss it its ad was a spoof on how Ripling's corporate spy said he was caught and so um clearly it's become a uh a small meme within within Silicon Valley.
13:27
Anyway, um, if you want to save time and money and not have any headaches in your back office, have you noticed, by the way, that when I point at you, it pops up on both cameras, so I can actually get two hands out at one point. Oh, you can. That's good. Um, there we go. Anyways, go to ramp. com.
13:49
Easy use corporate cards, bill payments, accounting, and a whole lot more all in one place. Ramp ramp ramp ramp.
13:55
Anyway, um using ramp is a joy. Yes, it is.
14:00
Um people have said that that you know corporate cards and spend management platforms are boring.
14:07
Yeah, but anybody that says that clearly has not used ramp.
14:09
We were we were just pitching ramp to we we did a photo shoot yesterday and uh we were pitching ramp to our to to the photographer telling him, "Hey, stop. We never miss a moment. We never miss a moment. We never miss a moment." Yeah.
14:23
Hey, those were He should be tagged automatically. He was by the end.
14:25
I mean, he was convinced by the end.
14:26
He was ready to he was ready. He was ready to switch. Yeah. And you should be too. Thank you to ramp.
14:32
Um, we have a story here.
14:35
Microsoft bets on agents to fuel next chapter of AI growth. Yeah.
14:37
So, we heard that that quote from Satcha Nadella that uh the amount of inference that's happening on Azure has 5xed and they're generating I think hundreds of billions of tokens at this point. Trillions of tokens. It's absolutely massive.
14:54
So the question is Microsoft's in a perfect place to deploy agents.
14:58
They have the distribution, but how who who's actually behind this team and we said yesterday they already have 70% of the Fortune 500 using C-pilot. Yes.
15:08
Which is to say what that means, right?
15:12
Evaluate what that usage actually looks like, but yeah.
15:14
How would we even do that?
15:16
I mean, because it it's it's they were obviously very successful with Teams, right?
15:19
they got teams into everyone and they really did switch over and spike the growth of that company uh and that and that product.
15:25
Um at the same time, you know, Google's been quoting saying that they they're using generative AI search results and it's it's kind of it kind of counts, but it kind of half counts and you have to like discount it a little bit because it's not it's not a consumer choice to move over to that product.
15:43
Just kind of happens naturally.
15:45
And so it's interesting to dig into Microsoft strategy because they seem to be the way I heard it described was that Sachi Nadella has obviously done a fantastic job as CEO of Microsoft and he's carved out a ton of territory in artificial intelligence.
16:03
The question is how much can he hold on to, right?
16:06
And and and this isn't to say that like seeding ground is a is a loss necessarily like it could be the strategic move, but there's this question of like even just to the point of I want to be a leaser.
16:19
How much how much capex do I actually want to be spending?
16:20
Do I want to own the land?
16:22
Like you could you could vertically integrate all the way down in the AI factory, but they're partnering with people. They're leasing.
16:27
They also own data centers.
16:29
own data centers. and and there's a question of do they you know Microsoft has an has an AI research team that um at one point was training models now they seem to be much more model agnostic and you saw that with Microsoft build where Satcha highlighted the importance of being able to choose your own model
16:47
which is something you can you can do on GCP and AWS as well but Microsoft has really leaned into that even with intelligent model routing within different yeah different different OpenAI and Llama instances and there's lots of different ways that they've plugged in so Let's dig into how they're betting on agents. This is from the
17:01
This is from the information.
17:02
Uh in March at a Microsoft all hands meeting, one of the company's newest executives, Jay Periq, laid out a rough vision of Microsoft's path forward in artificial intelligence.
17:12
AI models made by OpenAI and others were quickly becoming commoditized by more efficient models from Deepseek and Microsoft's own research arm that are that that performed nearly as well for a fraction of the cost.
17:23
said sitting alongside chief executive officer Sacha Nadella and chief technology officer Kevin Scott that meant it would soon become easier for companies to build their own AI applications and that Microsoft could cash in on that growing market by redoubling its efforts to sell them tools for doing so said Periq.
17:40
Um, in recent months, Nadella has made similar comments in staff meetings telling employees that Microsoft needs to focus on platform, platform, platform, a refrain hearkening back to a speech by Nadella's predecessor, former Steve Balmer from the late 1990s in which he uttered the memorable chant, developers, developers, developers.
17:58
And uh, if you haven't had a chance to go listen to Steve Balmer's interview on Acquired, it is fantastic.
18:04
Uh, they've been posting a bunch of clips.
18:06
I listened to most of it. It is. It is.
18:07
He's really opens up and it's it's uh it's just an incredible piece of history. Uh and it's really fun.
18:14
Um there's a clip going up right now uh where Balmer talks about when he took over as CEO from Bill Gates.
18:20
And essentially he him and Gates didn't talk for a full year. Wow.
18:27
Even though Gates said, "Hey, I want you not just to be a figurehead.
18:31
I want you to actually be the CEO.
18:33
That means I will report to you.
18:36
you're the boss, but they didn't really he the way he described it was like they just didn't really know how to how to deal with each other anymore, how to like how to work together.
18:43
It was very interesting, like like somewhat emotional even.
18:47
Um anyway, um Nadella is clearly, you know, paying paying homage to Steve Balmer um but also focusing on this idea of platform platform platform, which is similar to developers.
18:58
Like he wants developers on top of the platform, but uh you need you need to put a little twist on it.
19:03
I'm actually excited next week at demo day with YC.
19:04
I want to get a sense of how many companies are building on the broader Microsoft platform, Azure, etc.
19:13
We talked to someone at last demo day who was there handing out tons of credits. Yeah. Yeah. Yeah. Britain. Yeah.
19:19
It was kind of a I'm sure he'll be back there.
19:21
He runs their startup uh team.
19:22
And so yeah, I mean I can I I can imagine if you're I if the platform is offering a lot of flexibility around these tools like building on top of that platform lets you switch in and out of different models very quickly means you don't have to shift as soon as like a new model is going viral.
19:39
You're just like okay I just swap it into my my Azure stack.
19:40
I don't even need to set up billing on a new on a new platform.
19:43
Um the comments from Nadella and Periq reflect a subtle but important shift in AI strategy at Microsoft.
19:50
No big tech company has benefited more from the frenzy around AI than Microsoft, whose 3.
19:53
44 trillion market capitalization makes it the world's first or second most valuable company, depending on the week.
19:59
Nvidia took the crown on Tuesday. Wow.
20:00
Let's hear it for Nvidia.
20:02
So, so far, the bulk of the company's AI revenue has come through its relationship with Open AI.
20:09
And there's this really interesting Debart is just unbelievable. It's crazy.
20:15
He's he's doing unbelievable. Yeah.
20:17
So, uh, there's a good chart here from the AI bonanza.
20:20
Most of Microsoft's estimated AI revenue so far has come from its relationship with OpenAI, which includes revenue sharing and leasing Azure servers.
20:27
So, they're making about an estimated 10 billion this year from OpenAI.
20:33
and then three billion from other AI sales to bring their total AI revenue to 13 billion which seems significant given how what are the actual what what are the is that revenue that open AAI is passing back to them somehow. Yeah. Yeah. Yes.
20:51
So they get so they get a they get a revenue share from uh from OpenAI from selling OpenAI products from from vending uh GPT4 as an API.
20:58
Open also is paying Microsoft to lease Azure servers for training and inference.
21:06
And then I believe that they're entitled to a cut of revenue or profits up to that 100 million which I don't think I wouldn't assume has kicked in at all yet. Me either.
21:18
I thought I thought it was like earnings. Yeah. Yeah.
21:19
Um but clearly still that's crazy. 10 billion. Yeah. I mean it makes sense.
21:26
Yeah. I mean it makes sense. they they like OpenAI is training and the GPUs are on fire and they need to scale and so you know even if they were even if they had no relationship with OpenAI you would imagine that they'd be load balancing across the different
21:37
hyperscalers and and and trying to soak up GPU capacity wherever they could and so you know if you look at which is just crazy so it the numbers Microsoft invested a billion in 2019 2 billion in 2021 10 billion in 2023 and then 750
21:53
million late last year so and then 10 billion then they're they're generating you know obviously it's not necessarily super high margin but it's a lot it's a lot I mean the margin of of Azure is not low it's you know over 30% right so um they are particularly and they still own
22:12
half of 49% of openi global LLC it's amazing Satcha it's the best but again it's like you know he doesn't own it all and so as as as openai kind of goes more independent how much can Satcha like hold on to in terms of like the consumer AI market. If if if the if the if the
22:30
AI market. If if if the if the if the narrative around open AI as the un what did Ben Thompson call it like the unwilling consumer tech company or like the unexpected uh to consumer tech company like if OpenAI becomes the next
22:45
Google what will that relationship with Microsoft look like because they could I mean they're building their own servers with Stargate and so like that revenue could go away over the long term even though it seems like they will be partnered for a very long time. Anyway,
22:56
Anyway, we'll we'll have to dig dig into it more.
23:00
We want to have some Microsoft folks on the show and uh and I I would love to know how Azure is tracking against inference versus uh training loads because we didn't get that from Jensen in the NVIDI earnings call but that seems to be an important question that is on everyone's mind um as we hit the GPT 4.
23:21
5 and and and this and the pre-training scaling kind of wall.
23:24
Um obviously the the the the hope is that we shift to inference very smoothly and GPU demand continues to grow and the overall industry grows very quickly but uh it's still like an open question.
23:37
We don't have a lot of hard data on what's happening there.
23:39
Anyway, um Microsoft is particularly bullish.
23:42
This is from the information of course uh Microsoft is particularly bullish on a new category of AI applications called agents which will be able to carry out tasks.
23:49
is maintaining a spreadsheet to keep track of unpaid bills or patching websites after outages with minimal human oversight.
23:54
Agents are all the rage throughout the tech industry, not just at Microsoft, with other enterprise giants like Salesforce, Service Now, SAP rushing similar products to market.
24:03
The growth of agents could take off with cheaper AI costs leading to the rise of what some executives call the agentic web in which most of the world by autonomous AI.
24:13
We need to figure out and understand agent force which is Salesforce's digital labor platform what adoption actually looks like over there from my sense is like they're force feeding people it where it's like hey you like our CRM you will also enjoy our digital labor platform and if you don't buy it you will um you know we're just going to charge you more for CRM. Yeah. Yeah.
24:37
for CRM. Yeah. Yeah. There's a lot of these products that are that are seeing like rocketed adoption based on I mean it's almost like the bullcase for for for some of Google's tools that like V3 you know we were joking about that post yesterday that like some of the products
24:51
are amazing but hard to find but at least you know that if a Google model is going viral it's authentic like people really love it versus they're it's it's rarely just oh they just stuffed it in everywhere and like it doesn't really count like the numbers don't really count. um because it's like it's it's
25:07
um because it's like it's it's pretty hard to go and find these find these models versus you know if if Microsoft chooses to you know roll out co-pilot in in every installation of teams by default.
25:19
uh that could that that could trigger a lot of like daily active users, but are they really getting those tools or are those tools just kind of sitting in the background and then and then companies are going to other more focused more dedicated um startups or businesses for those those agentic workflows.
25:36
Um anyway, for Microsoft, taking advantage of the shift toward agents means making new inexpensive models available on Azure and as as an alternative to larger models, embracing open source protocols that make it easy possible to build agents and launching new products that let c let customers set up their own customuilt agents.
25:55
Uh uh Periq said at an event connected to Microsoft's build conference, "Our goal is to build a new stack that allows anyone to build AIdriven applications and agents and orchestrate them."
26:07
At the center of the shift, uh is Periq, a former Meta executive who joined the company October in an unnamed role. Very cool.
26:14
Just like, "Hey, we just want you." Off the org chart. Off the or chart. It's the playbook. Yeah.
26:20
In January, he became head of a newly formed unit called uh Core AI that unified groups from across the company, including the company's developer platform, GitHub, its internal developer division known as DevD, and several teams that previously ported to Azure head Scott Guthrie, and were focused on and focused on running AI models on cloud service.
26:40
Pique now oversees more than 10,000 staffers at Microsoft.
26:44
Let's hear it for massive org.
26:46
He's got almost 5% of the org chart reporting to him.
26:49
He's definitely on the org chart now.
26:50
Good luck getting him in the same room.
26:52
Yeah, you're going to have to rent a basketball stadium to have your staff meeting.
26:55
Uh, I'm sure Satia could arrange that.
26:58
Uh after this story was published, Nadella on Wednesday announced another reorganization to staff focused on agents consolidating executives running LinkedIn, Office 365 and business application applications under uh executive vice president Reeshu Ja whose groups will those groups will aim to sell out of the box agent applications to customers while Periq's unit focuses on getting companies to build their own agents on Azure.
27:22
So, a little bit of a divide between what's being vended into the uh the the the Office customers and what is more on the Azure side and and and enablement of developer workflows in on on top of Azure.
27:37
Um I'm surprised Microsoft hasn't slapped some agents in LinkedIn yet.
27:42
You know, they got to should be able to have a link an agent that just replies.
27:45
That's where Clippy needs to come. Yeah. Bring us Clippy.
27:47
We got to bring back Clippy.
27:50
I think we can make it happen.
27:51
It's It's We're so ready.
27:53
I sat on the show and the entire time we'll just be like for Clippy.
27:56
I really think I really think it could make like actually we don't have any questions.
28:01
Like we really We're just pitching you.
28:03
We're just going to be pitching you on on uh I mean yeah Microsoft is one of those brands that's like it's still cool.
28:10
Um but it but it's not fun, you know?
28:13
It's a little bit it's a little bit like serious business and just having a little clippy used because it's such a it's so efficient behemoth and reliable and like and and and it's reinvented itself multiple times.
28:27
But they also have Xbox, you know, they like to have some fun.
28:29
They like to play some Call of Duty.
28:30
They literally own Call of Duty.
28:31
Like that's We need to uh we need to get a racing simulator here at the new studio.
28:38
Speaking of speaking of games, and a golf simulator in the green room to just let Yeah.
28:42
let guests, you know, start to get warmed up.
28:44
Flight sim flight simulator.
28:46
That's been like a 30-year project.
28:48
Uh anyway, if you're if you're designing um for anything really, get on Figma. Figma.
28:54
com, think build bigger, build faster.
28:56
Figma helps design and development teams build great products together. Go to figma. com to get started.
29:01
It is the backbone of the show.
29:04
I want to see if they even put customer lo Okay, so they do.
29:06
You're you're really into the customer logos. I am.
29:08
I just think people need to understand like the the range.
29:14
Coinbase, Dribble, Dropbox, GitHub, Herman Miller, Microsoft, New York Times, One Medical. Yeah.
29:26
Racketin, Slack, they got them all. They They got them all.
29:29
And you should be on Figma, too. Go check them out.
29:31
Try some of their new products. They are fantastic. Yeah.
29:35
Um they are going to be uh so this is another quote from Danny Fish a Janice Henderson investor portfolio manager who oversees two funds that hold a total of $800 million in Microsoft stock.
29:46
He says there are going to be software companies that are able to embrace and adopt that and there are going to be software companies that are going to find highly disrup that are going to find it highly disruptive to their models.
29:56
Microsoft ability to embrace that will be really important.
30:00
So you want to you want to offer enough tools to empower the companies, but you also need to allow the the flexibility so you don't lose companies who migrate off platform because they're just like I'm going to build everything myself with new, you know, I'm going to vibe code a bunch of agents and I don't need you for anything.
30:14
You you want to have like the full continuum and seems like uh Microsoft's in a in a position to kind of like index the market.
30:19
It will be interesting to see we should start pressing pressing pressing more of the founders that come on that are building kind of like agentic enterprise workflows and see how they're positioned against uh Microsoft.
30:31
Are they seeing Microsoft deal cards go up against them when they're when they're pitching um or or is it like purely additive?
30:38
cuz I feel like um for probably for the Fortune 500, you get a very white glove experience with Microsoft and they tell you every product they're working out, but uh in in more of like the SMB self-s serve market, it might just be a situation where you you know see a viral video or get an intro from an investor and then you start spinning up whatever.
30:57
Yeah, I really want to get a better sense of what B2B agents are getting uh have sticky usage, right?
31:03
I can imagine the obviously you see it in developer tooling.
31:09
tooling. seems like it's getting there in legal sales I think is happening but at the points that I notice it are when the person you know the the the agents are sort of messing up and and saying you know hey uh I I enjoyed hearing you
31:27
on X podcast talking about Y subject with Z person like actually saying that I got an email today uh of somebody that that said Jordy, it was great hearing you on X podcast talking about why this Wait, wait, but it hadn't populated it. It didn't populate X podcast. Why? That
31:45
It didn't populate X podcast. Why?
31:45
That doesn't even feel like an LLM hallucination.
31:50
That feels more like an if statement gone wrong. If statement gone wrong. Wow.
31:55
Well, I mean, maybe maybe uh for for the one person that went on a podcast that's just named X to talk about Y person named Z because there's Y Combinator to talk about There's probably somebody who goes by the name Z out there, right?
32:10
It's actually surprising that nobody's said drop the combinator. Just why? Just why?
32:17
I mean, that's what happened with the YMCA. They call it just the Y. Yeah. Yeah. Just the Y.
32:20
Go vertical with Y combinator.
32:24
Periq was impressed by the team of of Microsoft employees who developed autogen and open source framework for building AI agents and aimed to move more of those employees into his unit according to research including researchers that were previously within
32:36
Microsoft's research unit led by Peter Lee but a Microsoft research vice president who oversees generative AI research push back on Periq's attempt to move the researchers to his organization according to someone who spoke to her. So the information is kind of digging
32:46
So the information is kind of digging into like all the internal politics of a 220,000 person organization as you're trying to build a different team uh kind of build a a green field AI agent strategy.
32:58
There's obviously a lot of chips moving around the board, a lot of a lot of uh you know internal resources changing hands and you want to hold on to your uh to your best people.
33:06
Uh the uh the internal politics at Microsoft must be staggering given the scale.
33:12
I mean, it's like it's it's like Washington DC in size, so it makes sense. A nation state. Yeah.
33:18
I mean, 220 people 220,000 people is like a small city.
33:23
Like a mediumsized city, actually. Yeah.
33:25
It's basically the size of Qatar, I think. Wow.
33:27
Like the actual residents.
33:31
Does Microsoft have a 747 yet?
33:31
They probably should get one. Yeah. Yeah, they should.
33:35
they should. I heard I heard a funny thing that apparently the apparently the guitar jet was up for sale for like three years and I think maybe no one no one wanted to buy it because it had been like so overly retrofitted to be like opulent that people were like there's
33:50
too much gold there's too much gold that's flowing it's it's hurting the gas mileage like this plane is now too heavy like we can't do it so there's 380,000 Atari citizens so quite a bit more but uh Microsoft could look to acquire a a small nation state and rebrand it as Microsoft Microsoft land. I mean, it's the ultimate it's the
34:12
I mean, it's the ultimate it's the ultimate out ofome ad is just to be on the map. Yeah. Yeah. Yeah.
34:16
Well, speaking of out of home ads, go to adquick. com.
34:20
Out of home advertising made easy and measurable.
34:21
Say goodbye to the headaches of out of home advertising.
34:23
Only adqu combines technology, out ofome expertise, and data to enable efficient, seamless ad buying across the globe.
34:29
We did a photo shoot yesterday.
34:33
We're going to be going up on a billboard, baby. I can't wait for this. I need photographer.
34:36
I need to get these ASAP. I'm so excited.
34:40
We'll be dropping on the timeline. Stay tuned.
34:45
And please uh subscribe to us on X.
34:47
Follow follow follow the TVPN on X.
34:47
I don't know why you would see this and not be subscribed, but there might be people in the RSS feed or something.
34:55
Maybe you've been you may maybe it's like, ah, it's too much content.
35:00
A, let the algorithm sort it out, but B, just give us a follow just for a little bit.
35:04
We're so close to 64,000, which is another doubling.
35:07
We'll be doing something to celebrate. So, thank you.
35:09
Uh, and uh, in other news, Snowflake is uh, buying crunchy data for $250 million.
35:15
This is from the Wall Street Journal.
35:17
The cloud data company aims to attract uh, customers who want to build their own artificial intelligence.
35:23
Let's give it up for a base hit for these investors that uh, 250.
35:26
It probably didn't return the investor's fund entirely, but it's still a fantastic outcome hopefully for the for the team.
35:35
And so we got to we we we have to do like a database day or or get a bunch start talking to some of these database folks because uh Snowflakes, Snowflake, Data Bricks, Palunteer, they're all making serious moves in this space and they're all kind of moving to different layers of the stack.
35:51
stack. like data data bricks is now more of this like data lake like data unification layer and then Palunteer sits on top as like the ontology layer actually un understanding how all of the different data interacts and so they're kind of playing nice and um data uh snowflake and palunteer used to be kind
36:12
of comped directly to each other but they've they've diverged in the public markets um but still uh you know snowflake's like a fantastic and like incredible story of an incubation at Center Hill Ventures that went massive and IPOed and went into the tens of billions of dollars. Data Bricks is also
36:27
Data Bricks is also up there in the tens of billions of dollars.
36:30
A lot of people are waiting for them to IPO and so it's a very interesting dynamic.
36:33
Now Data Bricks bought Neon, a similar database startup um in a deal valued for about a billion.
36:40
Now Snowflake is buying Crunchy Data. Um great name. Yeah, Crunchy Data.
36:44
the the the the the naming schemes in like enterprise SAS like deep down in the are are just wild and entertaining.
36:52
Like data dog remains one of my favorite startup names of all time. I mean it's so funny. Data data dog.
37:00
I mean you have uh last YC demo day we met the founder building Pig. Pig. Pig was great too.
37:05
We got to check in on Pig.
37:08
We got to check in on Pig.
37:10
I just I don't exactly remember what they do but but I but I remember the name. Yeah. Yeah.
37:14
I It'll only be a matter of time until we see like the the I'm now searching Pig AI and luckily I was able to find it. pig. dev. There we go.
37:23
An API to launch and automate Windows desktops. Let's hear for Pig. Let's hear for Pig. Just a great name.
37:30
I think, you know, if Pig is successful, it will inspire a generation of of companies named after animals. Monkey, goose. Yeah. Horse. Horse. ai. Potentially.
37:42
We don't have any animal name companies.
37:45
Although we do want uh we do want animals to start sleeping on eight sleeps.
37:50
That's something we're pushing for.
37:52
We think animal testing is going to make a big comeback in the mattress market.
37:55
Eight sleeps are so comfortable.
37:58
We're going to have your dog should be sleeping on eight.
38:01
How'd you sleep last night?
38:01
I think I had a rough night.
38:03
I was up at 5 on the grind. Let's see.
38:06
I actually didn't I didn't get the hours in.
38:09
You didn't get the hours for six hours and 50 minutes. 80. I got an 84. So close.
38:17
I got six and a half, but then So close, but so far. Yeah, it happens.
38:22
It happens to the best of us.
38:22
Anyway, um it's pretty sunny. Funny.
38:24
My um uh my son like came in our room at at probably like 2:00 a. m.
38:30
I was I was half asleep, but then he just slept like um uh uh like perpendicular perpendicular to to us. Yeah.
38:40
So he was getting two different temperatures like half his body was getting you know my wife's temperature half the other.
38:47
Um, which had been a funny just weird weird experience, but he's he was sleeping well.
38:54
He was sound asleep when I when I left, of course. Yeah. Yeah.
38:56
I remember showing showing my son the uh the aid sleep app and I just had you know how you can see like the left side and the right side and like what the temperatures are and he was like and me for the center and we're like no they don't make that yet.
39:09
And also go back in your room go back in your room.
39:10
Um, I had to pull up startups named after animals because there are actually a few.
39:18
Mailchimp, HostGator, Mailchimp, Muleoft, Post Hog, Post Hog. Post Hog is like pig. It's the same thing. A hog and a pig.
39:25
Post hog is one of the greatest post Yeah.
39:28
I mean, it's it's an iconic It's like postmodern. Like post hog. Like we are post hog.
39:33
We are now in domesticated hogs, which are pigs.
39:34
So pig in many way is the post post hog. Yeah.
39:39
I don't even know if they do the same Uh, Task Rabbit, Survey Monkey, Hippo Insurance. It was a big The zebra. Yeah. Attractive. Okay, now we're going. Now we're hallucinating. We're hallucinating. Fat Llama. Appearl.
39:51
A peer-to-peer rental platform. Fat Llama. It's a great name. Lunch Badger. A cloud services name. Cloud. These are hallucinated. This is not real. No, it's in. It's in a Crunch base. Let's go.
40:05
I mean, if it's in crunch base, it's Well, is it in crunchy data?
40:10
The thing that stands out about pig is that it doesn't add anything before or after. It's just pig. No, it's like ramp. Yeah.
40:16
Or like Vanta, automate compliance, manage risk, prove trust continuously.
40:19
Vanta's trust management platform takes the manual work out of your security and compliance process and replaces it with continuous automation whether you're pursuing your first framework or managing a complex program.
40:28
Uh, back to Crunchy Data.
40:30
Crunchy Data has roughly a 100red employees scaled pretty quickly.
40:32
They'll join the company once the deal closes which is expected in to close in the next couple weeks.
40:37
Crunchy data will be part of an offering called Snowflake Postgress.
40:40
The vision here is that Snowflake Postgress will simplify how developers build scale uh build, deploy and scale agents and apps.
40:47
Very relevant to that Microsoft uh story we were talking about earlier.
40:51
Uh with that in mind, it was important to acquire a company that's not just that was not just engineering and quick experimentation.
40:57
Crunchy data is a cloud database cloud-based database provider that helps large businesses and government agencies use Postgress.
41:04
Postgress without needing to manage infrastructure themselves.
41:06
I've used Postgress before.
41:08
There are some it's crazy was founded in 2012. Wow. Overnight success. Overnight success. You love to see it.
41:17
Congrats on the $250 million sale. How much did they raise? I wonder. Um I'll find out.
41:22
There's a couple other um deals.
41:26
Snowflakes offers a platform for storing, organizing, and analyzing data across multiple cloud providers including AWS, Azure, and Google Cloud.
41:33
The company, which went public in 2020, grew rapidly during the pandemic as more companies migrated their data to cloud storage from onremise data warehouses, which is it's crazy that we're still in like the the cloud migration era from onrem data warehouses.
41:47
Uh this is like cloud data has existed for a decade and still happening.
41:52
And uh if the AI roll out is any any this any similar or if it tracks similarly, it's like we could be talking about agents for the next decade and just continually rolling these these products out um in a very very slow takeoff scenario.
42:06
Crunchy data only raised 14 million. There we go. Very efficient. Very efficient.
42:12
Well, speaking of other efficient businesses, there's a story from our friend Chris Best in the information.
42:17
He says, "Apple's app store changes, quote, have been fantastic."
42:22
And this is why I wanted to highlight the the the uh uh the efficiency.
42:29
They he said uh Substack was accidentally cash flow positive in the first quarter of this year. Fantastic.
42:37
Growth has been translated to revenue.
42:40
Best said the company which was founded in 2017 was accidentally cash flow positive during the first quarter of 2025.
42:46
That said, Substack is not focused on profitability right now.
42:50
He said as it hopes to continue investing in growth.
42:54
Said that Substack apps you uh users are heavily in the US where Apple has been ordered by federal court in Northern California to allow app owners to offer users alternative payment mechanisms allowing Apple to allowing apps to bypass Apple's 15 to 30% fee for inapp purchases.
43:11
So had Apple been taking Yes.
43:14
If you went in the app and then you subscribed to a uh a Substack and you used the inapp payments flow, they would take 30% of that on an ongoing basis.
43:25
I didn't even know that was possible.
43:26
I've never signed up for a Substack in the app. Yep.
43:28
And so that's crazy because this this Yeah.
43:33
I mean, it's a digital product, right?
43:36
And so it is it Well, it's interesting because Audible has found a way to get around this by credits and stuff.
43:43
credit system, which is just so annoying.
43:45
Yeah, I think that that's a bigger negotiation because it's Amazon. Yeah.
43:49
Uh and I think I think I think Apple had more leverage over Substack, but we should have Chris back on the show and ask about how it actually works.
43:56
But yeah, I mean, if you're I feel like most Substack users would be actually fine going through a web flow.
44:02
Yeah, it's not as much of like an impulse purchase.
44:04
You really, you know, you're you're you have a relationship with the Substacker that you're subscribing to.
44:09
You kind of understand and there's and it's pretty easy to message.
44:12
I feel like the average Substack user probably understands a little bit more about the Apple App Store desktop respectors. Exactly.
44:22
There's a lot of corporate athletes, emails, corporate athletes, you know, people that are sitting in front of their computer all day long.
44:27
And what better place to buy things than the computer.
44:31
Buying things on the computer is really historically and still today just a fantastic experience.
44:36
It's fantastic, especially when you have your sales tax automated with numerals. Numeral HQ.
44:39
Put your sales tax on autopilot.
44:42
spend less than five minutes per month on sales tax compliance.
44:45
Um, you know, you may may be able to avoid the Apple tax, but you will not be able to not this one.
44:52
Uh, avoid state sales tax on software and consumer goods. Sales tax AGI. Yes.
44:59
Anyway, Substack has now has more than 50 creators who are making millions of dollars per year on the platform.
45:05
Best said the platform overall has more than 50 million five million paid subscribers.
45:12
So yeahund million dollar.
45:15
I didn't realize they raised so they raised so much but that's great that they're profitable. Yes.
45:17
So I mean uh I I think this is such a great story where they this is a business that that raised ahead of where they were. Sure.
45:29
And then a lot of people wrote them off because they were like, "Wait, this is just a media business with a take, you know, a take rate and and all this stuff."
45:37
And and what Chris has been able to do over the past year, you know, when we had him on the show, I was I was kind of asking, you know, has it felt like Substack has broken through really into culture? Yep.
45:49
Uh becoming like a a real brand itself outside of, you know, Twitter, right?
45:54
Because they basically got like Elon came for them hard. Totally.
45:58
And that was like somewhat warranted because they launched like a competitive product.
46:02
I thought the competitive product happened like as a response to the link ban.
46:06
I thought that was the sequence of events.
46:08
It was it was links being getting deprioritized somewhat and then Substack launching a post competitor.
46:16
Um, which I don't even know if I don't even know if I agree that that was like the right move.
46:20
I haven't I haven't really played with that product.
46:22
Well, you know, it was kind of obvious that X was going to ban links. Yep. Either way.
46:26
Yeah, that hurt Substack.
46:30
It hurt the writers on Substack.
46:33
Uh and was the right the right decision for for X. Yeah.
46:36
Um but uh but yeah, so there was a link suppression.
46:43
Um and then and then uh and then Substack Hamish uh one of the co-founders called Elon Musk a propagandist with more conflicts of interest than El Chapo. Spicy.
46:55
Um but uh ultimately um interesting.
47:00
I I had never I never I never picked this up.
47:03
What uh apparently Musk had had made some type of proposal to buy Substack in 2023. Yep.
47:08
But probably after that they restricted links.
47:13
Y um and and and and Twitter formerly had another publishing platform.
47:17
I I don't remember the name, but they rolled that in and then and then I think they closed it down and now you just have the ability to write pretty long posts.
47:25
You can post articles, X articles directly on X.
47:30
And I think that it's it's weird because I we have yet to see like we are an example of a of a media business or a show that just said we will play the game by Elon's rules on X.
47:47
So we we by we we don't we don't think about links at all.
47:51
We think about what are the what are the products that X loves text posts memes images video uploads live streams and we do those very very well and that's what we focus on and we don't so so we're not intention we're like leaning into what the platform loves.
48:08
We should actually consider doing articles on X.
48:10
The question is, what would it look like if you tried to build a Substack like business?
48:17
Like you're you're just a writer and your output is articles and you use X's subscription tools and the articles feed to have that experience of one free article a week and then one paid a week and you try and build up the book of subscription business on your X account.
48:38
I've seen some people that have subscriptions turned on.
48:39
I think I actually technically have subscriptions turned on.
48:42
I think only Gary Tan is like the one who who subscribes to me. Shout out Gary Tan.
48:45
Thank for Gary just it just like throws me a couple. Yeah.
48:48
Ultimately the thing what's interesting about Substack is they were uh you know effectively benefited from Zerp in terms of accessing capital and people started to write them off because I don't think I think at times their growth was not bestin-class.
49:01
They were just sort of chugging along.
49:03
chugging along. And if you uh if he had originally had this idea of hey we're going to build this publishing platform uh that's going to allow independent you know sort of uh citizen journalism and writing to flourish uh and then we're going to launch this sort of social
49:20
network with streaming and all this stuff that uh at certain times would have been hard to believe but they actually have executed that to a tea and now when you go on Substack it it does feel like a social network you know based on you know email as a sort of backbone. So, I'm excited to follow
49:34
So, I'm excited to follow their progress more and eventually get uh really set up on Substack ourselves. Yeah.
49:42
So, next up we have Augustus Dico coming in.
49:46
There's an article in Semaphore today.
49:47
Uh China boosts use of cloud seating to combat droughts and we wanted to have him come in and break it down for us.
49:55
The headline is uh China stepped up cloud seating in the face of severe drought.
50:00
The country the country's gain grain growing regions in the north have been parched for months leading to concerns over the harvest.
50:06
Uh though some scientists are skeptical over cloud seeding's effectiveness and environmental impact, several countries have begun deploying it.
50:14
China is already the world's leading user of weather modification firing chemical compounds into clouds to spark p precipitation and has conducted more uh conducted 20% more than by this time last year.
50:28
uh apparently causing a one-third increase in rainfall.
50:30
That seems pretty significant.
50:31
So, let's bring in Augustus Dico and have him break it down for us. How you doing, Augustus? Good to hear from you. I don't know.
50:40
That sounds like the cruise ship or something. Yeah. Yeah. Welcome.
50:44
That's that's the sound that that August that's in Augustus wake, you know, whenever he goes anywhere.
50:48
Uh it's great to have you. Yeah. Thanks, man. That and size gongs. That and size gongs. Um update soon.
50:55
Wait, you got to listen to this one.
50:57
journalists on the horizon. Stand by. Great. Nice.
51:00
Who Who was the guy that was responsible for posting the uh stay tried to stay focused on the mission gif a while back when I got that text?
51:06
I got like a hard eyes text about like, "Oh, I liked your appearance on TVPN."
51:11
Good social media intern there. Yeah. Yeah. Yeah.
51:15
But the team's growing over here.
51:17
We We got a good crew coming together.
51:19
Uh there's vibe coding happening over in this in that part of the studio.
51:21
There's a lot of production stuff going on. Uh it's been a fun time. Um, sweet.
51:25
So, uh, yeah, break it down.
51:27
I'm I'm here to ring the alarm bell.
51:29
Um, totally transparently, uh, there are really big problems on the horizon both with China's domestic weather modification program and then their international one.
51:38
Um, the article that you read or that you're referencing talks about how China is actively retrofitting their Wing Long Twos, right?
51:46
Um, talk about like nominative determinism. Cool name for a drone.
51:50
Um, but it is the uh essentially Chinese equivalent of the MQ9 Reaper.
51:54
Um, and they're using it for cloud seating and weather modification operations all across the country.
51:59
A lot of that is to fill up the snowpack in Tibet and then use that as a natural water tower for runoff for all of their urban, industrial, agricultural, and environmental assets in China.
52:10
So that's that unto itself is insane.
52:12
To recontextualize people, the Chinese Meteorological Administration has about a $300 million budget for weather modification.
52:20
They have 38,000 employees exclusively working on weather modification.
52:24
And they have two universities that offer bachelor's degrees in weather engineering, not meteorology, not atmospheric science, specifically engineering the weather.
52:33
Um, so they are driving extraordinarily hard on this domestically just for their own water supply, just to green deserts, just to keep their cities and industry humming.
52:40
The problem is uh the international implications of this, right?
52:45
We are in a and like I critique people all the time for saber rattling with China needlessly, but the Wing Wong 2 has been sold and is being operated in countries across the world, namely Saudi Arabia and Egypt for defense applications, right?
53:00
So that has problems for defense, but also for weather modification.
53:04
The CMA has explicitly said they want to export cloud seating as a means of soft power to control water supply and weather across the world.
53:13
They already collaborate very closely with the Thai Royal Rainmaking Department.
53:16
Um, and they can easily retrofit these drones that they've sold for defense applications.
53:21
Royal Rainmaking Department. Wow, great name. Great name.
53:26
Um, but they can easily retrofit these drones for weather mod as well across the world.
53:30
And then not only control the shipping and receiving and these ports, not just the energy infrastructure. Nice.
53:37
Um, but also the water supply and weather.
53:40
And, uh, President Lyndon B.
53:42
Johnson said, uh, whoever controls the weather controls the world.
53:44
Um, and right now we're trending towards a world where China controls the weather and subsequently the world.
53:49
Um, and it's really Rain Maker Technology Corporation representing the United States against the CMA.
53:55
Um, so President Donald Trump, if you can hear us, uh, Secretary Rubio, if you can hear us, uh, the State Department should be involved in this soft power conflict on weather modification internationally.
54:04
H how are you thinking about the the current pitch for weather modification?
54:09
Because it feels like this uh this news out of China has a few different hooks.
54:13
uh as well as agriculture and drinking water.
54:17
Much of China is reliant on hydro power for electricity.
54:20
Uh Sichuan in the southwest gets 80% of its power from dams, meaning that droughts can lead to electricity shortages.
54:27
I I mean I know we have the Hoover Dam, but is hydro electric power important in America?
54:35
Obviously power is top of mind for everyone, but in America it feels like the narrative has shifted to nuclear and solar.
54:41
Um, but is there a world where we could be getting more out of our existing hydroelect electric assets and there's maybe a narrative there or not? 100%. Right.
54:50
Nuclear is awesome and I'm super excited for Valor Atomics to turn a thousand reactors online in their gigasytes and produce the world's energy.
54:58
But it's going to take at least a year for that to start and it'll probably take years longer still until that's our main form of stable base load.
55:05
Solar's great, but we have nighttime still and reflect orbital hasn't solved for that problem.
55:08
So uh we need stable base load and we need clean base load and hydroelectric power across hundreds of dams in the American West produces hundreds of gig gigawatt hours and we can refill our own dams to increase more stable clean base load with cloud seating where these dams are drying up right now.
55:26
we we can increase supply.
55:27
You know, 80% of uh you mentioned Sichuan, 80% of Colombia's power comes from hydroelectric as well and they're going through a drought right now and so they have rolling blackouts because there's not enough power there both domestically for our own energy production.
55:41
We could use cloud seating to produce more hydroelectric and then internationally again as a means of collaborating with other countries, let's call it, and ensuring that like they have American interests in mind.
55:52
We can produce more water and hydro for them.
55:53
Can you talk about the Chinese approach to cloud seating versus what you're doing in the American approach?
56:00
I feel like um a lot of times when we see a competitive dynamic emerge between China and America, there's only a small tweak between the way Instagram reels are served versus Tik Tok or, you know, DJI drones versus GoPros.
56:15
It's usually just the scale and efficiency and reliability of of Chinese technology, but there isn't usually that much of a of a shift in the underlying strategy.
56:24
Um, are they using the same chemicals?
56:27
Do you think we should be using different chemicals from their mix?
56:29
Uh, are they using different drones or are they shelling this stuff into the into the into the atmosphere with howitzers?
56:36
Like, are there is there anything that we can learn that might not be IP protected that we could safely port back?
56:41
Is there anything that we should change based on what we're hearing from over there?
56:45
Um, China is throwing the kitchen sink at weather modification research.
56:50
Um, so they're doing drone based aerial cloud seating.
56:54
They're doing ground generator based cloud seating.
56:55
They're doing uh acoustic uh cloud seating research, meaning they have these huge 130 decel uh speaker systems where they just blasted up clouds.
57:04
They put them all in Tibet.
57:06
Uh so even though it's like destroying the ears of Tibetan villagers, they're trying to shake water out of the clouds.
57:12
Um so that they have a bunch of other photonic stuff.
57:15
Um they're doing a ton of research, but really the the big and important aspect of this is their sophisticated uh military retrofit of drones for long endurance missions.
57:27
Um their radar research for detecting phase change in cloud.
57:32
Um, and then lastly, I think the thing that they have like the most serious edge on the United States and anybody else in the world in is their uh ice nucleation agent and their particulate.
57:42
Um, they're doing a bunch of uh nanoparticle design.
57:44
So, super super small scale particle coating.
57:47
It's titanium dioxide on top of these salt crystals among other things that are way more efficient at nucleating ice and subsequently creating snow or condensing stuff in cloud.
57:59
Um, Raymakaker is doing research into that right now, but that's where China far and away has the the biggest lead on on the nanoparticles that they're using.
58:06
How about your challenges at home with various states and regulators? What's the update there?
58:14
Um, so uh 31 states proposed legislation to ban all forms of weather modification this year.
58:20
Um almost all of them dropped that legislation.
58:23
Uh because one um because of you because of you.
58:29
Was it did you We did Me and the Rain Maker team did a lot of work around state capitals.
58:35
I've got like a regular barber shop in Tallahassee and a few other state capitals um to to tune the mullet mullet up before I testify.
58:41
Um but uh the one state that did ban it was Florida.
58:47
Um, Florida made weather modification a class two felony.
58:50
Um, so if I were to work there, I'd go to prison for five years. Wow.
58:53
Um, and that I think unto itself is not a huge problem. Fine.
58:58
Sure, you're depriving Floridaian cattle ranchers and orchards uh from having as much water as they want and they do have wildfires and droughts.
59:06
So, it hurts the state of Florida.
59:08
Um, but the real problem is the canary in the coal mine uh in terms of American political sentiment, particularly Republican political sentiment. Right?
59:15
sentiment. Right? There's this whole conversation around the tech right and who is pro- innovation and who's not is the Trump administration pro- innovation seems so for the most part but um unless we have clear federal regulation on weather modification and a public stance in favor of this then we're going to lose control of the weather to China
59:33
yeah what about um what about desalination so it's something that we were doing in America we kind of fell off it feels like the nuclear story and it just feels like I'm gonna hear a story in the next few years of like, "Oh yeah, China just figured it out and now they have like a bunch of deselination plants and we're behind on that too." Mhm. Tracking it at all. So desalination Mhm. Tracking it at all.
59:54
So desalination is largely held up by the California Coastal Commission.
1:00:01
Um and then like HOAs basically that block the construction of del.
1:00:06
Um an interesting stat that because it smells bad.
1:00:08
No, just because it looks ugly. Diesel smells fine.
1:00:12
It's just like a big Well, it looks I should It looks industrial, right?
1:00:16
It looks It looks like a big beautiful oil refinery, which I happen to be a big fan of personally.
1:00:20
Um, but you know, folks in Newport Beach less so. Um, diesel is great.
1:00:25
Uh, and if we come up with some really sophisticated new reverse osmosis membranes or catalytic desalination methods with good electrical engineering, um, we can make it more efficient.
1:00:36
But the problem with del still is that we have to move that water from the coasts.
1:00:40
like it's a non-starter for Nevada or Colorado or Utah to get desalinated water.
1:00:46
Um what you can do though is cloud seed obviously, right?
1:00:51
You could produce water anywhere where there are clouds uh with our tech.
1:00:54
Yeah, I remember Blake Masters was saying when he was he was running in Arizona, he was saying like the future of California is nuclear power del and then we need to uh reroute the Colorado River to hydrate the the inner the the inner states.
1:01:09
Um is there a world where you could build I mean you mentioned it looks like an oil and gas refinery.
1:01:15
Could you build a like an offshore oil rig essentially or is it just like it's not economically dense enough?
1:01:22
Water is not the same as oil.
1:01:24
So, you're not going to be able to pay to put it on a truck and then bring it in.
1:01:27
It just doesn't make sense to do it that way. Yeah, exactly. It must flow.
1:01:28
I was I was talking to some commodities traders the other day and I was trying to like come up with some crazy derivatives for water. Water right there yet.
1:01:39
You know, it's it's it's it's cents for a barrel, right?
1:01:41
And it's just as it's almost as heavy as oil.
1:01:43
So, you you can't convey this.
1:01:45
You know, 13% of all of the electricity in California is used just moving water around. What? That's crazy. Wow. Yeah.
1:01:53
Like the Central Valley exists just because we pump all of that stuff from the Sacramento Delta and American River down into the valley and elsewhere.
1:02:03
Like there's this huge huge unnown problem which is like because we don't have enough water, we have to dedicate so much of our energy resources just to moving it around where we could just be producing in the Sierras.
1:02:15
Um and that's what rain maker's doing. Interesting.
1:02:19
Any anything else you're tracking from China?
1:02:22
Um so I guess one one thing that I will say in terms of like this soft conflict is we had a customer meeting um in the Middle East um and a day before we had the scheduled meeting they said hey sorry we have to go do a last minute
1:02:35
trip to China um to go talk about the uh Winglong 2 um so uh we're trying to qualify rain makaker vehicles in the Middle East right now um so that we are clearly at par with their system capabilities and then can selected for the Chinese. Um but yeah, the challenge
1:02:52
Um but yeah, the challenge if you if you look at the precedent for sort of statebacked companies out of China, they're willing to sell at a loss for years.
1:03:01
And so how you know the you know I I have no insight into whether they would try to attempt that in in cloud seating, but it wouldn't exactly be surprising, which I'm sure is a competitive dynamic that you're thinking about.
1:03:15
every single international office that we go into to talk about our work, we see a stack this high of purchase orders from China, um, and then one that's like two pages thin, uh, that's that's from American companies.
1:03:28
So, yeah, it's absolutely a problem. Wow.
1:03:32
Uh, part of this article highlights this idea that, uh, uh, grain growing regions in China have been parched this month for or four months.
1:03:41
Are we in a particular like global drought? Is this unique to China?
1:03:46
Are we experiencing drought in America?
1:03:49
Like what is the state of drought generally?
1:03:51
You mentioned that there are different pockets around the globe.
1:03:54
But um is this a particularly bad year?
1:03:57
Are the trend lines bad overall or I mean we saw the fires and that felt in a it felt very visceral in in California and Los Angeles, but you never really know when you zoom out where we are on the trend line.
1:04:09
line. So we we had we had out of distribution high amounts of precipitation this past year in California and yeah um and even still we had the wildfires right um you know the the thing that I think is a good reference point is the California
1:04:27
Department of Natural Resources water supply strategy they explicitly plan for half a million to a million acres of farmland in the central valley to turn into desert in the next 5 years just because there's not enough water. So
1:04:38
So even when we have boom years and the reservoirs are all full and there's tons of snowpack and everybody gets to go skiing in Tahoe, there's not enough water for current demand.
1:04:45
And that's in part just because population is growing, right?
1:04:49
Like we built the Central Valley Water Project to turn the Central Valley from a desert and swamp into the most productive agricultural region in the world and and we wouldn't have the US population that we do now if it weren't for that water project.
1:05:01
And so we we just need to produce more if we want to maintain agricultural and economic growth. Makes a ton of sense.
1:05:06
Well, good luck out there.
1:05:09
We hope you can uh strike some big deals on the back of this news.
1:05:12
Uh, you know, we got to be competitive.
1:05:14
Thank you for fighting on America's behalf. We'll talk to you soon. Come back soon. Later. Cheers. See you. Bye.
1:05:21
Next up, we have Brad from Cobbot coming in the studio talking about collaborative robots.
1:05:25
I'm very interested to talk about the the sim toreal gap which we covered yesterday.
1:05:31
Uh, we will welcome Brad to the studio. How are you doing? Welcome to the show. Thank you for joining. What's new? Oh, fantastic outfit. Welcome to the show. Amazing. Looking great. Uh what's the occasion?
1:05:48
First, introduce yourself, please.
1:05:49
But uh explain uh why the fantastic outfit today. Are you doing real work? That's right. Yeah. Yeah. Hi, I'm Brad Porter.
1:05:58
I'm founder and CEO of of Collabor Robotics. The uh Yeah.
1:06:00
This is how you tell whether uh whether a robotic company's really in production or in the field if they're wearing their safety vests.
1:06:08
And uh I was at our uh our deployment with Maris yesterday and so uh had had this handy and thought uh there we go.
1:06:17
I'd bring it out for you guys. Yeah.
1:06:19
So I mean to the degree that you can talk about it, uh what exactly are you doing for Maris because that sounds like important work.
1:06:26
Yeah, we're uh we're helping them in in transload operations in moving in unloading ocean uh freight containers and loading out onto tractor trailers.
1:06:38
Um the you know they load carts, industrial carts uh full of kind of up to 1500 lb worth of generally boxes of product uh for retailers.
1:06:47
Um, and then we help with the moving the carts because moving those heavy carts around all day long is is pretty pretty taxing work.
1:06:56
And we've got You said 1500 pounds.
1:06:58
How big is one of those boxes?
1:07:01
Uh, like I I'm familiar with like a 55gallon drum.
1:07:03
I'm familiar with like a pallet of goods that you might see in an Amazon warehouse. How big are we talking? Yeah.
1:07:09
So, think of this as as the types of boxes and and that would that would flow to a to a retailer, right? To a big box retailer.
1:07:17
Um and so they're unloading those from ocean containers onto carts that are about um 3 and 1/2 ft wide by about 6 and 1/2 ft long.
1:07:27
Um and so they just load up as many as they can on the cart and then take it to you know usually these are getting dispatched out to big box stores.
1:07:36
Um and so you know there might be six tractor trailers that are getting loaded up to go to six different stores in a in a region.
1:07:44
Um, so we're working um in the in the Sar Washington area, so out of Seattle port um and helping basically get distribution out to Pacific Northwest.
1:07:56
Can you talk about some of the differences about unloading at a port versus what uh Amazon's KA systems does within the warehouse and and some of the different challenges that you face versus what KA is doing?
1:08:08
I imagine that some of the there's some learnings that cross over, right? Yeah.
1:08:14
So the way you can think about logistics is there's there's inbound flow, you know, products coming from manufacturers around the world, a lot of it coming from China.
1:08:24
Um, and then that ends up in some distribution warehouse ready for people to to buy.
1:08:30
So it might end up in your in your local big box retailer.
1:08:35
You can just go and, you know, buy a fan off the shelf.
1:08:37
Um, or it ends up in an Amazon warehouse, uh, an Amazon fulfillment center.
1:08:42
So the inbound side of that is to unload the ocean containers and then bring it to, you know, some someplace where it's being stored or bring it, you know, to a to a retail store or to an Amazon fulfillment center.
1:08:56
And then an Amazon fulfillment center, yeah, that KA network or now Amazon robotics um what they call kind of the Hercules drives uh is their storage array.
1:09:06
So, Amazon will have multiple mezzanine decks of um of those ka pods full of all kinds of things that you might buy from Amazon.
1:09:17
Literally can have a million different SKUs in a in a building.
1:09:22
Um and then when you when you order it almost immediately, the the system knows where it has lots of those and they're stowed across Amazon's network.
1:09:34
It quickly calculates where is the most optimal place to deliver this to you.
1:09:39
And then a robot goes and gets that shelf and brings it to uh a picker, brings it to someone who pulls it out of those shelves, puts it into a tote, and then those totes get routed to a pack station, gets packed, thrown uh and then it gets sorted to a truck and then usually, you know, either to FedEx or UPS or Amazon's delivery network or to USPS.
1:10:05
Amazon can kind of deliver into any of those outbound delivery networks.
1:10:09
And so, so yeah, three phases coming in from the manufacturers, stored and ready to be bought and then shipped to you.
1:10:18
How how uh where are we at in the kind of AI journey of these collaborative robots?
1:10:24
Um I imagine that there's tons of work that can be done with just hardcoded business logic.
1:10:29
drive two feet forward, take a left, and it kind of just is almost like a conveyor belt on on independent wheels.
1:10:37
Um, versus, you know, the far future where the robot has a brain and is just making completely independent decisions and decides where it goes and pro and problem solves and reasons and and we're on the cusp of that, but I imagine that there's a there's a journey that we're going through.
1:10:55
Um, and so walk me through where we are in terms of that journey. Yeah.
1:11:00
So, we we've made a lot of progress from the days where you just like followed a tape line on the floor. Yep. Right.
1:11:05
Um now robots can can generally sense and perceive and navigate um commercial environments autonomously quite well.
1:11:15
usually, you know, at human walking speeds, uh, maybe a little faster, but, um, the kind of self-driving problem is reasonably well solved in, uh, in commercial spaces at those type of speeds.
1:11:30
And that's generally done with a LAR um, and maybe, you know, stereo depth cameras.
1:11:35
And so, you know, LAR based slam or um is how it localizes and then navigation and planning and that can be done in a way where it can detect humans, obstacles, navigate around things.
1:11:49
Um, and that's the the the capability our robot has and it can do that in hospitals in um, you know, ultimately airports, stadiums, in and around people quite safely. That technology works.
1:12:03
that what's coming now is you can talk to robots and the robots will make the the highle plan and instruct that um where we need to get to is what we can do with our hands, right?
1:12:18
Where you know like open up your AirPod case and pull out is a very complicated set of motions that we do without thinking about it.
1:12:25
We don't quite have that capability yet. Got it.
1:12:27
Talk to me about the LAR supply chain and cost structure in your business.
1:12:31
Um uh it's been a controversial debate point for a long time in autonomous vehicles.
1:12:37
Um but if the economic model works, it feels like it's just pure value ad.
1:12:42
Is the cost of LAR uh getting lower?
1:12:46
Are you thinking about solid state LAR coming down the pipe or is it already available?
1:12:52
Um, are you banking on a reduction in LAR costs over the long term or does your business model just by nature of how much value you're adding, you're fine paying 50k or something like that if that's the number? I don't know.
1:13:06
Yeah, liars are in the kind of they used to be 50k, they're in the kind of $3 to $6,000 range right now. Yeah.
1:13:12
Um, and so you're right, you do have to add enough value.
1:13:17
You're not going to put that on your Roomba, right?
1:13:19
Um, but you can put that on an industrial robot um and uh and get a payback.
1:13:27
Obviously, we want to keep seeing those costs continue to come down, but uh but it's not a it's not a prohibitive um element for a lot of the work that needs to get done out there.
1:13:40
Can you can you talk about uh form factors broadly and and what guided you your thinking towards the proxy the initial product and other products in the suite feels like you guys distinctly chose not to do humanoids even though I'm sure various VCs thought hey why
1:13:58
don't you have you guys thought about doing humanoids demo have you seen this viral video from Boston Dynamics yeah um so I'm curious uh you know kind of the the decision-m that went into uh the current and and uh pending form factors. Yeah. So I mean as much as like the Yeah.
1:14:12
So I mean as much as like the humanoid hype seems to have been peaked in the last you know 12 18 months uh I think Elon did a lot to kind of fan that humanoid robots have been around for quite a while.
1:14:24
Agility systems has been uh electronic these guys have been at it for a while.
1:14:28
And so um so when I was leading robotics for Amazon um the we we studied deeply uh humanoids in in 2018.
1:14:39
I looked I I went through my like hype phase on humanoids in 2018.
1:14:41
I got really excited about them.
1:14:43
Uh what agility was showing at the time and legged mobility looked like it could actually work and it it does.
1:14:49
Um and so I had my team at Amazon do a full analysis of everything we weren't going to automate in another way where a humanoid could help.
1:14:56
And I remember reviewing this paper, there were 40 diff 40 different use cases where a humanoid could be great, right?
1:15:05
Um and and then I looked then we looked at all 40 use cases and we said actually to solve these problems, we don't need a humanoid.
1:15:12
Um and in fact a humanoid is kind of too complicated. You really want wheels.
1:15:16
You kind of want to move more than three and a half miles an hour.
1:15:19
And is the number is the number of motors a potential issue too from a degradation standpoint?
1:15:26
It's like, you know, we've talked to to other robotics founders that say, you know, I'm using, you know, robotic arms in my facility and we already have to replace those motors all the time.
1:15:34
And then a humanoid might have an order of magnitude more and actually be less productive than some type of robotic arm.
1:15:42
So the number of motors really does drive the overall cost.
1:15:45
It also drives the complexity of the controls, right?
1:15:50
And so you get into a world where you need AIdriven controls.
1:15:51
But the problem, the real problem that people don't talk about very much with humanoids is uh getting strength out of rotational motion is very hard, right?
1:16:00
You um you're effectively because you're just doing a short throw.
1:16:06
You don't have the kind of momentum flywheel effect as you get the torqus rolling on your electric vehicle, right?
1:16:11
You're moving through very short distances.
1:16:13
So all the power is basically your electric magnet and then your rare earth magnets.
1:16:17
and and so you end up needing bigger and bigger motors to get that kind of power.
1:16:21
Um and so and then you want to wind them with the absolute highest density you can and so they end up and then you want to run them at almost the peak current that you can to get the most strength.
1:16:35
So, the problem is humanoid robots either have to put this way big motor that doesn't look right in the shoulder.
1:16:43
Um, but what they're typically doing is they're they're hand winding the motors, they're pushing the current to the max and even then they're getting maybe 60% the strength of humans and they they burn those motors out.
1:16:56
So, the motors are very expensive and they burn them out very quickly and they're still not as strong as a human.
1:17:01
And so, it just it we need some breakthrough.
1:17:04
You know, the pneumatic, like, you know, Boston Dynamics had that Atlas robot that like could do back flips and everything.
1:17:11
Pneumatic is 10x the uh the power of an electric, right?
1:17:15
And but you can't you can't really make that system reliable in production.
1:17:23
So, do you think there you think there are more consumer use cases for the humanoid form factor, maybe around the home?
1:17:30
How do you think about applications outside of industrials?
1:17:34
You know, I I I have struggled to find someone someone mentioned one to me the other day that that seemed great, which is like is walking your dog.
1:17:42
I think humanoids walking the dog would be would be quite interesting, quite cool.
1:17:45
Um I suppose you could have a quadriped but uh um otherwise I am I am not bullish.
1:17:56
I do think there there's some some cool robots recently that are more kind of friendly.
1:18:00
looked like they were kind of playing a game with a kid.
1:18:02
Like I I think kind of the emotional companion idea is quite interesting.
1:18:06
Um but yeah, getting the strength is tough.
1:18:12
I mean, even just thinking about the human arm, like the force that's generated from the human arm is from like the bicep muscle, which is much bigger than the actual joint.
1:18:21
And so if you put a motor on that joint, you you're not humanoid with just absolute cannons.
1:18:27
There is there is a a humanoid company that's trying to create the muscle fibers and pull that and that's sounds like some of the pneumatic projects and maybe it'll be a hybrid.
1:18:34
Um in terms of training and AI development there's been this talk about the sim tore gap.
1:18:42
tore gap. I don't know how closely you've been tracking this but obviously generating data for robotics has been very difficult but now there's this new paper that semi analysis was talking about yesterday um all about training in simulation uh basically Unreal Engine you build the robot virtually you have it walk around learn as much as it can
1:19:02
in in thousands of years of artificial data then there's going to be a gap between what it experiences in simulation and reality and and so what you do is you take what it's learned in simulation and you and you run that on a robot in a cage basically wired up with a with a power cable so that it can run forever and and it and it tries to do the moves that it learned in simulation. It messes
1:19:23
It messes it up but then that generates more data that feeds back in.
1:19:27
Does that seem like, you know, power generation, all the all the mechanical issues aside, does that seem like an interesting path to go down for actually solving the the algorithmic and like the AI piece of understanding these uh how these robots will will actually choose what what motors to move at what times.
1:19:47
And do you have any experience generating data? Just kidding. I'm just kidding.
1:19:51
Uh former CTO of Scale AI, if anyone's listening that's that's not familiar.
1:19:58
Um, no that's so, so the challenge in robotics first is how do you get I mean it's the same in large language.
1:20:07
What's the pre-training phase, right?
1:20:08
How do we get some base level in pre-training large language models?
1:20:12
It's to kind of understand how words are likely to follow each other, right? Just statistically.
1:20:17
Um, so motor actions, what's what's likely to, you know, to cause the arm to move forward and and things like that.
1:20:24
But the hard part in um in any AI system are the edge cases at the end when you're when you're interfacing with with the real world, right?
1:20:38
And you know, fortunately, we have all this large language model data.
1:20:44
We have all this data from the internet to give us reference examples of what the real world of language looks like, right?
1:20:50
right? Um and so we refine on that and then we use human preference to refine even further and that's how we get you know chat GBT in in the uh in the robotics world data from simulation data from multiple robots um data from tea
1:21:08
operation all of these are kind of techniques people are using to feed some data into you know what's kind of the pre-trained base model that gets some statistical sign correlation but when it comes comes to learning the edge cases, right? When it comes to, hey, that
1:21:21
When it comes to, hey, that doorork knob is higher than this other doorork knob or the doororknob does, you know, turns upward instead of downward.
1:21:29
Um, you uh you and I actually selfplay to figure that out.
1:21:35
We come up to a doorork knob that doesn't It's funny, we have a door at Kobat that you you can either push the handlebar or there's a handle.
1:21:40
Well, everyone tries to push the handle and then the door doesn't open and you have to push the um and so so humans get confused too and we do this kind of refined selfplay and I think right now we're very much focused on the pre-training phase just how do we get enough data to have something that like roughly moves its hand toward the door. Sure.
1:22:00
To really solve this problem though, we've got to learn how to self-play in the real world um like you or I do because there's all kinds of novel stuff we're going to run into um solving real problems.
1:22:14
Well, good luck with that.
1:22:16
It sounds like an easy task, but uh I'm sure you're up to the task and uh it's been fantastic talking to you.
1:22:23
Yeah, this has been super insightful.
1:22:24
Yeah, I mean it's such an exciting industry because it's really just like like we're still just on the early the early part of the S-curve and there's going to be fantastic advancements. So, good luck.
1:22:35
The future's going to be amazing. Awesome. Thank you guys.
1:22:38
Appreciate you coming on. We'll talk to you soon. Thanks, Brad. Thanks so much. Talk to you soon.
1:22:42
Next up, we have Keon from Nucleus coming on uh with a big announcement.
1:22:46
Something like 10 years in the making, close to it, maybe seven years. Uh we'll bring Keon in. Let's play soundboard. How you doing? Welcome to the show. That's a great intro. Their tweets are flying. Oh my god. You guys seeing this? Yes. You seeing this? Seeing this. Break it down for us.
1:23:08
Explain what's happening.
1:23:08
There's nothing like a launch day.
1:23:09
I'm trying to figure out guys, is this is it data or is it Theronos?
1:23:13
Cuz people can't they can't make up their mind. Oh yeah.
1:23:15
You know, they're trying to figure it out.
1:23:18
They're trying to figure out what's going on.
1:23:19
Let's give some context to the audience.
1:23:21
Nucleus has launched Nucleus Embryo, the world's first genetic optimization software.
1:23:25
Basically, parents can, you know, give their children the best start in life.
1:23:28
They can pick their embryo based off of physical characteristics like eye color, IQ.
1:23:33
They can go to disease risk like cancers or heart disease.
1:23:35
Basically, we really believe parents can get old information that exists about their embryos and they can pick however they want.
1:23:42
Um, for me personally, you know, it's been 10 years in the making.
1:23:45
The journalist actually covered it today in the Wall Street Journal was a journalist that covered my gene editing in a warehouse in Brooklyn 10 years ago. Yes. Let's Wow. Overnight success.
1:23:53
It's uh you know it's long time in genetics. Yeah.
1:23:58
So so so break down the state-of-the-art because like embryo screening exists.
1:24:02
I think most parents in America at least if they the means do some sort of screening uh while the embryo is growing. Is this purely for IVF?
1:24:12
Is this just going a layer deeper?
1:24:15
And then is I want to talk about the regulatory and FDA component as well.
1:24:19
Yeah, let's talk about it.
1:24:19
So, oh, basically if you go to IVF clinic today, you're a couple.
1:24:22
The vast vast vast majority of clinics.
1:24:24
The first thing actually understand is that the IVF process is principally controlled today by clinicians or doctors.
1:24:29
Honestly, couples don't have as much liberty in our perspective as they should. It's their baby. It's their embryos.
1:24:34
They should have the right to those that information and they should be able to pick off any vertical.
1:24:38
However, today in the clinic, what generally happens is people test embryos for very rare and severe genetic conditions.
1:24:45
For example, like a chromosomal abnormality like Down syndrome for example or even a condition like cystic fibrosis or T-ax or PKU, right?
1:24:53
These are conditions that are very rare um that maybe someone might have a carrier for cystic fibrosis but it's it's pretty rare.
1:24:59
Um then there are conditions that we've all heard of heard about things like breast cancer, things like coronary artery disease, the things that actually kill the vast majority of people today, right?
1:25:06
chronic conditions kill a vast majority of people today.
1:25:09
Those conditions are just not tested for in the clinic even though we have very good science actually that can make those predictions.
1:25:14
How do we know this as a as a DNA companies?
1:25:15
Well, that's what we do, right?
1:25:17
We build models that predict disease and the way you test those models in adults.
1:25:20
So, we go from adults to embryos is actually because we can basically well validate these models to show that they work in both the embryionic context and in the adult context.
1:25:28
And so, what we're really doing is we're going from okay, instead of just looking for really severe like down syndrome cystic fibrosis, why not do breast cancer?
1:25:36
Why not do heart disease?
1:25:38
Why not do colorectile cancer? Right?
1:25:39
Why not do schizophrenia? Why do Parkinson's? But then why stop there?
1:25:42
And this is really the important thing because ultimately, you know, if you think about diseases and traits, the extreme version of any uh trait is actually a disease, right?
1:25:49
Height is a good example of this.
1:25:51
One extreme end is like, you know, John for example, he's like Markin syndrome almost.
1:25:54
Then the other end is like me, dwarfism, right? But on both ends, okay?
1:25:58
So, so you know, so you know, IQ is another example of this.
1:26:03
One end is like, you know, autism.
1:26:05
the other end it can actually be some sort of you know cognitive basically challenge that people have and so when you think about it when we start realizing that people have drawn a line in the sand saying you can get you know rare diseases you can't get common diseases but then they really said you can't can't get any traits like height even though the best predictor we have
1:26:22
today actually in the world the best polygenic predictor is for height so as a company we've kind of completely reimagined this and said wait a second what's going on here you should have access to the entire stack rare diseases we do cystic fibrosis is common diseases like breast cancer and also traits all the way up to something like IQ. Yeah. Yeah.
1:26:39
So I mean uh that test are you just giving people the data because I imagine that once you get into particular recommendations that's more what I would expect a a licensed doctor to need to do.
1:26:52
Well yeah my sense is that they you can allow people to get the data from their doctor and then and then feed it into Nucleus. Is that correct?
1:26:58
So that's that is correct.
1:27:01
And actually we have a couple there was like 10 announcements today. You know how we do it.
1:27:03
We like to do 10 announcements in one day.
1:27:04
We are actually very very excited to announce uh a huge partnership with Genomic Prediction.
1:27:11
So Genomic Prediction is actually the oldest uh embryo testing uh a company that exists.
1:27:14
They've done genomwide test on embryos for almost a decade at this point.
1:27:18
And I think they've done over 120,000 couples for PTA, which is specific kind of test.
1:27:21
And so we're actually partnering with them.
1:27:24
So we make it very easy for genomic prediction customers to request their files and actually port it over to Nucleus.
1:27:30
But really, this isn't just for genomic prediction customers.
1:27:32
Anyone who's undergoing IVF can go to their clinic and say I want my embryo's data.
1:27:38
You can take that data, you can upload it to nucleus and all of a sudden you know the application of your D of DNA makes this technology universally uh basically universally accessible.
1:27:45
Now, how much of how much of the benefit is is actual uh algorithmic analysis bringing in other data points to contextualize the data versus just better UI and better hydration of existing text because we we uh we we had we had a friend on the show who was talking about um getting some medical results from a doctor.
1:28:08
The doctor's office was closed.
1:28:10
It took two days to until the doctor was going to be able to interpret the results.
1:28:14
He was able to just take a photo, upload it to ChatGpt and say, "Hey, is this is this a you know, is this really really bad?
1:28:21
Should I be panicking because it seems somewhat out of the range?"
1:28:24
And Chad GPT was able to say, "Hey, you still got to talk to the doctor, but this isn't this isn't the craziest thing I've ever seen.
1:28:29
This isn't way out of distribution."
1:28:31
And so that's almost like a pure UI layer, but extremely valuable.
1:28:35
I know it might not be like the right narrative for some people that it's like not as innovative, but I think that like all that matters at the end of the day is giving people benefits, right? Always both. It's always both. You you you Yeah.
1:28:46
Fundamentally technology just for technologies sake is not what silicon is about, right?
1:28:50
Silicon is about making something that people want. Okay.
1:28:51
And people can actually use.
1:28:53
So you think about the the nucleus innovation. It's it's it's twoprong. Okay.
1:28:55
One is in the informatics, right?
1:28:57
You know, I've been building this for 5 years.
1:28:58
I I almost I would argue to myself that I probably spent too much time, you know, developing the science, right?
1:29:03
Because science in is in a nutshell isn't actually very useful.
1:29:06
You need expanded access to it.
1:29:08
So on that point, we do multiple different kinds of analyses is they make it such that we can actually provide the most comprehensive analyses that exist today.
1:29:15
But moreover, and this is really I think a key point to your point, John, is people understand them. People can see them.
1:29:21
I mean, you can pull up the platform.
1:29:22
I'm not sure if you guys have shown it already, but it's very easy to sort compare your name embryos.
1:29:29
You can stack rank your embryos.
1:29:31
You can understand what the score means.
1:29:33
We lead with overall risk or we tell you for example instead of saying you're in the 99 percentile for genetic risk for a condition which you know what does it actually mean?
1:29:38
We say hey you have you know 5% chance or the like of let's say schizophrenia or some other condition.
1:29:43
In other words by even overall risk people have much greater intuitive understanding of the results we're communicating to them.
1:29:47
We have genetic counselors on hand.
1:29:48
So this really is a what are we showing here?
1:29:51
Are we showing we showed Yeah. Yeah. Yeah. We pulled up here.
1:29:55
That's another thing that's that's a fun one. That's an Easter egg. That's an Easter egg.
1:29:58
That's the that's the kind of approach that we're taking here.
1:30:00
And I think consumers are responding to it, right?
1:30:04
People want to have access to their data.
1:30:05
The clinician, the doctor shouldn't decide what ambate implant. You should. Okay.
1:30:09
So, talk to me about what requires FDA approval.
1:30:11
Obviously, uh new medical devices like if you were developing a machine to take in an embryo and sequence the DNA, I would expect that the FDA would want uh an approval for that medical device.
1:30:23
But if you are taking data and just showing it to a customer in a different UI, that feels like probably a very light FDA process and there's probably a continuum in the middle where once you're making a recommendation, they they have rules around that, right?
1:30:42
We as a company do not tell you which embryo to implant.
1:30:43
You know, basically parents, the couple has complete agency to decide how they want to use the information to implant their embryo.
1:30:50
Moreover, let's be clear, height, right?
1:30:53
I mean is can a height analysis be a medical device you know that doesn't even make sense right IQ height these traits for example we all you know traits are something that I don't think necessarily belongs in even the kind of infrastructure of think about medical care right these are things that go beyond medical care these are things like you know that that people just kind
1:31:09
of intuitively know and that there are DNA tests done every single day TOC um for these analyses because they're not disease analyses right so we do both diseases and traits to be clear my point is many of these innovations you have to wonder like you know should should the government say if someone can or cannot pick their embryo based off height. That
1:31:23
That doesn't seem right to me.
1:31:25
I think it should be in the complete liberty of of the individual to decide that. Yeah.
1:31:27
But I mean, we're we're a democratic country and so if if you know a huge swath of the population says that the FDA should review that type of uh test or that type of analysis, analysis, it could happen.
1:31:43
I mean, the the FDA reviews all sorts of different stuff.
1:31:46
And so I I guess the question shifts to like do you expect a change from FDA on uh the way these uh these analysis tools are regulated?
1:31:59
Um I think right now the most important thing is just putting these high quality rigorous scientific results in people's hands and then helping them basically have healthier children, helping them give their child the best start in life. Yeah.
1:32:11
Um, you know, I I think that generally speaking that, you know, people should have more liberty, more choice in in medicine.
1:32:17
I think the broader longevity trend actually touches on that point as well. Yeah.
1:32:20
Um, so that's what we're excited to do at Nucleus. Yeah.
1:32:23
I mean, the the fact that you're partnering with a company on on actual on the actual like medical device side, like they are doing the sequencing of the embryos, like that really takes it out of the Theronos question entirely.
1:32:34
In my mind, I think feel like you should beating the drum there a little bit more.
1:32:37
It's like like we didn't say we created some new device, but I don't know if we ship. That's the difference. We ship. It's locked, baby. Go look at it. Go use it.
1:32:45
That's that's the evidence. Why is there a footing? Okay.
1:32:52
Um I love the visual of John and his wife selecting between embryos and it's like 610 or 72. Tough choice.
1:33:01
Well, if we go with the 610, uh, he has, you know, potentially fly commercial once in his life.
1:33:05
We could actually, we could actually play this game right now.
1:33:08
Okay, here we're going to play a game right now.
1:33:10
I'm going to put in the chat. Pick your embryo. com.
1:33:14
Okay, everyone listening to this. Pick your embryo. com. I'm going to go to it. Oh my god. Here we go. Little Easter egg here. Okay, let's see.
1:33:20
What's more important to you, John?
1:33:23
Intelligence or muscle strength? Come on. Oh, absolutely. Muscle strength. Let's go.
1:33:26
We're the future bodybuilding. Let's go. Okay.
1:33:28
John would John would take a He would He would happily have a 52 son if he had, you know, top .
1:33:32
001% bodybuilding genetics. Exactly. Yeah. Okay.
1:33:39
So, here lifespan or height? Uh, come on. Lifespan. Lifespan. Let's go. Let's go. Let's go. Let's go. Let's go. Maybe low depression.
1:33:46
You got to be golden retriever mode.
1:33:47
You got to be uh You need low depression. You need low depression. Let's go low OCD.
1:33:51
I don't mind bouncing around a bunch. Okay. What's taking anxiety? Uh, let's go. High risk takingaking. There we go. Okay. Wait.
1:34:02
This is some generative stuff going on. This is great. I got Nadia, too. Enduring athlete. Let's go.
1:34:09
Physically strong, cautious, built to last. Yeah, this is great.
1:34:14
Is this driving a lot of uh a lot of attention, a lot of downloads? Is this going viral yet?
1:34:17
This seems like something that's designed to be shared.
1:34:20
We just dropped it right now.
1:34:20
Technology Brothers, we got you the exclusive. Let's go. There you go. Let's put it out there.
1:34:24
You can go and pick your embryo.
1:34:26
People say, "What's it like? Maybe not doing IVF yet. No problem. Funny.
1:34:27
Only only 9% of people choose Nadia. Okay.
1:34:32
Well, we're we're we're contrarian. We like that here. That's fun. Yeah. It's great. Oh.
1:34:37
Well, uh congratulations on the news.
1:34:40
Congratulations on the launch.
1:34:42
Um yeah, the pace is wild.
1:34:42
Last last thing, what's going on with uh Have you seen these just blood billboards? Oh, yeah. They're all over LA.
1:34:47
So, so there is there's someone who's running a campaign right now, Justice for Elizabeth Holmes, claiming that Theronos was not the scam people think it was.
1:35:00
And there's a there's a documentary coming out and there's billboards all over LA for Just Blood. Like, it's just blood.
1:35:05
It's not that big of a deal.
1:35:05
And John, to be clear, there's an exclusive on Technology Brothers next week about from this person, right?
1:35:11
They they're going to tell their story next week.
1:35:13
Just to make sure you you invited them already.
1:35:15
We're we're we are we are toying with the idea.
1:35:17
They someone reached out to kind of connect us.
1:35:19
We're we're thinking about doing it, but we're not we're not 100% sure that it would be appropriate for the based on the website.
1:35:26
I don't know if it I don't know if it's appropriate.
1:35:27
Yeah, it doesn't look like it was designed with Figma. So, I don't know. We can't quite do it.
1:35:33
It's a little bit The team definitely doesn't use linear.
1:35:35
Yeah, but they they they claim that uh that Elizabeth Holmes has been proven innocent.
1:35:39
And so, it's a bold claim.
1:35:41
We we like we like to see people making bold claims. By what? Uh jury is my question.
1:35:46
The jury of someone who knows HTML.
1:35:50
Keon uh always a great time. Energy is fantastic. Electric. Electric. Thank you for coming on. Firing us up.
1:35:57
Congratulations on the launch.
1:35:59
We will talk to you soon on Twitter for sure. Okay. So, we'll see you there. Bye, guys. Bye.
1:36:02
He's going through launch day right now, which is just like, you know, 40 notifications every minute forever. I love it.
1:36:11
Well, next up, we have Kathleen from Valos coming into the studio. Welcome to the stream. How you doing, Kathleen? Nice to meet you. Good.
1:36:17
Thanks for having me on, guys. Great to have you.
1:36:19
Uh, would you mind kicking it off with just a little bit of an intro for those who might not know? Yeah, absolutely.
1:36:24
So, I have been at Palunteer for the last seven years.
1:36:27
I built our life science practice there.
1:36:31
So taking some of the same platforms that Palanteer uses in defense and intelligence and a bunch of industries and then bringing those primitives over to pharma and biotech and researchers who need to study uh biomedical data in a secure and collaborative way.
1:36:42
So really building out end to end uh drug development workflows all the way from early discovery.
1:36:49
But then I left a couple months ago and now I'm working on something new.
1:36:52
Um so yeah, happy to chat with you guys. Yeah. Awesome.
1:36:54
I wanted to get your immediate reaction and and help us kind of contextualize the news. Yeah.
1:37:03
That came out yesterday or the day before, which was that two Chinese nationals have been charged by US federal authorities with conspiracy and smuggling after attempting to bring a dangerous biological pathogen, which I'm going to botch the name, into the United States.
1:37:16
This fungus is classified in scient scientific literature as a potential aggroterrorism weapon due to its ability to devastate key crops such as wheat, barley, corn, and rice uh causing a number of issues there.
1:37:28
causing a number of issues there. So um I wanted to kind of get a you know uh you don't have to go into too much detail but kind of a highle background on bioteterrorism broadly kind of maybe some like prominent examples and then um even just get your immediate reaction to
1:37:47
the news you know is this kind of thing surprising to you or are you surprised that we're not hearing more headlines about it uh all the time yeah totally so I think the to the second question first it's like a bad surprise obviously but we are absolutely entering this era of an elevated risk of bio threats. Uh so
1:38:04
Uh so sadly I don't think it's something that we should be that surprised about.
1:38:08
Um I think this story in particular and we don't know that many details on the story so I won't speak too much to it but the one it highlights how easy it is really to have any biological material pass across borders that obviously both natural or unnatural.
1:38:24
And then the second is that any kind of agricultural pathogen is also an enormous threat.
1:38:30
So I think when people think biod defense and bioteterrorism they think anthrax or small pox um which is obviously horrible but the idea that you could introduce a pathogen that would devastate one of our like primary crops that would have massive health impact but it would also completely destabilize our economy um and send us into something far worse than what we would see with COVID.
1:38:48
So I do think when when we and governments think about biod defense, it's very much both in terms of human health um and agricultural health.
1:38:56
But uh to your other question on what what's the primer on biod defense?
1:39:02
What are we worried about?
1:39:03
I think there's always been this um so we've always had the problem that people study the most pathogenic uh organisms in the world.
1:39:12
They usually do it in a biosafety lab, a BSL lab.
1:39:14
These are all over the world.
1:39:16
Some regions are obviously more secure than others.
1:39:18
So there's always this threat of something natural leaking or being used maliciously.
1:39:23
But basically two things changed recently.
1:39:25
It made that a lot worse.
1:39:26
One is that our ability to edit genomes and actually start changing those pathogens.
1:39:32
There's a bunch of new tools to do that.
1:39:34
So like things like crisper which you hear about usually in the medical sense of editing a genome. Yeah.
1:39:38
Are available on pathogens.
1:39:38
um synthesizing new nucleotides or new DNA to do that with also now commercially available like we could order uh nucleotides or you could print them out in some cases in a desktop printer.
1:39:49
So that made it now we're dealing with things that nature has never seen and then in the last couple years with large language models or biospecific language models we have this idea of AI uplift.
1:40:02
So, it means that someone um like us, not to underestimate your biological skills, but could actually be coached into how do we use these tools um guided all the way into making something that is actually way worse than anything nature has ever seen.
1:40:15
And this isn't like science fiction.
1:40:17
Like, uh Anthropic talked last week, I think, about Claude 4.
1:40:22
They did internal safety trials.
1:40:22
they noticed an uplift that goes way beyond just oh it's easier to Google a paper about how to do this and really into that um allowing a novice to access something dangerous.
1:40:33
So they released Cloud 4 with you know new security standards to try to combat that but a lot of models aren't like that.
1:40:38
Not everyone has those security standards you know.
1:40:41
Yeah, I mean this this has happened like uh uh what is it?
1:40:43
Timothy McVey looked up how to build a bomb with fertilizer and built and basically blew up like a massive government building and the the the wreckage from that was insane.
1:40:54
Uh if you think about what it would take a terrorist to uh work in a biosafety lab, it's obviously very complex but getting easier as you can have an LLM coach you through the process essentially, right?
1:41:04
Is that is that roughly the the the nature of the threat?
1:41:08
Yeah, both in terms of the actual steps to take and then what to change about a virus that would make it either evade some kind of countermeasure which is like some kind of medicine that we have for it um or be more infectious.
1:41:21
And I think the um the real like what has a lot of attention and weighs on people's minds is that we're not talking about like a state sponsored nuclear program that you need these like massive budgets and facilities to make something like that.
1:41:34
Like it's a laboratory and a computer.
1:41:36
Um, so it's a really different dynamic of threat than what we've been dealing with in the past.
1:41:41
And it kind of self-replicates by itself.
1:41:42
That's like the whole goal.
1:41:44
the whole goal. It's an interesting story to me because obviously over the weekend there was the Ukraine story around operation spiderweb and that was relatively asymmetric and that you know whatever the cost of the operation the trucks and the drones just you know even if it was hundred million dollars which
1:42:02
seems really really high right even if it was taking you know huge teams it destroyed you know a billion dollars plus of of assets on the other side and then when you talk about bow warfare and bioteterrorism It's like, okay, one or two people with access to a lab could potentially do billions and billions of dollars of of damage. Do you think this
1:42:21
Do you think this is a wakeup call or this will be a wakeup call for the government broadly?
1:42:27
How do you expect um you know, what are the different kind of uh ways in which the United States can defend itself from these types of um you know, obviously this wasn't whatever is being reported isn't a direct attack, but a but a potential uh threat in the future.
1:42:45
Yeah, the I mean the good news is uh DoD here, MOD in the UK, there's a lot of defense organizations already thinking about this and really trying to get ahead of it.
1:42:53
Uh it means that we have way more to do, but at least it's on the agenda.
1:42:57
And I think it comes down to so you're like the best defense would be to prevent this from happening.
1:43:02
So putting better safeguards on models, putting really strong um regulation on synthesis.
1:43:09
So who can synthesize what? How do we track that? that's great.
1:43:14
There's actually some legislation in the works on both of those things.
1:43:17
But the that works domestically and it works to contain an accident, but if we're actually talking about international collaboration, like those kinds of regulations are not really enough.
1:43:27
So the way then we get into thinking how do you deter something like this to your point like what is the the defense against this?
1:43:32
And there's really like three pillars that go into it.
1:43:37
The first is how quickly can you detect that something happened?
1:43:40
So in this case or in future cases do we immediately know that something new is circulating that it's uh high risk to us that it maybe will evade any kind of countermeasure that we have and can we know that before it's an outbreak and you know we're sampling from a hospital and then that leads into the second one which is how fast can we design or update a countermeasure so like an antibbody or some kind of biologic to combat that.
1:44:05
Um, and that is really like if we can diagnose and develop as fast as possible, then these weapons are much less powerful because you're not talking about billions of damages.
1:44:15
You're talking about a couple cases that are quickly contained.
1:44:18
And the last part of course is attribution.
1:44:20
So if you can actually say where did this come from? Did it come from nature?
1:44:24
Did it come from engineering?
1:44:25
Did it, you know, uh, did it come from a nation state?
1:44:27
That lets you bring the rest of the DoD and the State Department and our allies towards preventing something in the future.
1:44:34
And if you can get get that cycle down to hours really in terms of detecting, stopping, and then attributing, then you actually have a really robust profile for defense.
1:44:43
And there's a bunch of new tech that's going to help make that better.
1:44:46
So this is scary, but it also is there's good news on the horizon, too.
1:44:50
Can you talk a little bit about the bio practice at Palunteer?
1:44:55
I mean, uh there's been a lot of uh potentially like misunderstanding or misinformation about how Palunteer works.
1:45:03
works. uh you know I think most people who understand the company at this point understand that it is a it is an ontology platform that s is on top of a large database but I think what I'm struggling with is I understand if you're using like the Airbus uh case
1:45:18
study you have a database with all the different parts of the airplane and then Palunteer understands helps you understand how the different pieces and lead times for you know this screw and this seat seat belt and this engine part fits all together. So if you're demand
1:45:30
So if you're demand planning or figuring out how to manufacture airplanes, that's a very helpful tool to understand your supply chain, that makes sense to me in the very concrete widgets business that is airplane manufacturing, although it is obviously very complex.
1:45:42
In the bio or pharmaceutical space, I don't really understand the nature of how large these data sets are.
1:45:48
Are we talking about trial data or manufacturing?
1:45:50
Is it all of these above?
1:45:53
Like how does how does all that fit together?
1:45:55
um when you're thinking about applications of of understanding large data sets in just the bio world broadly. Yeah, definitely.
1:46:06
Um so the some of it is more similar to what you're you described with Airbus where we're talking about manufacturing.
1:46:13
We're talking about something that's like it's a process with a lot of moving parts.
1:46:16
How do you s make these all synchronous and update uh when you need to?
1:46:21
There's also obviously biologics manufacturing.
1:46:23
Um, some of it is more in the logistics end, which Palenter also talks about quite a bit.
1:46:27
So, when you're running a trial, making sure that patients and the medicine that they need are in the right place at the right time with the right support staff.
1:46:32
That's actually looks pretty similar to coordinating flight routes or staffing a hospital.
1:46:39
Um, the part that so those parts are super similar to the rest of Palunteer.
1:46:43
The part that's probably most unique is when we're actually talking about that trial data or talking about patient level information.
1:46:47
And there it really gets back to some of the the core concepts of Palunteer is how do you work with multimodal data and see patterns across it.
1:46:56
So if we're looking back historically on all trials that we've run and we want to start trying to identify what kind of patients have the best response to this or what are potential side effects, can you start linking together data that's from like a medical record with samples from a lab with sequencing data if you have it?
1:47:12
Um, and can you do that in a way that is completely secure, completely auditable, and complies with all the regulation in the space?
1:47:20
Um, and that's really the the niche that Palunteer maybe it's not always well understood to fit into, but uh that is what that platform was built for.
1:47:30
Can you talk a little bit more about uh the I don't know like the long-term future of what you're building?
1:47:37
I know that you can't go into it too much, but uh uh like there are a bunch of different vectors and opportunities around um what we're building in bio and what we're trying to prevent?
1:47:48
There's almost like a uh like like there's a little bit of game theory going on here.
1:47:52
So, um, what would you like to see the United States, uh, really really dominate going forward and where are the biggest opportunities, uh, to both increase biocurity and then also help accelerate the the developments that we need, the good stuff and that is what I'm most excited about.
1:48:13
All of these technologies that are really scary when we're talking about them like this actually do have the potential to give us this enormous global advantage and our bioeconomy and how we respond to these threats and make medicine too.
1:48:23
It's two sides of the same coin.
1:48:25
Um I think the two areas that we're most excited about.
1:48:28
Um one is on that detection pillar.
1:48:31
So right now a lot of the ways that we understand what's going on around us is super analog.
1:48:38
We have like a list of pathogens we're looking for.
1:48:40
we test and we get a yes or a no if those exist and not much more than that.
1:48:44
We've already seen a shift in this space towards sequencing.
1:48:48
So actually getting the DNA sequence from anything that's in the environment or any sample.
1:48:53
So that both broadens the scope of what we're looking at rather than just having tunnel vision on the pathogens we know about which means that we can start detecting unknown unknowns and things that we've never seen and make a risk assessment.
1:49:04
Also, using sequence data lets you have this much deeper level of insight into what the risk is rather than just there's a virus here.
1:49:13
You could say these mutations make it more adapted to this type of host or make it more dangerous or make it evade a certain type of medicine that we already have and then you can take all that information that you get on this more robust detection layer and use it to drive countermeasure design.
1:49:28
to drive countermeasure design. So we now have like every biotech that you guys talk to probably talks about programmable therapeutics where we're moving to this era where you can update um based on how the targets update and
1:49:40
if we have this deep level of intelligence we can also start thinking about rather than have a stale stockpile of medicines that were made 10 years ago can we actually see a threat immediately change the countermeasure um and then start deploying that immediately. So, I
1:49:54
So, I think getting that cycle down really tight, that's what um the future is hopefully. Are we testing enough?
1:50:00
I know when I go through TSA at LAX, someone who was doing uh research uh maybe for the CDC I talked to uh was saying that like they basically just focus on LAX because that's like the biggest hub of virus.
1:50:16
It's going to freak everyone out who flies.
1:50:18
Uh it's like if you're going to get sick, like if you go to LAX, that's where it all starts. It's like a petri dish.
1:50:23
Um, I know that they swab my hands for I think it's like bomb making materials, but should they be swabbing my hands for new pathogens?
1:50:30
Should we be doing more in the in the detection?
1:50:34
Should we be sequencing random farmland to see if there's new invasive pathogens that could be targeting our crops?
1:50:39
Uh, like like how and then like how do we even pay for that?
1:50:44
Is that something that the the government should put the bill or or uh corporations should be incentivized to pay for?
1:50:49
Uh what does the actual like upping the amount of data that we're ingesting look like?
1:50:56
Yeah, I mean I'm I'm always gonna say more data is better. Yeah.
1:50:58
But but the question is like yeah, how do we how do we get more data?
1:51:02
How do we incentivize more data?
1:51:04
Buy more data uh you know test for more data. Yeah, absolutely.
1:51:05
So I think there of course we should collect more the type of sequencing actually helps drive down you the cost because you can target a wider range of things you might worry about rather than setting up individualized programs for specific pathogens when maybe that's not a threat.
1:51:24
But I actually think the with the caveat of more data is always better.
1:51:29
We actually do collect a lot of this data today.
1:51:31
You don't necessarily see it because some of it is in wastewater or environmental samples like you're saying.
1:51:36
um we don't always extract that much intelligence from the data that we do collect.
1:51:39
So we we know that a sequence exists.
1:51:42
We might not necessarily know what that means in terms of health or the impact of that variant or that mutation.
1:51:47
Um so I do think there's there is benefit and there's like cost-effective benefit for collecting more but a big piece is just on of the data that does come in.
1:51:58
How do we build the right models and the right software to interpret that? Makes sense. Makes a lot of sense.
1:52:04
Uh, I think that covers it for now.
1:52:06
I would love to have you come back on when you're ready to talk more about uh specifics on on what you're building.
1:52:12
And I feel grateful that you and the team are doing what you're doing and just work a little bit faster, please.
1:52:21
Um, but uh, thank you for coming on and and giving us some insight here and uh, congrats on uh, on starting the the founder path. Thanks very much, guys.
1:52:31
Great chatting and hope to chat soon. Cheers. Thank you.
1:52:33
Next up, we have Roy from Cluey coming back for an update.
1:52:39
He's hired 50 interns, I think, or something close to it.
1:52:42
He said they're bringing every intern on.
1:52:43
They're bringing every intern on.
1:52:45
We got every intern coming in.
1:52:47
Well, welcome to the studio, Roy. How are you doing? Oh my. Let's go. There they are.
1:52:55
I think we're overpowering you.
1:52:57
Can uh can can you hear us? Yeah. Yeah, we can hear you. Yeah.
1:53:02
Make sure we're zoomed out all the way so we can see everybody.
1:53:03
We got a small army here. This is incredible. Uh, how big is the team? Kick us off.
1:53:07
How many you got at this point?
1:53:09
The team is 11 full-time plus the interns.
1:53:11
How many interns you got so far? Interns, bro.
1:53:16
We're closing in on 50, brother. Let's go. 10 there. That's amazing. Congratulations. What are they all doing?
1:53:24
How do you manage everything?
1:53:26
Uh, is it just is it is it purely social media?
1:53:28
Is that what you want them to focus on? growth. Yeah. Yeah. Growth marketing.
1:53:32
Like the only goal of the company is get 1 billion eyeballs onto Culie.
1:53:35
So, we have unrestricted creative freedom and permission to do anything and everything.
1:53:42
Uh just just make the company go viral.
1:53:44
Every single person you see behind you has over a 100,000 followers on some social media platform. Wow. 200,000 plus. Wow. That's remarkable. Me too now. Me too.
1:53:52
Now, yeah, you're up there. Oh, yeah.
1:53:54
Yeah, you probably popped. Uh uh what's working?
1:53:59
What platforms have actually been uh been driving the most growth?
1:54:00
I mean, I'm sure you've run a lot of tests.
1:54:03
What have you learned that's uh that you can share, bro? Ben, take it away, bro. Let's hear it. UC has been really good.
1:54:10
We just hit 10 million views today.
1:54:11
Um 10 million views on eight days. Wow. There we go.
1:54:13
Hoping to get 100 million views in the next month.
1:54:18
What what what platforms specifically are the most fertile ground for uh targeting your specific customer?
1:54:24
because you can imagine that uh there's a lot of folks who are AI curious on X but then there's a much broader more viral audience more general audience on platforms like Tik Tok, YouTube, Instagram.
1:54:34
What's working and what is uh is the next next platform that you're going to be focused on? Yeah.
1:54:39
Well, we're trying to go viral on every platform regardless.
1:54:42
Um but the main thing right now is Instagram reels.
1:54:44
Oh, Instagram reels in interesting.
1:54:46
And what is the main value prop that you're hitting people with?
1:54:50
Is it still the cheat on test thing or have you evolved at all? What?
1:54:54
Still actually interviews? Yeah, interviews. Okay.
1:54:56
And uh has there been I This was controversial when you launched it.
1:55:02
Is it still controversial in the comments? Are you getting flamed?
1:55:05
Has anyone big dunked on you?
1:55:05
And has that driven virality?
1:55:09
Is that actually a net positive?
1:55:11
Instagram is not like Twitter.
1:55:13
Like you could post the craziest on Instagram and they will still will not think it's controversial.
1:55:18
So to make it controversial, like we have to engage in bait some other way.
1:55:22
Like it's cheating tool is controversial on Twitter, but on Instagram, you could you could have like a white guy say the N word 10 times and it's still not controversial enough.
1:55:27
Like like you need crazy on Instagram. That's what we crack.
1:55:32
Every single person here has like very great viral sense. Yeah.
1:55:33
And you watch the reals that do go viral.
1:55:35
You see there's like ways that we've engagement beta the videos and this is what we'll keep doing to uh probably a billion views a month is probably how how long does it take to figure out if an intern is cracked?
1:55:44
Is it like an hour, two hours?
1:55:46
Just how much time do you need?
1:55:48
For me personally, me personally, probably like 10 minutes, but for anybody watching, probably would take like one one or two weeks. There we go. There we go.
1:55:55
Uh, how do you guys think about how do you guys think about product marketing?
1:55:58
Obviously, you're just going viral everywhere, getting all this attention.
1:56:00
How do you make sure that it that it uh I love that he's shaking his head on doesn't think about it's not about the product, it's about the attention.
1:56:12
Attention make anything go viral. Yeah. Um, yeah.
1:56:15
But but but how do you the side of the street, you know, you uh make some UGC videos, make some Twitter posts, you know, you could sell anything.
1:56:22
You know, in 2025, product doesn't matter.
1:56:25
You know, I could jack off off the side of a building, sell some videos of it for 20 bucks each, make $2 trillion. It's crazy. Two trillion. That That's intense.
1:56:32
Um how do you guys think about burn? Uh is it on your mind?
1:56:36
Is it on your mind at all?
1:56:36
I don't know if you saw the last tweet, but as of literally like two days ago, we're still we're still cast positive. We're still profitable. We're still profitable.
1:56:45
Let's give it up for the profit. Let's hear it. Let's hear it.
1:56:48
Uh it sounds uh so you're charging for the product and people are paying.
1:56:52
Are they at all satisfied or do they feel uh like they got satisfied, bro? Like the product works.
1:56:56
You're either using this as a consumer and it's working cuz like like you're passing your interviews and or if it doesn't work, you're not going to complain to me because I'm going to go right to your employer and tell them, "Yo, guess who's complaining about using the product?"
1:57:08
Like I will get you blacklisted if you complain really.
1:57:12
Um, where how are you thinking about how do you how are you guys thinking about product evolutions?
1:57:15
What do you want to add to the product?
1:57:16
Obviously, uh, you want to help people cheat on everything.
1:57:20
Where where are you going to help people cheat next?
1:57:22
We don't care about like the product is going to be led by the virality of the content.
1:57:26
We have video ideas right now that we're going to try to push for different use cases.
1:57:30
We're going to see which ones go consistently the most viral.
1:57:34
If you can make something go more viral, then like you can just build the technology after you have all the attention.
1:57:38
So, we'll figure out the exact use cases and exact niches we're going to quintuple down on once once these guys get to work.
1:57:44
What what uh what formats on Instagram reels are like the most modern in terms of uh consistently viral?
1:57:52
Like you mentioned like man on the street interviews.
1:57:53
What do you do for a living?
1:57:54
That's always been a fertile ground.
1:57:56
What about uh I see a lot of those like mobile game ads that look like you know you're fighting down some sort of bridge and then you go into the game it's actually just uh match three.
1:58:06
Um what what what are the different formats that you like to pull from?
1:58:11
Every week there's two new ones and at any point there's probably 10 to 20 viral trends that is happening and these cycle so quick you need to keep your finger on the pulse these things will like expire immediately.
1:58:22
You need to be on the ball and like like if I if I told you right now by the time people watch this on YouTube like it would have all been expired.
1:58:28
Well, we're live so so give us give us the latest and greatest like what's going viral today.
1:58:33
Well, right now we got 10 million 10 million views using a Snapchat format which viral for like the last three years to be honest. Okay.
1:58:42
And and I I I think that like we just have to get people who continuously scroll TikTok like six hours a day. Yeah.
1:58:50
But what's the what's the actual format that you use? Like describe the video. What is the hook?
1:58:57
Like break it down for me like you're explaining the like the art behind the viral format. There's a caption. It starts with a face.
1:59:04
Usually a handsome dude or a pretty girl.
1:59:06
They're saying, "Damn, this interview is starting with the interviewer starting with the hard questions.
1:59:10
I should have been a CS major, not a business major."
1:59:12
That's engagement because people are saying like, "Bro, like CS is way harder than business."
1:59:16
Then it turns around the interviewer asked like like, "Hey, how are you doing? Why should we hire you?"
1:59:21
And then this guy uses Cluey to generate a response, but he can't read the response.
1:59:24
So he reads it hella autistically like, "Oh, I re revel in detail."
1:59:28
And and then that's that is like another conversation point like people are cooking on the guy because he uh he can't read properly.
1:59:32
The guy is like a doing a really dumb interview using That's great. That's great.
1:59:37
How are you guys using uh AI generated content internally?
1:59:41
I know a lot of these the videos that you guys are creating are just typical social media vertical video.
1:59:48
Do you have an intern that's just generating uh basically copy and pasting making a bunch of other V3 or any of these tools rev anything clicking? Not not yet.
1:59:56
I think there's still like a 10% left before they cross the uncanny valley.
2:00:00
And the biggest thing is that people need to think your video is real.
2:00:03
That like like that is the difference between 100k views and 10 million views if people think that's real.
2:00:10
Totally AI CEO is bearish on AI.
2:00:14
What about uh what Google needs like 10 more Chinese researchers to like figure it out and once once they push out the latest update then then then BO3 will be there.
2:00:23
But right now we need real people. Yeah. Yeah. Yeah.
2:00:25
Uh, well, I I mean, what about just using AI as like stock footage replacement?
2:00:30
Not not as the leadin for the video, not the entire video, but just like sprinkled in to illustrate a point, you know, an establishing shot of like a building, a helicopter pulling into a building.
2:00:40
Like that that historically has been kind of something that you would reach to uh, you know, Adobe Stock video for.
2:00:47
V3 feels like it's there, but are you not drawing on that at all yet?
2:00:52
If there's a viral format when we need it, maybe we'll use it.
2:00:56
But right now, like it's it's really brain dead to go viral on Instagram. Yeah. Formats are not hard.
2:01:03
You don't need a helicopter.
2:01:03
You need a guy, a camera, a really shitty camera, and you need a computer. That's it.
2:01:06
I mean, what about like those those kind of like AI mashups like Harry Potter Balenciaga or the uh the kangaroo with the plane ticket getting on the plane like like AI content can go viral when it's really uh when it's like inspired almost by a human.
2:01:21
It's not it's not entirely AI generated, but it's using the tools effectively to create something that's like still catchy.
2:01:26
Uh do you think you'll be using any of that anytime soon? Probably very soon. We're scaling up.
2:01:33
Like what you see right now is probably about less than 1% of what the size we will be by the end of this summer. Like we are profitable.
2:01:38
We're not trying to be profitable.
2:01:40
We just keep making so much money. We can't help it.
2:01:43
So we're really scaling this up to I'm not even trolling you.
2:01:45
1,000 creators are going to be shipping out content.
2:01:49
We're doing a complete internet takeover. Okay.
2:01:50
So, so, so, so, so why in house? Why? Like, like like what?
2:01:53
Why do they even have to be employees? What?
2:01:57
Couldn't you turn this into like a multi-level marketing scheme or something?
2:01:59
A pyramid scheme where we're going to do this again.
2:02:01
Oh, that's what you're going to do. Okay. MLM MLM.
2:02:06
Are you guys worried that you could be infiltrated by journalists?
2:02:09
I'm sure they're circling the house right now.
2:02:12
The hit pieces are going to come.
2:02:12
You know, we're we're doing a softball interview right now.
2:02:15
I mean, the person who's brave enough to try to do a hit piece on the Culie Army is It's going to be scary.
2:02:22
I bet they're dying, too.
2:02:24
Look, more more eyeballs is better.
2:02:27
There's no companies that ever died from a founder being too controversial.
2:02:28
You got deal infiltrating with genuine spies and they're still doing fine, bro. You got worker 17, guys.
2:02:36
They're they're they're still kicking.
2:02:38
Like, no company ever dies from being too controversial.
2:02:39
You die because you don't make enough money. Yeah. Yeah. Yeah. Yeah.
2:02:43
Uh speaking of making money, uh what's the pricing model right now?
2:02:47
Are you doing anything on price discrimination?
2:02:49
And is there a super high tier?
2:02:50
If you get a whale, what does a cluey whale look like?
2:02:52
Can I spend $2,000 a month on this service?
2:02:54
Yeah, you should add a tipping feature, too.
2:02:58
People should be able to tip you guys if if they have a good experience and they get the job.
2:03:01
Like really financialization, pay as you go, high interest rate loans. Just really push it.
2:03:06
Make it sports gambling in there.
2:03:07
Maybe just throw it all in. Yeah.
2:03:07
I mean, it's $20 a month for consumer, $100 a year.
2:03:12
And and our our topline revenue is really being driven up by enterprise. enter.
2:03:16
You're gonna have to talk to your sales team to get a custom pull.
2:03:19
But you know, like there's a lot.
2:03:22
Wait, are you are you serious?
2:03:22
What what what are the is that more on the sales side?
2:03:26
What who are the enterprise? So you sell the SDRs.
2:03:28
You guys laugh cuz you think I can't sell enterprise cuz I'm No, I I don't believe it.
2:03:32
I trust like these 20 these Fortune 500 CEOs.
2:03:37
Like these are like 35 year old dudes who sit there scrolling through Twitter laughing at my post. Yeah. Yeah. Yeah. Yeah.
2:03:42
No, it seems seems legit. It makes sense. No, I I I believe it.
2:03:44
But but I mean you're not going even higher tier.
2:03:48
Like what's the $2,000 a month clearly vision for for consumer?
2:03:50
There's a lot more we can do with more compute, but right now we're like to be honest, I didn't expect to grow this fast. The team is quite small.
2:03:58
I'm like spent a lot of time trying to hire more more competent engineers.
2:04:02
We have a lot of backlog tasks that we need to fill out, especially for this last contract that we signed.
2:04:06
So we're full-time focusing on the one big guy that we got right now.
2:04:09
And after that um the then we'll we'll try and scale this up.
2:04:12
But right now, we're focused on the one one big client that we signed. Yeah.
2:04:15
Uh talk about your compensation strategy. The people want to know.
2:04:19
Uh you you said uh you can raise infinite capital and you're so confident. I believe you.
2:04:24
Uh but but I'm curious to to get some more insight there.
2:04:28
Bro, I feel like it's so to be a company.
2:04:31
Sorry, am I allowed to say that? No, you're not allowed.
2:04:34
No, this is a family-friendly show.
2:04:36
It's very stupid to be a company like trying to race to the bottom to see how little you can pay your employees.
2:04:41
Bro, if I'm making hella money, we're all making hella money.
2:04:43
Like, like it's I'm trying to pay them more to see if man, like maybe tomorrow we'll start being like cash flow negative, but make money, bro.
2:04:52
Like, I I would like to pay these guys what they're worth and the output is insane.
2:04:55
We did 10 million UGC views in what, like eight days?
2:04:59
Like, like you don't see this sort of traction in any company and you don't see killers like this in any company unless you paying these like what they're worth, bro.
2:05:05
Like, I don't know what happened to like 13 maxed out contracts. Maxed out contracts. Exactly. Exactly. Yeah. Uh what about devices?
2:05:13
I mean it seemed like this would be a natural fit for some sort of AI wearable or other platform.
2:05:18
Um is there an app coming or are you interested in what's happening with Johnny IV and OpenAI?
2:05:23
What's what's your take on the device world?
2:05:26
We're very interested in the hardware space.
2:05:30
We've got like a million things cooking on hardware.
2:05:32
We got people in the garage right now working on you don't you don't even know about, bro.
2:05:36
like like like we're we're bringing manufacturing back to America and it all starts at the Kulie garage. Let's go. I love to see it.
2:05:46
Nobody, you know, they doubted, but you guys are re-industrializing America.
2:05:47
You guys really are the brain down there.
2:05:51
They're working on brain chips down there. Brain chips. That's the future. There we go. There we go. The new Neuralink. Yeah.
2:05:57
I mean, I I you know, there's a world in the future where you guys actually just roll up Neurolink and OpenAI.
2:06:02
Bring them under the fluy umbrella. Definitely. It's possible.
2:06:04
I'm excited to offer acquisitions for for both of those companies. It's in the road map. It's on the road. All right.
2:06:12
This has been a lot of fun.
2:06:12
I'm excited for you guys.
2:06:14
Uh it is uh and uh I have no doubt that you'll go from, you know, 10 million views a week, 10 million views a week to 100.
2:06:21
Uh and I'm excited to see you guys hit that billion uh view mark very soon. So, keep it up.
2:06:26
We are all very entertained and uh rooting for you. Sh. I love the energy. Thanks, man.
2:06:33
We appreciate you joining. Later, guys. Keep having fun. Bye. You know what he needs? He needs linear.
2:06:39
He needs linear to manage all his interns. Sounds chaotic in there.
2:06:43
Needs to use the platform purpose built to design and build the best products on earth.
2:06:48
I bet they already I bet they already use linear to uh meet the system for modern software development, streamline issues, projects, and product road maps.
2:06:56
Go to linear and get started today.
2:06:58
Um, I wanted to go through the uh the Sarah Guo piece because we are having uh Sean on because he is throwing the AI engineer world's fair and he's going to be joining with some some other folks uh breaking down what's going there.
2:07:13
Um we can we can kind of revisit this Sarah post which we talked about later.
2:07:18
Uh possible topics for her keynote.
2:07:22
Apparently she completely revised everything but I really liked it.
2:07:24
She wanted to talk about AI native UX, not just chat skins, vertical AI surge, that was something we were discussing earlier.
2:07:32
Um, multimodal frontier video 3D audio uh retrieval plus long-term memory, synthetic data flywheels, obviously super important.
2:07:39
obviously super important. The sim tore real robotics push uh closing the autonomy gap robust agentic workflows cell scale digital twins and programmable biology compute geopolitics world models for zero shot planning and RL environments that actually generalize
2:07:54
generalize she has mapped out the like the the true surface area but more recently she uh authored a post an article on X uh directly on X you can read the full thing at her uh at her X page senormis on taste and I Thought this was an interesting post. Lulu uh
2:08:10
Lulu uh quote tweeted and she says, "Stripe returns errors in plain English."
2:08:14
Quote, "That card number doesn't look right, not error_invalid parameter because developers debug at 2 a. m.
2:08:22
Spotify's shuffle isn't random.
2:08:25
It avoids playing the same artist twice in five songs.
2:08:28
True random feels feels broken.
2:08:31
Engineered random feels right.
2:08:31
Notion's drag handle appears only on hover.
2:08:34
Six dots arranged in two columns.
2:08:37
Not three lines not always visible because permanence is clutter and six dots whisper, "Grab me."
2:08:43
While three lines shout, "I'm a menu." I love it. This is taste.
2:08:48
The relentless, almost painful ability to know what should exist, what shouldn't, and where quality matters.
2:08:55
It's the difference between shipping a product and shipping a point of view.
2:08:59
The best founders understand that taste is a competitive advantage that compounds.
2:09:02
It runs deeper than pixels.
2:09:04
It's in your codebase, your culture, your cap table.
2:09:06
And I'm just thinking about the clue lead taste, which is just like the most maximal possible. Just maxing everything. He's he's maxing. He's maxing. It's true. He's such a character.
2:09:17
I love that we can have him on and then have someone, you know, like, you know, public company CEO on the same. We got range. Yeah, we got range. It's fun.
2:09:24
Um, that was absolutely wild.
2:09:27
He um uh we should probably text the families and say, you know, don't play this in front of the kids.
2:09:33
But um but uh he knows, you know, the level, you know, he he even in every word, he's thinking of how do I get the max amount of attention out of each incremental word.
2:09:44
It's not just each incremental post.
2:09:46
He knows what he's doing.
2:09:46
Even I am like s systemically offending people in multiple ways in a single sentence. Yeah.
2:09:54
But at the same time, like he's entitled to be his own personality. Yeah.
2:09:58
And and I do think, you know, not to get completely sidetracked now that we're on to the next topic, but uh I do think, you know, an enterprising young journalist might want to take a crack at Roy just because it will turn into a feud.
2:10:14
Y it will probably be good for both of them.
2:10:16
It will be very It's WWE.
2:10:19
It's WWE paywall the article.
2:10:19
He's become a little bit of a heel of tech.
2:10:23
Like he's somebody that people like love to hate.
2:10:25
No, it it's like how do you take the YC playbook and just run the opposite of it?
2:10:29
Like it is it is the inverse of the YC playbook which is you know be loud before you're confident in your business model and confident in your product.
2:10:39
Um the question is just like it's getting yeah it can't all be style.
2:10:44
There needs to be some substance.
2:10:46
You have to actually build a product that people will be that people will love.
2:10:49
you have to make something people want.
2:10:51
And so, uh, we're not seeing that side right now, obviously, at least not in in public interviews.
2:10:55
Um, but I hope that he can turn it off incrementally and and eventually you probably need to turn it off 90% of the time because yes, attention is really really important right now, but at a certain point, you have to just deliver on the on the core metrics and the value prop or else someone else will come along and and and offer something with with Yeah.
2:11:13
They won't have all the all the style, but they will have the and I I I I think there's a venture capital firm out there that even now would give him an increment just seeing that interview see that and be like, I'll give this I'll give this team another 10 mil. Yeah.
2:11:30
To go keep figuring it out, right?
2:11:32
I mean, attention is incredibly valuable and but he's also but he's also being smart and that he's figured he basically has the world almost convinced that he's going to burn through every dollar he has in the next two months. Oh, totally.
2:11:46
Which is pretty but behind the scenes he's like, "Well, I'm actually making money." Yeah.
2:11:51
I mean, obvious Shipman at Friend was going through a similar thing where there was that big story about, oh, he raised like a $2 million seed round and he bought a $1 million domain and everyone was like, "Oh, he's so wasteful."
2:12:01
But then, of course, he would like finance the domain.
2:12:02
And so it was really just adding like an incremental like 30k of burn per month and it wasn't really like he didn't take that whole hit upfront and then he was able to kind of build the business.
2:12:11
Yeah, he's making a category bet on AI companionship and having friend.
2:12:16
com is a great is will lead to increased marketing efficiency over time as he starts to scale.
2:12:23
So well anyways we should go back to this completely.
2:12:28
You know what he should do?
2:12:29
He should get a he should get a watch on bezel. Go to getbessel. com. Really flex.
2:12:35
Your bezel concier is available now to source you any watch on the planet. Seriously, any watch. So go to getbzel. com.
2:12:42
Um, so Sarah goes on to say, "Think of it like running a restaurant."
2:12:44
It's so funny comparing the taste thing to this.
2:12:47
It's like the most serious post.
2:12:47
There's the most silly uh just silly hype train over there.
2:12:53
Uh, think of it like running a restaurant.
2:12:54
Anyone can follow recipes, source ingredients, and serve food.
2:12:56
But the difference between a forgettable meal and a Michelin star isn't just technique. It's the chef's pallet.
2:13:01
Their ability to know when something needs more acid, when a dish has one element too many, when to stop plating. Software is the same.
2:13:09
But products aren't just feature complete. They're composed.
2:13:11
And we see this with like uh a lot of the best software products.
2:13:16
The software is more like art than than science in terms of like the design the design and how you're how you're um interacting with the user.
2:13:23
Um, so Sarah says, "It's easy to say, hard to do.
2:13:26
Everyone claims to have taste now it becomes the new uh product market fit.
2:13:31
A term so overused it's lost meeting. Founders drop.
2:13:34
We're taste driven in pitch meetings. VCs nod knowingly. Nobody defines it.
2:13:38
Most companies confuse taste with aesthetics.
2:13:43
They hire a design agency, pick a nice font, and call it done."
2:13:45
But real taste runs deeper in the error message, the loading sites, the the loading states, the features you killed because they were merely good, not essential. Real taste hurts.
2:13:55
It's saying no to features that would triple your TAM.
2:13:57
It's spending a week per of It's spending a week perfecting an interaction that users will barely notice consciously.
2:14:05
It's choosing the harder technical path because the UX is 10 10 time 10% better.
2:14:11
If your taste doesn't cost you something, it's not taste, it's preference.
2:14:13
The pain compounds daily.
2:14:13
A Fortune 500 prospect wants a demo tomorrow.
2:14:17
Do you ship a halfbaked feature or the half-ass collateral or lose the deal?
2:14:21
The entire AI landscape reshuffles every 3 weeks.
2:14:25
Do you chase every new model or trust your vision?
2:14:29
Your competitor just shipped something flashy.
2:14:30
Do you match it or hold your ground?
2:14:32
Anyway, we'll have to have Sarah on the uh on the show to recap her talk.
2:14:38
But right now, I believe we have Shawn coming in to the studio to tell us about the AI engineering worlds fair.
2:14:43
We're very excited to have him join the show.
2:14:47
Thanks so much for hopping on. I wish I could be there.
2:14:49
I have the biggest FOMO I've ever had with any tech conference because the lineup seemed absolutely fantastic.
2:14:54
Unfortunately, I have a large family and and a studio that is covered in junk and we're moving in.
2:14:59
So, I appreciate you taking this remotely and we'll have to do the next one in person, but thanks so much for joining today. How are you doing? Hey guys. Uh glad to be back.
2:15:08
Um I It's funny cuz like your setup looks so great on camera.
2:15:10
I'm just imagining the mess that's off off camera. Oh, yeah. Yeah. Yeah. Yeah.
2:15:14
It's definitely a work in progress.
2:15:15
Uh but part of the brand is of course just showing the nice parts.
2:15:19
Um but but but congratulations.
2:15:22
Uh this is a many years in the making.
2:15:24
Can you give us a little bit of the the history, the plan, and then I want to go into some of the hot topics that you've been discussing today. Yeah.
2:15:30
Um I guess this is the fourth conference we've done.
2:15:33
uh we've done um and I started this basically uh right after the sort of chat GBT moment and talking with enough developers and understanding that the people who can wield LM APIs are going to be way more powerful than people who just chat with products and this is actually going to widen a lot because they can basically wield serverless intelligence.
2:15:56
Um and so I coined you know I sort of popularized the term AI engineer.
2:16:02
Andre Karpathy sort of endorsed it.
2:16:03
endorsed it. He said that um he does believe that you can get very far without ever training anything which is a big thing for him to say because I think the status quo at the time was you know it's high status to train models
2:16:15
research scientists but now it's actually consensus now that you want to work on rappers rather than models y and uh there are many many multi-billion dollar companies that um have spoken at AI engineer that reflect that fact and
2:16:29
yeah so like I guess and there was a big shift there where where there was there was a moment where uh fun like the vibe had shifted a little bit to like the application layer, but there was still the idea that if you were going to build
2:16:41
an AI legal startup, you were going to train an AI legal foundation model and now it's moved all to post- training, all to RL and how you're how you're prompting and how you're integrating and ultimately like the user interface. And
2:16:53
And so, yeah, it makes a ton of sense and I'm sure there's a ton of companies that are beneficiaries of that. Boom. Yeah.
2:17:01
Um, and it's not even companies.
2:17:01
It's more just like customers. Sure.
2:17:03
Like because the foundation model labs are never going to work directly with like your healthare system.
2:17:11
Like it's going to be like a bridge that um that comes up and and does that.
2:17:15
They're not the foundation model labs are not going to work with directly with lawyers.
2:17:18
It's going to be Harvey, you know.
2:17:20
I think so that that directly led into the rise of vertical AI enabled SAS.
2:17:23
And I mean that's consensus now. But yeah, I'm sorry. Right.
2:17:27
That's a Slack thing that by the way the conference is still going on.
2:17:32
We just had the morning keynotes.
2:17:33
Uh I have my chat with Greg Brockman later today. Fantastic.
2:17:35
And he doesn't know this, but I can I might as well just share this.
2:17:39
But like we have a nice little cameo from uh Jensen Huang coming by. Um no way. That's great.
2:17:43
So like we're trying to level up.
2:17:47
Uh the c the keynotes this morning were fantastic.
2:17:50
Like the Microsoft, you know, they they like really went on.
2:17:52
They're they're going so hard after the AI audience.
2:17:55
We were just reading about that in the information that platform platform platform platform platform says Satia. Yeah. Yeah.
2:18:00
Um I think they see this as their chance to overtake AWS. That's great.
2:18:05
But like AWS is also, you know, also sponsor and like and having a very strong presence with us.
2:18:10
So like uh we just want to be the vendor neutral uh place, right?
2:18:13
Like all the big clouds, all the big labs work with us.
2:18:15
We have the the MCP team here with the entire steering committee presenting as well.
2:18:21
Um, and we just wanted to be for developers like this is where you this like kind of the trade show you come to hire people, learn about what's new and upscale.
2:18:29
I have a couple more questions.
2:18:32
How long do you have exactly?
2:18:32
Five minutes or I can go till like uh 1:30. So Okay, great. Yeah.
2:18:37
So we had we had a guest and we we I moved into a different day. Okay, first uh perfect.
2:18:41
So um I want to know about the mix of attendees.
2:18:47
How many uh how many folks are are trying to start venturebackable application layer AI companies versus is there a new trend of someone who's building more of a vibecoded almost lifestyle type uh AIdriven business?
2:19:01
Uh are there folks from either companies that have established themselves and are now trying to bolt on AI or are there lots of folks that are uh working for large companies and just want to stay ahead and become AI engineers?
2:19:15
What's kind of the shape of the audience if you can characterize it at all? Yeah.
2:19:19
So, uh we we do surveys but we don't we don't know specifics to to to that high level granularity.
2:19:27
I would say about 50% are are people working at um medium to largesiz companies and trying to upskill and then the others are uh smaller companies and and you know our most popular title is like founder or CXO of like a a smaller startup and those are those are venture backable and I think mostly that's just a function of us being in San Francisco because obviously we we we will have that startup bias.
2:19:51
Um I do think that um one thing I I kind of don't really care about this whole lifestyle versus ventureback thing because uh for example I have a tiny teams track that speaking this year.
2:20:04
Tiny teams is something that I'm trying to push as an idea um of u companies that have more millions in ARR than employees. Mh. Right.
2:20:10
So your your revenue efficiency is so high because obviously if you pay each employee less than a million dollars, you're probably profitable and therefore you don't actually need the venture money except to pour into marketing.
2:20:20
Uh, and it's your choice. You can be profitable.
2:20:23
I have a sixperson team uh making more than more than $40 million.
2:20:28
Um, and uh, yeah, I mean it's it's absurd like the amount of leverage you can get with with uh, agents and also building AI products for other people to use.
2:20:37
You're where you're just kind of passing through or slapping a margin on top of uh, the tokens that you resell from the big model labs.
2:20:45
I think that really makes a lot of sense. Totally.
2:20:47
uh has the narrative of like, oh, if you're building an AI, like you're going to become uh obsolete by Google or or or or open AI is going to steamroll you in their next Dev Day keynote.
2:20:58
Has that narrative dissipated?
2:21:00
And what if so, what's driving it?
2:21:04
Is it kind of the pre-training scaling law wall that we're kind of seeing with GPT 4. 5?
2:21:12
Are you guys still I I don't know if people still talking about that.
2:21:13
I mean, everyone, you know, everyone's moved on.
2:21:16
Yeah, we've moved on to different time, right?
2:21:18
Like, yeah, I think I think most people have agreed like there are still new pre-trains happening, especially with the open source model that OpenAI is working on. Sure.
2:21:25
But yeah, I mean, I think we've just like we've seen it come and go enough times.
2:21:30
So, for example, we have OpenAI launching chatp codeex, which is just head on a direct Devon competitor. Yep.
2:21:39
competitor. Yep. But um you know Devon's not worried because they you know have been doing this for two years and they're much more polished in terms of the integrations and they have different things that their customers already like and so I think it's just like everyone
2:21:51
is going to need their version of a thing and so this is the sort of house storeband version of what CHG codeex could be and it's it's not competitive just because the the ocean is so huge for software engineering and Devon has at least established the category by being first there. Um and I think you
2:22:05
Um and I think you can see similar versions of that across across the the domain.
2:22:09
Um yeah, I haven't I haven't really seen anything there yet.
2:22:13
Um although it's not to say that it doesn't happen. It does happen.
2:22:17
For example, with the first wave of GPT3 startups like Jasper um uh but it I mean yeah so far there's no fear that in fact people very excited to meet the foundation model labs like this is where the engineers meet the lab people um and lab people train them on how to use their LMS.
2:22:32
I think that's that's a perfectly harmonious relationship to be honest.
2:22:35
I I haven't seen What was your reaction to the the news around anthropic and wind surf yesterday?
2:22:39
Uh a little bit of drama on the timeline, but so yeah. Yeah. Yeah.
2:22:45
Uh give us uh give us a walk through for those who might not be familiar of what happened and then I'd love your analysis. Yeah.
2:22:51
So the history of this is that um Windsurf is an independent company that uh basically kind of followed Kurser's footsteps and launched an AI agentic.
2:23:04
They've done very very well for themselves in a very short amount of time.
2:23:07
I think they launched we're the first podcast that they launched with in October or November last year.
2:23:09
Uh and and then they there's been rumors that they got acquired by OpenAI for $3 billion.
2:23:15
Uh those are rumors that are unconfirmed by both sides.
2:23:18
I've talked to both sides.
2:23:20
Yeah, it's it's I've been fascinated how everybody just takes it as fact. I take it as fact. He wore two polos.
2:23:27
Nobody Yeah, he wore two polos.
2:23:29
But again, at no point every every news outlet, legacy news outlets have been reporting it as fact, even though no side has verified it and nobody said it closed or anything along those lines.
2:23:41
Yeah, the the only thing I can say is um I Windsurf is speaking tonight right at right before Greg Brockman goes on.
2:23:47
Um, and there's a reason. Um, exciting.
2:23:52
We're not like we're not dropping a ton of alpha, but I'm just saying like I'm not I that's all I can say.
2:23:55
That's all I'm allowed to say. Yes, of course.
2:23:57
And I I think that so there's there's there's a there's a relationship there just as there is uh a fact that both cursor and windsurf have benefited a lot from their relationship with Enthropic because cloud 3. 5 and 3.
2:24:10
7 have for whatever it's worth been regarded as the best coding models. Yep. Open will disagree.
2:24:19
Gman will disagree but whatever like the community has voted. Yeah.
2:24:22
Um so overnight I think like two days ago one day ago uh anthropic cut off the first party access to claude to uh for cloud to windfur this is their top model and now they just don't have access as a as a first party tool.
2:24:34
You can, for example, still bring your API key and use your accounts on Windsurf, but it's right.
2:24:40
You can't just like like Windserf, I'm gonna pay you 20 bucks.
2:24:43
Use your account, whatever.
2:24:43
I, you know, I just don't want to worry about the the the rate limits and stuff. Um, that's gone.
2:24:49
Windserve, they just woke up overnight and and that's gone.
2:24:51
So, a lot of people are very upset.
2:24:52
Uh, this is a very big no no if you're like aiming to be any sort of like credible LM API provider to just cut off access.
2:24:59
provider to just cut off access. um Google hasn't done that even though you know if you take the the sort of rumors of the acquisition on face value um but I think this does lead credence that I think uh anthropic at least thinks that Windsurface is just a competitor product
2:25:15
now so I think like in so far as you're putting odds on whether the acquisition actually closes the odds have moved up interesting yeah the other the other factor that's interesting is how much does this benefit cursor if you're a developer who loves various cloud models and Now maybe you have a reason to go spin up. Do you know market share of
2:25:33
Do you know market share of cursor versus windsurf?
2:25:35
Have you seen like estim estimations? What is it?
2:25:36
I I think the numbers I've seen again I I really I saw it on on my timeline on Twitter.
2:25:44
I didn't save it and press but someone please look this up.
2:25:46
It's something like Windsurf is 5% of cursor. Wow.
2:25:50
Because based on the based on the market caps I would assume that windsurf 30% I know of cursor.
2:25:54
So the thing that you don't see there is cursor has entirely developed business on the IDE where Windsurf used to be codium which used to be like this GPU serverless GPU company that has significant LLM inference uh for enterprises like they they spent the last four years doing that. Yeah. Yeah.
2:26:12
Um so they're buying that team that revenue that product as well as Windsorf. Oh interesting.
2:26:16
Yeah that makes a lot of sense.
2:26:18
Um what uh how do you think that the uh agentic coding market is shaping up?
2:26:23
coding market is shaping up? we've been kind of talking about maybe it's three different markets like I can go to GPT uh like I can go to 03 and I can ask it for a question and it will just write code and write Python and execute it and I don't even have to tell it to write code it's just if I ask a question that requires code it will write some code
2:26:42
then of course they openAI now has codec and then they also potentially have windsurf or or you know you can think about the idees uh as a different um as a different uh entry point into that market uh are we seeing like a true permanent bifurcation between uh synchronous and asynchronous uh AI coding usage or do you think these all blend together at some point? Yeah, I
2:27:04
Yeah, I think that actually you're just catching up on something that has already been a thing.
2:27:10
Um that was the divergence and actually they're converging.
2:27:15
So they they already like it already started that you had the synchronous ones like cursor windsurf and then you asynchronous ones like Devon and Factory AI for example. Yeah.
2:27:23
And they were all sort of like inner develop what we call the developer inner loop which is hands on keyboard coding which is developer outer loop which is uh PR review and all that kind of stuff or like file an issue make a PR that kind of stuff.
2:27:35
Um so those are merging the explicit goal of codeex is something I'm going to talk to Greg about uh later today is that they will merge those paradigms just that they're merging the reasoning models and the the sort of instant thinking models uh with GPT5.
2:27:48
Um, and I I that totally makes sense.
2:27:53
It's technically really hard to do because once in one's on your machine, one's one's cloud.
2:27:56
And also they're just different set of user experience paradigms.
2:28:00
Is my uh video freezing by the way? It is. You look great. You look great. I see you guys. Great.
2:28:06
Uh, how how seriously do you take Google's uh coding agent jewels?
2:28:09
Do you think it's just an experiment that they're putting out there or do you think they'll actually invest in it and try to get real adoption?
2:28:17
You're trying to get me in trouble. You're conflicted. You're conflicted.
2:28:25
I'm sure I'm sure Google has has is a part of Google's great product wise.
2:28:30
They have had hits like no LM that then have failed to continue the momentum for whatever that I and I think that is I think that most people are extremely unfair to notebook.
2:28:43
I think they have really well.
2:28:45
It's just that you will never repeat the initial wave of excitement that they had. Totally. Just never will. Like Yeah.
2:28:50
And I mean Notebook LM seems interesting because that feels like a case study in if you're a startup and you're fast following an innovation that actually came out of a hyperscaler.
2:29:01
That's where you turn into a bullet point on the next dev day or something like that.
2:29:06
Because Notebook LM it felt like there there was this amazing tremendous momentum online.
2:29:12
Everyone was excited and then it was kind of faltering and everyone was like, "Oh, there's no app."
2:29:15
Uh, so maybe I'll build the app and I'm sure some people built notebook LM apps or notebook LM spin-offs and then it just took it just took six months and then there now there's an app and now and I'm sure they will continue to iterate on this and so yeah.
2:29:27
Yeah, I I I know to some degree.
2:29:29
Uh but but but I I think the lesson is probably like yeah there's probably some narrow window where you can cash grab as kind of like a cynical like ripoff startup but like realistically if you if you found out about the concept of turning uh a deep research report into an audio product from notebook LM like
2:29:47
you might be just behind the ball and so like like like you don't you don't have the right to own that market long term as opposed to you know I think dev and cognition like they have more of a right and and and factory they have more of a to uh continue to fight with Google's uh Jules product because they were really there earlier. They can they issued the
2:30:04
They can they issued the original hype and and and brokered a whole bunch of enterprise contracts and like kind of got a couple years down the road before.
2:30:14
So they're not like a fast follower to big tech.
2:30:16
They're actually uh big tech is maybe fast following them. Uh yes.
2:30:20
I I think I think it just boils down to execution and everything, right? Yeah, of course.
2:30:25
And I think like uh that's that's one of the reasons why we've actually somewhat broadened out from just engineering to we've added AI product management and AI design. Okay.
2:30:35
And I think that really good this is this is product management.
2:30:37
This is just straight up can you keep up the m momentum?
2:30:40
Can you listen to your users?
2:30:42
Can you come up with creative new stuff that keeps the momentum going and some some teams can and some teams cannot.
2:30:46
Um the only thing I can say with regards to Google and like the fate of jewels is um the initial the founding PM of of Nom left and she's speaking here with her own startup.
2:30:58
Um it's like the the it's really hard to keep employees of people who like start interesting products for Epic Lab because they will get any number any amount of money thrown at them. Yeah. Open it up.
2:31:14
$20 million seed round, 10 million of secondary if you quit your job right now.
2:31:17
Uh probably not that extreme.
2:31:17
Um I I wanted to ask you about uh Yeah, cheap.
2:31:23
Uh uh I wanted to ask you about GitHub.
2:31:25
Uh how much do you think Satia cares about, you know, GitHub?
2:31:28
Uh do you think he's, you know, really pushing? Yeah.
2:31:34
Is that an important wedge into the AI coding market?
2:31:36
because obviously they were first with GitHub Copilot.
2:31:40
Um, and you could imagine that that it's just a phenomenal distribution channel.
2:31:45
And when you look at the the value of and the way Microsoft executed around Teams rolling into you you're already using Outlook now you're using Teams instead of Slack.
2:31:55
You could imagine really great adoption.
2:31:57
Maybe not at the most cutting edge companies that are hyper online, but we could see one of those charts where we're like, "Oh, wow."
2:32:05
like Microsoft's really crushing it in in developer tooling.
2:32:07
Uh I mean uh absolutely I I think it's extremely core.
2:32:13
Uh what would you put GitHub co-pilot revenue at right now if you want to if you had to hazard a guess? 500 500 million. Yeah. It's right there. You nailed it. Let's go.
2:32:22
Standalone that is a publicly listed company. Wow. Yeah. Yeah. Yeah.
2:32:27
That's like the perfect Here's the thing here.
2:32:29
I mean here's here's the challenge, right?
2:32:30
Here's the challenge is that no one is talking about GitHub copilot. Is that a challenge?
2:32:37
It's inside of Microsoft.
2:32:37
So yeah, I mean no one's talking about No, I know.
2:32:41
But you know, nobody's saying like look what you know there nobody's sitting there being like I'm blown away by Copilot, right?
2:32:50
I I think I think uh I think that people who live in the Microsoft stack um are because that is what they have access to and like their company agreements are.
2:32:59
I think on Twitter there's a novelty bias, right?
2:33:04
People people always want the new thing.
2:33:06
They want to they want to support the little guy.
2:33:07
It is I support the current thing.
2:33:11
You want to support the current thing.
2:33:13
But okay, here's the mental framework I want to leave you guys with on on that stuff is um data gravity and the mo is a thing like data has attracts the compute attracts the money attracts stickiness like once once my is one place everything else moves towards that data
2:33:30
and um if my code my code is the the most valuable form of data I have that's that's the most uh expensive uh data to acquire um so if my code already lives with you then it's much easier for my coding agents to also with you and uh and so so they they have a home turf advantage. I think the GitHub
2:33:45
I think the GitHub acquisition back in the day was one of the smartest and most enterprised acquisitions uh in in developer tools history.
2:33:52
Um so like I think I think you should you should take that point of view.
2:33:55
So like that that moat is so strong like any like try try to be like a no-name random dev tool startup that that comes up and says hey give me access to your codebase like like I need to run agents on it.
2:34:06
They're like, "No, like I I'm already doing Yeah.
2:34:08
What hap what happened with codeex yesterday?
2:34:12
There there was one user that was reporting that uh private repo information was leaking from one to another. Was that user error?
2:34:18
Did you did you track that at all?"
2:34:20
Now, you were probably getting busy.
2:34:21
No, I I was I was running the conference.
2:34:23
Uh yes, and individual things are going to happen.
2:34:28
Um I think GitHub's reputation is is going to last through that.
2:34:31
I think totally it has to be really egregious for for that kind of stuff to slip through. Yeah.
2:34:36
slip through. Yeah. I I mean I think like more broadly I think that there's a standard stack of what we are calling what Andre Karpathy is calling the LM OS right that I think you guys are kind of going and learning about but I think it's it's pretty well established by anyone who's who works in software agents coding agents which is you want a
2:34:55
sandbox you want a browsing environment you want branching you want u fine-tuning on your codebase like there there's just a standard stack of things that cloud codeex jewels cognition um um factory and anyone else in this field, they're all converging on the same thing and it's just who does it best and who reaches who who solves it for their customers best. Um I think that's that's
2:35:15
Um I think that's that's going to be the the name of the game for now. Makes a ton of sense.
2:35:19
Uh do we have to roll over to the next person?
2:35:20
Uh no because I I heard I mean I guess my question on data gravity is uh it it seems like uh if we compare the evolution of video generation models Google has a really amazing cornered resource in YouTube because that data is so big. It's so many tokens.
2:35:44
It hasn't been it's growing and it also hasn't been like exfiltrated to the public web.
2:35:48
Um, but it feels like GitHub, although it is a massive data set, it's not nearly as big.
2:35:54
I think it was I I think somebody clocked it at like a hundred million tokens or something.
2:35:59
It's it's just not it's just No. How big is it? Way bigger.
2:36:01
Well, public repos because I I I don't think that they can train on run private repos, right? Or can they? Uh, who? GitHub GitHub. Yeah. GitHub. GitHub.
2:36:10
I don't think they would let themselves. Yeah, they wouldn't. Yeah. Um, I don't know.
2:36:15
100 million just sounds way too low.
2:36:17
And that's my that's my gut reaction. Sure, sure. Keep going. Keep going.
2:36:19
Uh so so so the question is like how durable is the data moat at encoding versus uh in video generation purely I'm just thinking about the the raw data set size of YouTube has to be orders of magnitude larger than GitHub and so I would imagine that it's it's Microsoft probably has a less durable advantage versus Google's like kind of march down the video generation pathway if truly uh restricted by data.
2:36:51
Not all data is comparable like that.
2:36:53
So you're comparing verifiable data to unverifiable data.
2:36:56
So code can compile and if it runs it if it runs you can do more code like that.
2:37:01
Um and that creates the RL loop that lets you generate synthetic data for for more code.
2:37:05
Uh with videos you have what you have and you can train on that and uh so you're technically not even allowed supposed to train on that but who knows what Google's doing behind the scenes.
2:37:16
Well, you think Google can train on public YouTube videos, right? I don't know.
2:37:19
I I I imagine that like I I've posted videos on YouTube.
2:37:23
I imagine that I've opted in at some point.
2:37:25
I think it there was like a big fuss with MKBHD and all the other guys Yeah.
2:37:30
making up a fuss about this. You remember? Sure.
2:37:32
So, I'm just I don't know.
2:37:34
I'm not a lawyer, but like I'm sure is a completely untested clause that just has to go to court. Yeah.
2:37:38
I just want to say if uh to Google if you remove John's rate limits on VO3 you can train on our back catalog full permission please.
2:37:47
Yeah, I I think I think there is that diversity of of opinion, right?
2:37:52
Like do you want to just like give yourself to the machine or do you want to keep your keep your data to yourself and it's it's like there's no in between like you're either one or the other.
2:38:00
It's it's a very strange dichotomy like it's not a spe like many things in life are a spectrum.
2:38:05
This one is not like you you run into someone they're either Maxi privacy maxi or they're like an AI maxi. Yeah. Yeah. Yeah. No, that makes sense.
2:38:12
What what else is on the cutting edge of of of debates that are kind of raging at the conference uh this year versus prior years?
2:38:20
We've we've lived through like the P Doom debate.
2:38:22
We've we've we've we lived through like the Leopold and Brener era.
2:38:27
We've we've shifted to the geopolitics debates, but what's kind of on the frontier of like hottest topics to discuss?
2:38:34
Uh first of all, Leopold was right and uh I don't know if people know that he was on Ilia's team when the the the sort of board drama happened and like he was like directly connected to it.
2:38:46
Anyway, he we haven't lived through it.
2:38:49
We are living through it. It is happening.
2:38:50
It is directly leading to the geopolitics because he was foreseeing that and he goes exactly right. Yeah.
2:38:56
The way I would the way I would uh I would kind of characterize at least my takeaway was that his piece uh was a uh was a little bit of a pivot from the paper clipping doom and a shift to geopolitical competition in AI.
2:39:09
Is that is that a mischaracterization? No. Nailed it. Okay. But like he was right. Yeah. No. No. I I I completely agree. I completely agree.
2:39:17
I'm just wondering if if if it's like if the consensus is that he was right, then the book is closed.
2:39:23
We're not debating that anymore. What are we debating?
2:39:25
What's more are the debates? Yes. Yeah. How to do great AIPMing. How to run a tiny team.
2:39:29
We have a robotics track for the first time.
2:39:34
That is the Tesla Optimus is is speaking. Physical intelligence. Whimo.
2:39:38
Whimo just overtook lift. I saw that that already.
2:39:41
Um voice is the hottest thing um in in terms of multi modalities.
2:39:45
Um like everyone's sort of building with voice because I think it's like finally good enough. Yep.
2:39:48
And um I think maybe the last thing I will highlight to you is we are also emphasizing security for the first time.
2:39:56
Um security is like kind of a boring topic.
2:39:58
It's nobody really wants to talk about like how to secure your system but like they actually do now because they they have real money running through their their product.
2:40:04
So there's all that and then that is like roughly important equal in size to the excitement about MCP.
2:40:11
And so we have an entire MCP track with the anthropic team here. Very cool.
2:40:16
because it's uh because they're nice enough to to to come by and that fills up the the whole ballroom that we have.
2:40:21
So, are we going to get payments in MCP?
2:40:23
Is there a subtrack for putting stable coins in there or something like that?
2:40:27
Uh there there's nothing in the official spec, but we have a number of people and we have a speaker talking about the sort of MCP economy that's being enabled.
2:40:33
Um I do think that remote MCPS and authorization give you the foundation for basically just remote agents that you can buy and hire.
2:40:42
Um, and that that's that's that's effectively what it is.
2:40:47
It's just that we're still in that stage where there's still these things are still trivial enough that you can actually just write your own.
2:40:52
So, you have to overcome that build versus buy for this to actually kick off. Okay.
2:40:55
Um, and uh I would just say like I'm not at all a crypto person, but I would say that the crypto people have been ahead of us here and they haven't found that much yet, you know. No, makes sense.
2:41:08
They're but they're actively hunting and if anyone finds it first, it'll be them.
2:41:11
It's it's circles presenting here.
2:41:13
Solana presented at my previous conference and uh they're all they're all on it.
2:41:16
Stripe is also on it with with their stable coin thing.
2:41:18
Um something's going to happen here. Yeah.
2:41:22
Uh in in robotics, are we getting to a point where we're starting to see an ecosystem of companies pop up like we've seen?
2:41:30
No, it's just it's what what do you Yeah.
2:41:32
So they all vertically integrated. Sure.
2:41:34
They all building every their stack.
2:41:36
I don't see like any horizontalness if that's what you're going for.
2:41:40
That's exactly what I was asking. Yeah. Yeah. It's so weird.
2:41:42
I think um you know there's just so much custom needs that you you have to sort of reinvent the universe every time.
2:41:47
The one uh thing that that does reuse bots is a cloud chef uh which which is um you know a very young company but I I ran across them when I tried their food.
2:41:57
So they're they're a kitchen robot.
2:41:58
They do demonstration learning from a single shot uh from let's say like a Michelin chef and you just give them you give it the ingredients and and you will uh it will just flawlessly just kind of repeat that that cooking for you and you can hire it for $12 an hour.
2:42:12
So it's meant to directly replace uh human labor uh which is very expensive and very labor intensive and unreliable.
2:42:19
Let's let's put it that way. Sure.
2:42:20
Um and uh they they use they can they can live on top of any other robot arm or Tesla Optimus substrate um which okay um because it's more about the the robotics sort of uh framework than it is about the individual uh hardware.
2:42:34
And I think like that is the first time I've seen that happen.
2:42:38
I'm like relatively new to this, but like this is the first time I'm like I'm like, "Oh, okay."
2:42:41
Like he's actually pretty confident he transfers across any any system as long as they have the minimum required set of device drivers basically.
2:42:50
Um, and I'm like, "Yeah, that's cool." Like, yeah. Yeah.
2:42:52
So, he's he's speaking tomorrow.
2:42:53
What about other pieces of the robotics data stack?
2:42:55
Are we thinking about like data brokerage or or kind of like a scale AI or maybe scale AI actually working to generate more robotics data?
2:43:05
Anything on like the sim tore gap?
2:43:06
I saw some semi analysis summary of a paper that was pretty cool.
2:43:12
So I would say that as an as a sort of industry practical conference those are just in the domain of research right now. It's in research. Yeah. Yeah.
2:43:18
It's it like there's there's a lot of papers out there.
2:43:21
Um Jim Fan had a fantastic talk at Sequoia Ascent that I I recommend everyone watch if you haven't seen it about the physical touring test. Yeah.
2:43:28
How you need to uh just do a lot more simulation.
2:43:31
And there's a there's a lot of work on being done on this.
2:43:33
Um, I think Google Genie is the other one that I would re recommend people do.
2:43:36
There's like a really interesting tiein between like the generative video world and the robotics world that you wouldn't necessarily expect until you like spend some time in that.
2:43:45
Um, but we just haven't focused on it because people cannot get jobs in it.
2:43:49
Like I I want people to get jobs at my conference.
2:43:52
That's like almost the whole reason of bringing the companies, I bring the engineers, they meet, they they fall in love with each each other.
2:43:58
That's like to me that's like the most fulfilling thing.
2:44:00
Can you give a a highle update on the hiring market today? What what fact?
2:44:07
What's fact, what's fiction?
2:44:07
Uh, you know, it's one of those things people love to, you know, talk about how bad things are and yet every company that I know is like can't find people and maybe there's a talent kind of gap there.
2:44:20
But, but what's your read?
2:44:22
Yeah, I mean uh I I think both are right, which is weird.
2:44:27
Uh this is actually the topic of my conversation with uh Greg Brockman later today because I I half the people I I meet at my own conference are worried for their own jobs, right? They do ask about this.
2:44:38
It's not like they're not they're blind to it, but they're just like I'm here to see what else I I need to do or how how how do how does my skill set need to change for the future?
2:44:47
And nobody knows, you know, everyone says the trit thing, which is like, oh, I'm going from I write the code to I manage things that write the code.
2:44:54
So I become an engineering manager, but like what does that mean in practice?
2:44:57
How much knowledge do you need to supervise an agent?
2:45:01
You actually kind of need a lot uh when it goes when it goes off track.
2:45:04
So like it that's super unknown.
2:45:06
Um I do think that the juniors that the like fresh out of college students, they're a little bit cooked unless they're unless they're unless they're good.
2:45:18
So it really amplifies skill issue.
2:45:19
It really really my skill issue. Yeah.
2:45:22
Yeah, it's a hard thing to to to you know try to uh tell somebody you just need to be better five times smarter than you are and then the job and then you'll have like a bunch of offers.
2:45:36
But I also think I also think that the the solution here and and we we just hired uh an intern who already shipped a new product. He he started uh Monday.
2:45:47
He just shipped uh something relatively simple but it's cool to guest directory.
2:45:52
He shipped he just sent us a V1 of that product that he built last Thursday.
2:45:56
We saw it and we were like, "This is super cool.
2:45:58
Why don't you start on Monday?"
2:46:00
And it's that one kind of interaction that can get your foot in the door that allows you to develop skills and build relationships that, you know, hopefully we, you know, we work with him for a long time, but even, you know, I'm sure people are already seeing him at TBPN and and we'll probably try to poach him, uh, before the end of the summer.
2:46:20
So all it is it is like a changing shape of the software engineering employment because like I feel like like the vibe coders can come into organizations that might not have a full engineering department and then have an impact because of the way the tools work but uh it's certainly it's certainly a lot of change very very quickly.
2:46:38
I would I I would say that's not the consensus yet that hire a a like chief v code officer. Yeah.
2:46:43
Because like humans are still very much needed to to patch the gaps that the models are not good at and it's really painful when they do go wrong and they do go wrong.
2:46:54
You know, I think like all the hypes that you see on Twitter uh fails to admit like what happens like one month, two months, three months after.
2:47:00
And that's honestly what you get paid to do as a software engineer.
2:47:03
Make maintainable software, not Yeah.
2:47:04
green field all the time.
2:47:06
I've even seen people repurpose old old projects and say, "I I vibe coded this in 24 hours and" and everyone's like, "Wait, no, this was published in 2019."
2:47:14
Like, "You spent months on this."
2:47:16
Uh, and you're just hyping this up. That's misinformation.
2:47:22
Um, I I I can't I don't pay attention to what people say there, but I do care what people report for themselves that they they've done at serious companies.
2:47:32
So one thing I would highlight there is Booking.
2:47:34
com did about 10 um company years worth of migrations uh in 3 months with uh with their sort of automated um uh code migration that they that they did with source graph.
2:47:44
So um we did a talk with them at my conference in New York uh in February which if you want to look that up.
2:47:50
Um so like those that company has been around for like 14 15 years.
2:47:55
That's that's a serious code base, right? Totally.
2:47:57
They're just reporting their success.
2:47:59
They got nothing to to gain from selling you on it.
2:48:02
They're just they're just happy about it and they want to teach others.
2:48:05
So you want to look for those.
2:48:05
They're not trying to sell you anything.
2:48:06
They're just looking at reporting like their progress at a real company.
2:48:09
Um and it's that's really what I try to optimize for for the for the conference.
2:48:13
It's it's hard because not everyone's incentivized to do it.
2:48:16
You just have to create an environment where they get something out of it by meeting their peers at at a thing like this. Well, give us the plug.
2:48:24
Where can people watch your talk with Greg Brockman later today?
2:48:26
Uh yeah, it's on YouTube. YouTube. engineer doot.
2:48:32
And yeah, I'm What would you ask him?
2:48:35
I mean, you're having him on eventually.
2:48:37
So, you know, what's like the the way that you would open the conversation?
2:48:43
Oh, that's a good question.
2:48:46
Um, for Greg, there's a lot, but um, I mean, the question that I feel like is at the top of a top of mind in like Ben Thompson world this week is, uh, the shift from training to inference and how and how workloads are are are shifting.
2:49:07
This was in the context of an NVIDIA earnings, but Greg obviously has intimate knowledge and so um in uh OpenAI is obviously going to do larger and larger training runs, but they're also doing tons and tons of inference. How is that shifting?
2:49:22
And then what I want to know, and I don't know how much he can speak to this, is is is like where how will if we get to a world where we're in 80% inference, 99% inference, something really extreme, does that change the type of data centers we're building, does that change the chips that we're demanding? Are we moving to A6? Are we moving to FPGAAS?
2:49:44
Something like that to like speed up the actual inference workloads?
2:49:47
At what point do we actually bake these models down if they're deemed like good enough and then we're like orchestrating them?
2:49:55
If we hit a plateau, that might make sense.
2:49:56
Um, at the same time, if there's a lot of promise on algorithmic progress and we're expecting like, oh yeah, we're going to leave the transformer behind eventually.
2:50:03
Uh, well then, yeah, you don't want to bake all that down.
2:50:05
So, I think that that's like it's a little bit more of like a semi question, but it it is an interesting question for him as well as he's like seeing the workloads. It is very relevant.
2:50:13
Um, he he does have a role to play in Stargate that I'm not super clear about.
2:50:17
I'm going to ask him about and uh yeah, I mean I I I I I would say that nobody's betting on the end transformers except for very small very small number of people that are just experimental.
2:50:27
Um and uh it's it's really about just scaling the RL.
2:50:30
Um yeah, I mean there is this interesting uh thing that's happening.
2:50:35
I believe images in chat GPT seems to be using a number of different algorithms combined.
2:50:42
Uh so there's a little bit of diffusion in there potentially.
2:50:44
There's some uh there's some transformer based stuff.
2:50:45
I don't know if you have a if you're pushing back on that, but that's I'm not there there's a there's a strong hypothesis based on the hints that they have dropped from the people that worked on it. Yeah.
2:50:54
That there that there's multiple algorithms at work, right?
2:50:56
It's it's a it's a diffusion head with the transformer backbone. Yep.
2:51:00
The rest of it is just regular 40 and that's so good at chat and has has decent world knowledge. Yeah.
2:51:06
There's not like a ton of complexity there beyond that that we know of. Yeah.
2:51:11
And so what I'm interested in is like we saw the demo from uh from Google on diffusion, textbased diffusion, 900 tokens a second.
2:51:20
Uh I've heard good reviews about that.
2:51:22
I've heard kind of mixed reviews on that.
2:51:24
on that. Maybe it's not a path, but uh what does what does the future of an LLM look like if we're if we're applying the same path that we've seen in images where we're seeing an ensemble of models come together to create a better like
2:51:39
even more multimodal multi multi- uh like multiarchchitectures uh because it feels like we're moving it we're moving further and further away from the single big transformer being the answer to everything and and so what does that mean? Are we going closer or farther
2:51:54
Are we going closer or farther away from the single big transformer that is like the the god in the weights?
2:52:00
Yeah, this is where I can offer a little bit of coaching for uh for your audience and maybe you you guys have talked to a lot of people. I don't know.
2:52:06
This is where people usually say the term mixture of experts and they're wrong to use that.
2:52:11
Just and sometimes they say mixture of ages, mixture of experts.
2:52:15
They're they're they're incorrect.
2:52:15
Um anyway, uh we have talked with a lot of uh the the frontier lab researchers. Yeah.
2:52:23
No, none of them believe that that is the way forward.
2:52:25
Um that is that is an optimization for the current thing which is which which one mixture of or mixture of architectures.
2:52:32
That's mix of architectures.
2:52:35
Um let uh okay um yeah so experiments for example with Jamba with from 1821 in Israel where they mix for example like a mamba uh layer with with transformers.
2:52:46
Um that seems to be promising but even Shazer a correct character was just use scaling out transformers.
2:52:53
Um I I I asked a variant of the question you just stated to Nome Brown in an upcoming podcast that we did. We recorded it. We haven't released it. He's the same way.
2:53:01
He's like, "Anything you're trying to do fancy around mixing up weird architectures is just not going to scale.
2:53:08
Just scale out the the the basic thing."
2:53:09
Um, and he's pretty strongly convicted in that.
2:53:12
I I know people I can't name a Gemini who is also pretty strongly convicted in that.
2:53:17
Um, I have reason to suspect otherwise.
2:53:19
Obviously, like you can't shut this shut down anything because like it might be true.
2:53:23
Um, I'm just telling you what the people working on this would say because I've asked. Okay.
2:53:28
Well, I mean, if that's AI, the last thing I wanted to kind of cover, I I would want to get a sense of, you know, prediction markets are pricing in that uh Google continues to dominate on benchmarks.
2:53:38
How much is OpenAI even going to how much do they even care about being at the top of benchmarks?
2:53:47
Is it is it purely an ego thing or is it just about uh you know, usability and and value for users and code quality and things like that?
2:53:55
It's usability and value.
2:53:57
Um uh you I think you need to notice that the benchmarks that Google talks about at Google IO are no longer the benchmarks that OpenAI measures itself on and um you know I want I want to stay friends but like there are some that are more sus than others. Sure.
2:54:15
And um I think like at the end of the day the the the customer will win.
2:54:19
Um, the only one that that has not that's clearly been caught out kind of lying, kind of like being whatever like underperforming, I got to say it, it's Apple.
2:54:30
Um, when they launched Apple intelligence, we were all very excited about the Apple intelligence paper, which is a which is a beautiful document that had no EVLs apart from their own internal eval.
2:54:39
So, they were not accountable to the standard set of evals that everyone else has, right?
2:54:42
So, um, step one is you you should try to hold yourself to some public benchmark.
2:54:46
I flawed just do it you know but and then uh so so so Apple did not do that um and and then like now there's a question of like are you holding yourself to a benchmark that can be gamed um and there's been accusations of L Marina being gamed and unfortunately that's what Google has put all it ships on they may do that for Gemini 3 but for 2.
2:55:05
5 that's what they did u and probably because they spent the last year doing it it's all revenue from here on out like that is the revenue is the benchmark GitHub copilot 500 million size gone for GitHub revenue is also tricky.
2:55:21
Yeah, you guys know how like Gemini I don't know if you know Gemini gives a billion tokens a day per human. What? Your personal API token. Okay, I'm kidding you.
2:55:33
You can you can you can you can set up a camera right now, run Gemini 2.
2:55:37
5 flash on it, uh run like a frame a second and ask you to do whatever you want for free. Wow. Um it's absurd.
2:55:45
So for Google, thank you.
2:55:48
Intelligence too cheap to meter. We love also. Okay.
2:55:50
Like that chart that you saw at Google IO where there are tokens going like this.
2:55:54
Now you know why it's free. Yeah. Yeah. Yeah. Yeah. Yeah. Makes sense.
2:55:58
I mean it's they also have a very very good model. Don't get me wrong.
2:56:01
But I'm just like revenue is a short-term play.
2:56:03
Like you're trying to maximize revenue now.
2:56:05
You might cut off cut off yourself from getting the largest data set in the world which is the uh the sort of early adopter humans that are going to give you their data because they're like train on me daddy.
2:56:15
I I'm you're giving to me for free like whatever train on me.
2:56:18
Um and and that's a that's a deal that lots of people I think a theme that we we talked about is lots of people are going to make. Yeah.
2:56:25
Is free why they they can eat your data. Codeex is free. Why? Yep.
2:56:33
Um so I I talked about this in in one of my recent posts like the the the marginal cost of software has is is has gone down to zero and like uh can it go negative?
2:56:43
Will I start paying you for uh for me for you to use me? Absolutely.
2:56:47
Because then you become my labeler. Yeah.
2:56:49
You're you're a labeler I can never pay for.
2:56:51
You know, usually have to like hire someone in like the Philippines or something. We've maxed them out.
2:56:55
So now I got I got to I got to pay someone in Silicon Valley to label my my software.
2:56:59
So I might as well just give you my coding agent for free or like even honestly just pay you for good feedback on my coding agents.
2:57:05
And why should that why should it stop at like $20 a month when like there are people in New York who I've heard by the way get they're like for they hire former bankers they hire former hedge fund guys pay them 500k a year to label. It's wild.
2:57:21
Well, thank you so much for hopping on.
2:57:23
Uh good luck with the rest of the conference.
2:57:25
Let's make this a regular thing. Fantastic conversation.
2:57:29
Thank you for taking the time to your event and we'll be there next year whether you like it or not. Fantastic. You're invited for sure. We'll talk to you soon.
2:57:36
Thanks so much for Thanks, Sean. Bye. Cheers. Uh, fantastic.
2:57:40
Well, uh, we got to tell you about Wander. Find your happy place. Find your happy place.
2:57:46
Book a wander with inspiring views, hotel, great amenities, dreamy beds, top tier cleaning, and 24/7 concier service.
2:57:50
Is a vacation home, but better, folks.
2:57:53
I want the TBPN army to clear out every single wand of summer.
2:58:02
Mark clearing order inbound.
2:58:06
Um, this is an interesting one.
2:58:06
Oscar has joined the Fortune 500 for the first time.
2:58:10
I know what you're thinking and I agree. What? Only now?
2:58:11
And Josh Kosner says, "I'm deeply proud of the tenacity and pers persistence of the Oscar team.
2:58:18
We've been humbled many times.
2:58:18
From down 95% from our IPO to now entering the Fortune 500.
2:58:23
I am excited for all that is ahead of us." Wow.
2:58:26
Congratulations to Josh Kushner.
2:58:28
That is a phenomenal run. Amazing.
2:58:31
And somebody else was quote tweeting this saying that um there's only two people who have incubated a Fortune 500 company who are actively investing and it's Palunteer with Peter Teal and founders fund and Josh Kushner with Thrive and and Oscar.
2:58:44
So uh what a phenomenal run.
2:58:47
It is rare but the incubations are starting to work and the incubations will continue until morale improves will continue until the global economy is worth is growing at 10% a year.
2:58:59
Um anyway, uh every person who works in tech needs at least non five non- tech friends to interact with closely on a weekly basis so they can understand how the general public actually thinks says Kit Volulta.
2:59:09
You posted this from a different account. It was a copy pasta.
2:59:14
Somebody copied the whole thing.
2:59:14
So I found the original shared you shared the fake one. The ripoff. I shared a ripoff.
2:59:20
Yeah, you shared a ripoff and it got community noted and I found the original one right here.
2:59:23
Wait, so somebody just stole it? Yeah. Yeah.
2:59:25
They didn't even screenshot it and put it in banger archive.
2:59:29
They actually but 20k on this post.
2:59:32
The repost has 15k likes and so it got but it got community noted. This is a copy paste.
2:59:36
I you feel like that could be uh that could be like automated in axe where it's like this is the exact same text as has been posted elsewhere.
2:59:44
Uh you should just know that you should maybe go to the source on this if somebody's posting the exact same thing.
2:59:48
Except you know Josh Diamond would get that every day because he's constantly boasting good morning we are going to win.
2:59:53
And so he would have the note every single day.
2:59:54
Um, JT uh, Jira tickets says, "Web design, you mean dig digital physioamy." Very funny.
3:00:00
Oh, you put another one in here from JT.
3:00:02
Guy who is burnt out from his six figures, five hours a week big tech job.
3:00:06
You got to just hit the hit the hit the roof and just uh get some sun, you know.
3:00:11
Anyway, Pavle has a good one.
3:00:13
I don't think Whimo can work in New York City because during rush hour, you need to break the law to get anywhere.
3:00:19
And I'm not sure how you could get away with embedding that in software.
3:00:22
And IB says, "Travis would have gotten it done. Never too late."
3:00:27
Somebody says, "Full self-driving gladly speeds for me." Yeah.
3:00:30
Uh, I thought this was good from Will too thick scoops. Uh, okay. Sounds good. Thank you.
3:00:39
And it's Jersey Mike's order pickup at 12:48.
3:00:42
Dutch government collapses over migration dispute. Why is this? This is a Gmail.
3:00:46
Oh, it's just rolling up multiple emails or something.
3:00:50
This is Apple AI summary. Oh, yeah. It's an Apple AIM.
3:00:52
intelligence is the summaries are so good. It's genius. 75,000 likes.
3:00:57
That's a lot of attention on that.
3:00:59
You got to upgrade, folks.
3:01:01
You're missing out on entertainment.
3:01:01
Pay Apple to help me produce Apple intelligent viral bangers. Yeah.
3:01:05
I want to wrap up, but we got to say congratulations to Jesse Michaels. He got on Joe Rogan.
3:01:11
He said it was the honor of a lifetime to sit down with a goat himself, Joe Rogan.
3:01:15
Many people have called him a young Rogan. I've called him that. He is fantastic.
3:01:19
He's been on a fantastic run with American Alchemy. his podcast.
3:01:22
He's climbing the charts.
3:01:25
If you're into aliens, AI apocalypses, Bob Lazar, nuclear weapons, secret anti-gravity research, go check out uh the Jesse Michaels uh Joe Rogan Experience episode. It's episode 2,31.
3:01:40
Rogan's putting up big numbers.
3:01:42
He's giving us a run, but we're catching up. We're catching up. That's wild.
3:01:44
Um in a couple years, lastly, just wanted to cover this quickly. AMD has acquired Bream. Okay. Reported today. Interesting.
3:01:54
Uh to help reduce Nvidia's market dominance when it comes to AI hardware.
3:01:58
Talking to George Hotz and and we're going to have him on the show, but uh he's been he's been back in AMD.
3:02:02
We got Dylan Patel coming on the show Friday.
3:02:06
We'll ask him more about AMD and their their plan to unseat CUDA as the dominant AI platform. Yep. And we have other news.
3:02:14
We are officially we have doubled the size of the show once again. Jordy, hit that gong. Hit the reel.
3:02:19
Hit the real gong for me. Grab this. We're going to gong cam.
3:02:24
Uh, we got 64,000 followers on X.
3:02:27
Thank you for everyone who's been around. Oh, there we go. That's great. Thank you for watching.
3:02:35
Thank you for supporting us.
3:02:35
It's been a fantastic journey. It is a pleasure.
3:02:38
We got the whole crew in the studio.
3:02:39
We got a whole bunch more cameras here. We're growing every day.
3:02:43
Damn, we got the studio cam here. The studio cam's going. Thanks for cooking. Thank you for watching.
3:02:48
Thanks you for all the support review on Apple Podcast and Spotify on Apple Podcast.
3:02:52
Yeah, I feel like we're doing low there.
3:02:54
Yeah, it doesn't really make sense.
3:02:56
But if you're on Apple and you're listening and you haven't left us a review, we'd love a review. So, thank you so much.
3:03:01
And we'll see you tomorrow. We have an awesome day.
3:03:03
We have a we have a bit of an AI day coming together with folks from OpenAI and Anthropic coming on.
3:03:08
Uh it should be a great show. So, stay tuned.
3:03:11
We'll see you tomorrow and have a great rest of your Wednesday. Goodbye.