0:06
[Music] [Music] [Music] [Music] [Music] Team Deathmatch.
Search this video
0:59
[Music] We are surrounded by journalists. Hold your position.
1:10
[Music] [Music] We came to this world to reach the stars. We came to this world. to shape a future.
1:56
We came to this world [Music] to reach the stars to shape our future.
2:08
[Music] We came to feel the new If you let me see. [Music] Feel the music.
2:51
[Music] [Music] [Music] 5.
3:18
[Music] You're watching TBPN.
3:20
Today is Thursday, June 19th, 2025.
3:23
We are live from the TVPN Ultra Dome, the temple of technology, the fortress of finance, the capital of capital.
3:31
Uh we have a great show for you today, folks.
3:34
There's some breaking news that's dropping right now.
3:39
I think we got to go to the printer cam. Really?
3:41
Because we have an update from friend of the show. Let's see if this works. Do I need this? Uh no.
3:45
It's we need a moment of silence, gong.
3:48
um but because it's about the Elon Musk news out of SpaceX.
3:53
But we got an update from friend of the show Ashley Vance coming in here hot.
3:58
He says, "I happened to be at Neurolink last night when Starship went boom and so was Elon Musk until well past midnight Pacific time.
4:07
He was in three-hour plus He was in a three-hour plus long meeting when the explosion happened. Meeting ended.
4:15
I assume that's when he learned about it and then he went back to work. Wow. An absolute dog.
4:21
I mean video was absolutely insane.
4:23
Uh we will cover it in a little bit.
4:26
You had shared a transcript from it.
4:29
Oh yes with me earlier this morning.
4:31
I had seen the the video of it going boom.
4:33
Y uh you shared this transcript where one of the engineers is saying, "Hey, really quick, Sawyer, we just observed a couple of uh vents coming from the common dome in between uh locks tank and the methane tank.
4:48
And from this angle, it almost looks like the methane tank is gone."
4:53
Then he actually says, "Question mark. Is that normal, Jack? Is that normal?" Yeah, it's crazy. I'm seeing some venting.
4:58
Uh is that is that unusual? Is that normal?
5:02
And the guy goes, "Yeah, it's probably normal."
5:04
And they literally say, "I have just You have the video. I I just sent the video. Let's play this."
5:11
I have no idea if this is actually related.
5:12
We'll have to get someone on the show to dig into like exactly what happened.
5:15
I'm sure there'll be a post uh uh like a postmortem on the explosion.
5:19
They did say famous last words and then shortly afterwards the entire It is a crazy crazy video.
5:23
And really quick, so we Let's see if we can pull this up.
5:27
Um in the meantime, let's tell you about ramp ramp. com. Time is money. Save both.
5:33
Easy to use corporate cards, bill payments, accounting, and a whole lot more all in one place. Go to ramp.
5:36
com to get started, of course.
5:38
Uh the other two major news stories we want to cover today, 10 billion the price to buy the Los Angeles Lakers, 15 billion the price to buy Meta and AI leadership team.
5:49
Fantastic post by Alex Conrad.
5:49
Uh let's I've just observed the third show of vents coming from the common dome in between uh the locks tank and and the methane tank.
5:59
Uh, and from this angle it almost looks like the methane tank is gone. Is that normal, Jack? Is that normal?
6:15
And then I actually edited out a little.
6:16
I'm seeing some some venting coming from in between the methane tank and a lock thing tank. Is that usual? Is that normal?
6:21
Yeah, it's probably normal. Key word probably. Famous last words. Yeah, weasel words.
6:30
They miss last week's work.
6:30
These guys are just broing out on a live stream watching watching like a static firing test. This is not a launch.
6:37
They're not going to they're not trying to launch the rocket.
6:40
They're just they're just putting on the t on the test stand firing it up to make sure that everything works and it just completely exploded.
6:46
Um we'll go deeper into that and some of the reaction and the news in a little bit.
6:51
But in the meantime, uh let's talk about the other piece of breaking news that came out of the printer just after we got off the stream yesterday. Oh, right.
6:58
Um this is news from uh the information.
7:02
Meta is in talks to hire former GitHub CEO Nat Friedman and Daniel Gross to join AI efforts and partially buy out their venture fund. Yes.
7:11
So there's a ton of details here.
7:13
So let's read through the information article and then we'll go to some of the reactions.
7:16
some of the reactions. Um so meta platforms is in advanced talks not just talks advanced talks has that been defined quantified what does that mean like like are we past coffee meeting is this a 30 minute is this a two-hour conversation how long are the talks until they become advanced terms you know exact numbers are being thrown
7:37
around very possible yeah uh so they they're thinking about bringing in Nat Friedman Daniel Gross to help lead AI efforts uh part of those talks Meta is in discussion about partially buying out Freriedman and Gross's venture capital firm NFDG, which holds stakes in top AI startups and is worth billions of dollars on paper. Uh, if the talks are
7:53
Uh, if the talks are successful, Gross would leave Safe Super Intelligence, which he co-founded with former OpenAI chief scientist Ilio Sitskiver last year.
8:02
At Meta, Gross is expected to work on mostly on AI products, while Freiedman while Freriedman's remitt is expected to be broader.
8:10
Both Gross and Freriedman are expected to work closely with Meta CEO Mark Zuckerberg and Scale AI CEO Alexander Wang or Wong um whose hiring by Meta was finalized last week in a $14. 3 billion deal. Big numbers.
8:21
Big numbers being thrown around.
8:25
I think I think this is gong worthy.
8:27
Even though we're just in advance talks, we got to hit the gong.
8:30
We try not to hit the gong for advanced talks, but scoops are scoops.
8:34
It's it's such a big number. Fantastic. strong hit.
8:39
Um anyways, um yeah, I guess some, you know, immediately a couple things I was thinking about when I saw the headline.
8:50
One, uh I I don't know if this is uh common knowledge, but my understanding was that some of the money from uh NFDG was Zak, right?
9:00
So, they were already investing on behalf of Zuck to some degree, so this shouldn't be a huge surprise.
9:07
Um and then I think the bigger thing is what does this say about SSI, right?
9:13
If if um DG is willing to leave SS SSI despite it being, you know, such a young company and already valued, I imagine DG stake is in the billions of dollars there. Yep.
9:24
So to leave that um and and go to Meta says something.
9:27
I don't know exactly what it says.
9:31
I I think it could potentially say a few things.
9:35
One is that maybe artificial intelligence is more of a sustaining innovation than a disruptive innovation.
9:41
And so that just by training a fantastic model, you're not immediately going to be able to overcome the network effect at Meta.
9:48
And so Meta is maybe potentially a better place to go, you know, really reap the rewards of artificial intelligence.
9:54
That's kind of a signal because, you know, no one at Google was really thinking about joining Yahoo, right?
10:01
There wasn't a lot of flow that direction.
10:03
Yeah, it was like we're on to something.
10:06
We are going to disrupt, you know, same thing with with Amazon.
10:10
I'm sure Bezos wasn't losing people to Barnes & Noble, right?
10:15
This is this is Barnes & Noble threw out a couple max contracts and got a couple mercenaries.
10:20
But yeah, I mean, if you're, you know, this was truly disruptive, you would think that you would say, well, I don't I definitely don't want to be with the incumbent.
10:28
don't want to be in the legacy player because they there's nothing that they can do to capitalize on the new wave of technology.
10:34
And so there's been this question about AI clearly an incredible technology.
10:38
Clearly, you know, like the greatest invention since it's up there with electricity and fire like it's it's really really cool. The computers talk now. It's incredible.
10:46
At the same time, what what is the market dynamic that drives how this technology will acrue value in various places? Who will the winners be?
10:55
A whole bunch of startups?
10:57
Will there be a monopoly player around that comes from a startup?
10:59
Will there be a monopoly player and then sustaining innovation in every other mag seven?
11:02
And this is the question that everyone has been talking about for years, for a couple years now.
11:08
This is what Ben Thompson writes about what what we talk about all the time.
11:12
And I think people are gradually waking up to the idea that like it's possible that a lot of the value creation at the foundation model layer will happen at OpenAI because of their consumer products, right?
11:24
And this aligns with with Sam's piece from last week, the gentle singularity.
11:28
It's basically saying like we created intelligence, and it's less weird than we thought, right?
11:34
Which is a step back from from um how things were were talked about and and is a stark difference uh than than maybe AI 2027, which is, you know, super uh extremely AGI pill and just saying, you know, we're going to continue accelerating. And I don't know.
11:54
I think I think it's uh I think it's right to kind of read into this and and um it's it's two things can be possible.
12:02
It's possible that Ilia will create, you know, very important lab with SSI, but it's also possible that that it they might never grow into a $30 billion valuation.
12:12
Is that where they are now?
12:14
They're currently priced at $30 billion.
12:16
So for for uh DG to to leave as a co-founder of that company, Sure.
12:21
I'm sure he he's getting, you know, he'll get a a 10 figure package if he goes to Meta. Yeah.
12:28
If this goes through, but uh he's also leaving, I would imagine, billions of dollars of of, you know, shares. Yeah.
12:36
This seems like very rumor mill at this point.
12:38
Like this could go a bunch of different ways.
12:40
It could just be talks and maybe they just come on as like advisers or something or they join the board.
12:44
Like Meta has a board that includes people that don't work at that company and and and they add a lot of value there. So that could happen.
12:50
It could also be that uh Meta winds up acquiring SSI.
12:54
That would be a wildly different take on this.
12:56
Or it could be that they leave.
12:58
The information seems like somewhat confident about this.
13:00
But it's only a couple sources.
13:02
SSI, there's multiple co-founders.
13:04
It's very possible that as even since the company was started, certain members of the co-founding team like Ilia, you know, want to build what what they see as super intelligence.
13:15
And it's possible other members of the team are like, "Yep, this is like, yep, more software.
13:21
We're we're going to vend this out in a bunch of different places." Yep.
13:23
And ultimately, um, you know, it it just I don't know.
13:27
I You don't typically see people get off rocket ships. I agree. Yep.
13:33
Um, co-founders unless there was an extreme, you know, rift.
13:39
There's another side to this which is there's a question about what structure even even if even if the goal still is super intelligence what is the what is the corporate structure and the and the capital formation structure that delivers super intelligence because we saw this with open AI when it was a nonprofit.
14:00
There was simply no way to marshall a $10 billion donation to a nonprofit for a large GPT 4.
14:07
5 GPT5 level training run. Yeah.
14:12
There was no way to marshall that type of capital.
14:15
Like the biggest pe the richest people in the world had already donated $und00 million and there was not really a lot of appetite for Yeah.
14:23
Next year I'm 10xing my donation. Yeah.
14:26
And so and so they had to become a for-profit. Yeah.
14:30
Then you look at at the flywheel of what it takes to continue to develop and continue to do these training runs and continue to invest in reinforcement learning and it feels like you need a data feedback loop and you also need a financial feedback loop to be able to justify more and more investments.
14:48
And so, so if we're and we're going to talk to Mike at ARC AGI and there's this interesting thing that we heard which was that the reason that the foundation models are not able to oneshot ARC AGI right now is because they're all just like doing a nice thing and not reinforcement learning on it.
15:08
But if they actually were did like some finetuning around it and they were like, "Hey, we want to knock this model off," they could.
15:16
And and what that tells me is that for any really well-defined problem like chess or Dota 2 or go or League of Legends, like you can go and say, "Hey, we're doing a specific training run for this one problem and it's going to get really good at it."
15:33
The the weird thing is is that the economy and the like the global value creation chain from humanity is um potentially extremely longtailed.
15:44
there's potentially not just like five skills like oh yes you know IMO level math and you're good and you generalize you you might need to go and dig into all these different different pieces of value and and having a feedback loop or
15:59
an economic model like what OpenAI has with their app that generates a ton of revenue or like what Meta has where they can deploy these products in all sorts of different ways and get billions of people using them very quickly. that
16:08
that actually might be a a more like it might be the only way forward.
16:13
You might not just be able to go into monk mode, come up with the perfect algorithm and then train it on some like medium-sized cluster.
16:22
You might actually need to just scale energy, scale data center capacity and scale users smoothly for decades to get there.
16:30
So I don't know that this is this is updating my like probability of super intelligence ever happening.
16:34
is is uh how many different uh labs that are losing billions of dollars a year can the capital market support and for how long? Exactly. Right. Yeah.
16:44
You have X AI, thinking machines, safe super intelligence. Yep.
16:48
Um you know, anthropic is is kind of and open AI are in their own categories and that they are generating a lot of revenue.
16:54
Um, it's also hard because it's not like biotech where if you come up with a machine learning algorithm or you come up with, you know, the transformer, you patent it and then you just make money off of it forever.
17:05
Like that's not the way these these innovations Yeah.
17:11
But like if that was the case, I would actually be maybe more bullish on SSI because I would say, well, Ilia is clearly like an incredible researcher.
17:22
if he goes into, you know, his team and comes up with the next great training paradigm. Yeah.
17:28
And then patents that and is able to license that to Google and OpenAI, that could be extremely valuable, but that's just not the way the structure of the market is is like drug development. Exactly.
17:38
Um anyway, it's a fascinating story.
17:40
There's been a ton of reaction to this.
17:42
Uh Nick says, uh this is somewhat related.
17:45
Uh, Cararpathy literally said, "Meta's llama ecosystem is becoming the Linux of AI and you're blackpilling."
17:52
And so this was kind of like a narrative violation.
17:55
A lot of people should get a little bit more into the article because it does does give some color.
17:58
So Freriedman has been involved in Meta AI's efforts for at least the past year.
18:03
In May 2024, he joined an advisory group uh to consult with Meta's leaders about the company's AI technology and products.
18:09
Earlier this year, Zuckerberg asked Freeman to lead Meta's AI uh Meta's AI efforts altogether.
18:16
The person familiar with the discussion said Friedman declined, but helped brainstorm other candidates, including Wang.
18:22
While Zuckerberg was skeptical Wang would leave scale, Freriedman convinced him a deal was possible, said a second person with knowledge of the discussions.
18:29
As the Wang hiring came together, Zuckerberg approached Freriedman again.
18:33
This time, Freriedman agreed to a deal of his own.
18:37
He is currently expected to report to Wang who is roughly 20 years his junior.
18:41
Both men will be a part of a small group of meta leaders that Zuckerberg refers to as as his management team or M team.
18:49
For Gross, the talks with Meta put him in an awkward position with SSI.
18:50
A startup form with the goal of building a leading AI uh company insulated from short-term commercial pressures.
18:57
So again, uh SSI's strategy from the beginning is saying we're not going to release anything until we create super intelligence.
19:04
I just think it might be the nature of the economy and the nature of of artificial intelligence and the structure of the market that might mean it is impossible to insulate yourself from short-term from investors such as Green Oaks, Andre and Lightseed Venture Partners has also raised money from Sequoia Capital.
20:02
They basically got everybody they got the whole crew together.
20:06
Freeman and Gross have invested in some of the buzziest AI startups including search startup Perplexity and robotics startup the bot company.
20:14
That's Kyle Bot's company.
20:14
The firm had more than two billion of assets under management as of the last year though that figure is likely higher now with the increase in value of some of their startups. Um so it's wild.
20:26
Anyways, um this feels crazy, but yeah, Nat Freiedman independently going to work at Meta is not that crazy.
20:33
I feel like the craziest part is uh you know is someone like DG going from SSI to Meta, but at the same time you know it's very possible that SSI and Meta could work out some type of relationship and and maybe that's not getting reported yet.
20:52
What's interesting is that both of these guys Daniel Gross and Natt Friedman were both at one time thought to be like future really really significant leaders in mag seven companies.
21:02
So Daniel Gross, he started an artificial intelligence company, I believe, went through YC and then or maybe he went to YC after, but he sold it to Apple.
21:12
And at Apple, everyone was kind of like, "Wow, now that he's in there leading AI at Apple, he's going to be kind of like this young incredible talent.
21:21
Maybe he'll be like the next Steve Jobs.
21:23
Maybe he'll like take over the company one day."
21:24
And people were kind of like waiting for that, but it didn't seem like Apple was really set up for this.
21:29
Accepted into YC in 2010.
21:30
He was the youngest founder ever accepted. Yeah. Yeah. Yeah.
21:32
and then he went back as a partner shortly after because he left Apple.
21:35
But there there is a different like fork in the road where Daniel Gross is like next in line to run Apple after Tim Cook if they were set up to empower someone young, which I don't think any of these big companies really are necessarily, maybe except for Meta.
21:49
Um, and then Nat Friedman has the same thing where he's he's CEO of GitHub. He goes into Microsoft.
21:55
You know, it was always a possibility that, you know, GitHub's really important.
21:58
It's this $500 million business. It's growing. It's codegen.
22:02
like he's set up in the tech industry like he could have potentially taken over at some point. Yeah. Yeah. Yeah. That's just co-pilot.
22:08
And so so there was a world where you could see them at the ranks, but we don't we don't think about it this way because most of the succession plans in in manager mode.
22:17
Big tech companies are more managers.
22:21
We don't tend to acquire founders and let them take the helm.
22:26
But Zuck, it's not like he's stepping aside by any means, but he's very much leaning into this idea of like there is something special about these founders, these people who have built companies, these people who are at the heart of the technology really really in the midst of things.
22:41
Get them on my side at any cost. And I I I love it. I think it's amazing.
22:43
I think Nat and Daniel both want to make a dent in the world, especially in the context of AI, right?
22:52
So, they're not going to want to go to Meta and just cruise and, you know, make ads 10% better, you know, that kind of thing.
22:59
Um, make it easier to generate.
23:01
You do this deal and then you just go and rest invest.
23:02
I don't think that's going to happen.
23:05
No, I can't I can't see it.
23:07
Anyway, it' be interesting, too.
23:09
I wonder, you know, would they continue to be able to invest, you know, independently or would they just, you know, or would there be kind of structure that that says like, no, you have to you actually have to just go all in on this.
23:20
If I was Zach, I would I would hope and expect for that, but who knows? Whatever. working on.
23:25
I'm sure they'll be using linear over there.
23:28
Linear is a purpose-built tool for planning and building products.
23:30
Meet the system for modern software development, streamline issues, projects, and product road maps.
23:34
And they got linear for agents, folks.
23:36
Uh Dylan Patel is uh doing a little meme on this.
23:39
Zuck, founder mode master plan.
23:41
Don't pay the PyTorch and LLM people enough.
23:43
Lose 20% of the torch people to thinky thinking machines.
23:47
Hire Alex Wong, Nat Natt Freriedman for 10 plus billion dollars to help you recruit talent.
23:52
inflection back the torch people at 10x their previous total comp.
23:57
Uh and so Dylan's obviously saying like you should have just bet on the same people earlier and kept them.
24:03
I unclear how much of it was you know really about pay but clearly that is not a gating issue anymore.
24:07
The the floodgates have opened the the this was something that that we was was identified earlier.
24:16
We we we covered a timeline post about this where it was like like what how how will Apple compete in a world where like they can't they can't justify paying anyone $10 million a year like if that's the new normal or that that that's like the value of some of these people that are going to do some of this research.
24:30
Um you're going to be kind of hamstrung and and it's not because you you're not spending $10 million on an organization.
24:38
It's because you're not tell spending $10 million on a person which is a crazy new thing. Yeah.
24:41
Sam uh Altman was was taking shots at Meta cover of the Financial Times today. I got it right here.
24:49
Basically attempting Sam came out and said Meta started making these giant offers to a lot of people on our team like $100 million signing bonuses and more than that comp per year.
25:00
Uh I'm really happy that at least so far none of our best people have decided to take them up on that.
25:07
Yeah, the the metag game in here is like wild. It's so good. 3D chess. Yeah. One of our best people.
25:14
He's just getting into getting into Zuck's head. Yeah. Yeah. Yeah.
25:15
Somebody had a good breakdown of that.
25:17
He says the strategy of a ton of upfront guaranteed comp and and that being the reason you tell someone to join really the degree to which they're focusing on that and not the work and not the mission.
25:28
I don't think that's going to set up a great culture.
25:30
set up a great culture. Alman added, "I mean, the only thing here is like tell that to the world, you know, the the world of like Wall Street and like hedge funds where like if somebody's just really good at making money, you'll just offer them like a maxed out contract to
25:45
come over to your team and it's entirely, you know, like motivated by the value that they're creating, but it's very trackable and it's a lot harder in in this case, but still, you know, it's clearly up there if you're moving the market cap like you can kind of tell." Um, Spore says, "Is Ilia SSI
25:59
Um, Spore says, "Is Ilia SSI already DOA if its co-founder is potentially about to be poached?" Good question.
26:07
Uh, Swix says, "These guys are already sent millionaires."
26:09
So, we're not talking about a hundred million signing bonus anymore.
26:12
It's the first 1 billion signing bonus in history. This is going to cost.
26:16
Zuck clearly in is in spend mode.
26:18
If you think about 100 million bonuses are high, this is a guy who lost 14 billion in 2022, 16 billion in 2023, 18 billion in 2024, and 20 billion in 2025 to invest in VR.
26:28
All he has to do is cut VR spend for 2025, and he has more money than Anthropic has raised in its entire lifetime.
26:34
Wow, I didn't put it in that terms.
26:36
Uh, never bet against Zuck long term, but I think we're in for another costly period of investment.
26:42
And we know what happened last time.
26:44
He went so hard on a thing.
26:46
We do not have the balls or imaginations to do what he is about to do. Yeah.
26:48
Can you imagine being Tim Cook running Apple three trillion dollar company making a poulry 74.
26:55
6 million in 2024 just looking after just going through the most brutal year of his time at Apple.
27:05
You know, like it's brutal.
27:07
Pulling the company back from, you know, back from the brink of of this like trade war. Yeah. Yeah. Yeah.
27:11
He's actually fantastically.
27:13
He's checking his payubs being like like he's like a scale who doesn't even work there.
27:21
He just got paid out bigger than me.
27:22
When you put it into context that that some 24year-old AI researcher who's cracked Yeah.
27:27
and like deserves a great role at a great company with great pay making more than the CEO of Apple. It's rough.
27:40
It's just it's just absolutely brutal.
27:43
So anyway, we still need to organize this protest, hit the streets for Tim Cook. We do. We do. Head over to Certino. Yeah, we do.
27:48
We should design some posters and Figma for it. Go to figma. com.
27:54
Think bigger, build faster.
27:56
Figma helps design and development teams build great products together.
27:57
Uh Nathan, while all these companies are duking it out, Figma is powering the design teams of all of them. Yes.
28:04
So, so it's kind of like one hand washes the other scenario. We'd love to see it.
28:09
Uh Nathan says, "Zuck and and Freriedman to lead AI for tens of billions of dollars was not on my bingo card yet."
28:16
I don't think and many people uh predicted anything like this happening.
28:21
I think everyone was was kind of saying like there there's probably going to be like some sort of V2 of the Llama strategy, but being so talent focused, I I I think was not on the table.
28:32
It was more like, okay, maybe they'd do an an acquisition of a foundation model lab or maybe they would just build an even bigger data center since they have abilities there.
28:40
And I think um and it's and it's been a diff a very different We don't we don't know much about sports, but there's probably I was trying to think is is like Luca Don Donsick or whatever going from from Dallas to the Lakers. That was a big surprise. That was a surprise.
28:54
SSI co-founder going to Meta. Yeah.
28:56
This is the link of of tech. Yeah.
28:59
Yeah, for people who know what reinforcement learning is. Exactly. Uh Luke Metro chimes in.
29:05
Dog, how much is Zuck paying?
29:07
Uh he's uh he's over at Anderol right now.
29:10
The meme has been, do you want to just sell ads or do you want to build something important at Anderoll?
29:16
Well, with with these pay packages, I think you're going to get some Anderoll engineers being like, I'm willing to sell ads.
29:24
I'm willing to optimize ads.
29:24
Actually, I see a lot of ads throughout my day.
29:29
Yeah, I've always been kind of fascinated by that protecting the world and ensuring like you know western le peace creating world peace is like noble but like at a certain dollar value ads are cool too. Yeah. Absolutely insane.
29:48
Well, I'm excited to see this unfold.
29:50
The way I mean some of it's interesting the way that this reporting is written.
29:55
It it it feels at times like it's already happened, but it's clearly not confirmed.
30:02
So, yeah, it could kind of go either way.
30:05
But, I mean, we heard the leaks about scale AI like like a few days and and there was some speculation about what was going on there and it became very real.
30:12
And so, you know, who knows?
30:14
Maybe it does become real, but we'll be tracking it here.
30:15
Uh uh Near Cyan says, "Ladies and gentlemen, Midjourney has done it.
30:21
Uh it's a new AI image to video model.
30:25
Uh Justine Moore from Andre and Horowitz mentioned this yesterday, but uh the posts have been going out on the timeline.
30:32
We have our intern Tyler Cosgrove in the studio today playing with Midjourney video. How's it going so far?
30:40
Can you give us a little review? Uh good.
30:41
It's been a lot of fun so far.
30:43
So you're in the MidJourney Discord right now? Yes. Fantastic.
30:46
So I've actually made I've made four videos so far. If we can pull those up. Yeah. Let's see.
30:50
I I'm excited to see these. He's in the Discord. He's in the Discord. Live from the Discord. He's in the trenches.
30:55
Uh, how is the how has the interface been?
30:57
You just upload an image.
30:59
Does it do the same thing that you get with a MidJourney image where you type a prompt and then you get four images and you get to pick one? Yes. Okay.
31:06
So, you get four video results.
31:08
Four videos and then you upres the one that you like basically. Yeah.
31:10
I think when you export it basically does that. Got it. Um Oh, okay.
31:13
Here you kicked it off with a with an image of us uh reading the paper. All right. Let's see.
31:21
So, so this was uh you know kind of bear domestication, right?
31:23
This is so so so did you include a prompt alongside the image? Yeah. Yes.
31:28
You add an image and then you prompt it. Okay. Cool. Cool. It knows us too well. Okay. That was the first one. Okay.
31:35
Let's see the second one. Love it. This is great.
31:37
Bear domestication is in our future. Okay.
31:40
This is us on our phones. And what is this?
31:43
We have a kind of an angel flying over.
31:46
Ooh, the very very bizarre.
31:48
I thought I was thinking Pegasus.
31:51
Kind of has bit of a demon vibe. Who? Who's the angel?
31:54
Uh, in the back that uh What was the prompt for this one? Let's see that one.
31:59
Angel wings or something.
32:02
Angel flies up behind two men as they look back and smile. Okay. Yeah.
32:04
The actual the actual video on us. Okay. What's this one? What's this one here?
32:10
This is us in the studio. A little meta.
32:12
Oh, that's extremely demonic. Aliens come and steal. This is super creepy. I don't like this. Boom. Bring the horseshoes. Bring the air horn back. Oh, that's weird. They steal the gong. Wow. Hold your position. Okay.
32:24
I think there's one more. Yeah. Let's play the last one. What you got? What's this one?
32:31
That last one was bizarre. Very very Yes.
32:34
So, the physics um are pretty solid.
32:37
Sometimes you you see a bit of the the same thing with V3 where like if a car is driving, right, you'll see the back of the car.
32:43
Okay, we got us standing at the pool.
32:45
Let's take a look at this. Okay, this one's cool. Okay, we're back.
32:49
The lighting on that incredible. Okay, finish strong.
32:51
10 out of 10 for mid journey flight again. There we go. I love this. This is great. That is cool.
32:58
It looks really good, too. Yeah.
33:01
Always bring your F-35 Joint Strike fighter into uh the club pool. Something like that. I don't know. But okay, that was cool.
33:09
I I I like that one a lot.
33:10
Yeah, lot less demonic when took us on a bit of a roller coaster there.
33:13
I did not like that angel. That was weird.
33:17
That was That was very weird. Bit creepy.
33:19
The aliens were very bizarre, but you've redeemed it all. You want it all back. Uh fantastic.
33:23
Uh give us a review of the uh like overview of the actual experience.
33:29
Uh how long does it take to generate these?
33:31
Are you hitting rate limits? How much is it cost?
33:32
to give us like the breakdown of like, you know, the consumer experience.
33:36
Uh, yeah, it's really good. I mean, I I think so.
33:40
I'm on the $30 a month plan, which is kind of the mid-tier one. Um, it's very fast.
33:45
I mean, it takes probably 10 seconds, 15 seconds. It's really fast. Okay. V3 is like two minutes.
33:51
So, you can iterate like super quick. Oh, that's cool. Okay.
33:52
Um, but it's very easy to use.
33:54
I mean, I I haven't used I So, I'm actually not on the Discord. I'm on the website. Okay. Yeah.
33:58
Um, but it's very easy to use. Yeah. Okay. Very cool.
34:00
And you can run them concurrently, so I could do multiple. Okay. At the same time. That's great. But yeah, really great. Awesome. Well, very fun.
34:08
We'll be tracking it more, asking people how how it's benchmarking, how it's working.
34:10
We'll have to get uh have some fun with those.
34:14
I had a lot of fun with the V3 ones.
34:14
We were doing the the crashing through the Hollywood sign.
34:18
Uh couple few too many bottles of Don Perry on the back of the Ferrari.
34:22
Yeah, I didn't like how there weren't guardware.
34:24
There seemingly weren't guard rails.
34:26
That's an AI safety issue.
34:28
I shouldn't just be able to You should be able to visualize yourself drunk driving bottles of champagne flying out of the car.
34:34
The quality was remarkable.
34:36
Uh anyway, MidJourney is having fun on X.
34:38
They say introducing our V V1 video model.
34:40
It's fun, easy, and beautiful.
34:43
A available at $10 a month.
34:45
It's the first video model for everyone and it's available now.
34:47
How many how many prompts do you get for $10 a month? I don't know. You want to look it up? Yeah.
34:52
I mean, I think it's actually unlimited, but it just takes longer. Takes longer. Okay.
34:58
I'll verify that that's insane when you put it into the context of those outputs are in many ways better or on par at least like from an entertainment value standpoint as VO and VO is $500 a month still gated on I could only do three per
35:16
day I have to come back they take two minutes a pop yeah speed of iteration is really really key I mean that's the whole Discord model is like is like get people iterating sharing ideas like to explore the space and figure out what works. Like even just from seeing those
35:30
Like even just from seeing those four, I feel very confident about its ability to render aircraft.
35:35
And so I'm probably not going to go and prompt a bunch more alien videos, but I'll definitely be prompting a bunch more F-35 videos because it seems to do that really, really well.
35:47
And so the more people you have making more stuff, the more you you learn the guardrails, learn how to use it creatively and and can actually make a better product.
35:54
Mid Journey was having some fun.
35:56
Uh Devin Fan from XAI says, "I know what I'll be doing this weekend."
36:00
And Midjourney says, "What weekend?"
36:04
And Will Depuse says, "Lmao."
36:04
Um Blake Robbins says, "Midjourney video is breaking my brain."
36:09
And everyone's having a good reaction to this.
36:11
So, always fun to have a new AI tool.
36:13
And we'll be we're talking to a couple AI folks uh on the show today.
36:16
So, we'll be running through that, getting their reactions and talking more stuff.
36:20
Uh Elon Musk posted the uh the very sad what is this?
36:28
The peipo peep pepe or something. It's the green frog.
36:30
He's smoking a cigarette. He's not happy.
36:32
Uh probably because RIP to 4 a. m. Brutal.
36:41
Um, uh, we Ashley Vance, we should I I I I think a a picture of Elon, you know, smoking a heater after one of his rockets blew up blew up would become a timeless meme.
36:56
That would be it would be worth kind of his comm's team kind of working on putting that together.
37:01
Maybe working with Ashley Ashley Vance to get that shot. Yeah, it could live.
37:04
So, girls say, "I can't believe he didn't cry at the Titanic.
37:08
Do men even have feelings?
37:10
Boys crying at the sight of ship 36 exploding. Very, very sad.
37:13
And then Elon says, "Just a scratch."
37:16
The entire thing blew up.
37:19
Uh it was just a flesh wound.
37:22
It was intense watching it.
37:25
Um I mean the the ball of fire here was immense.
37:27
So uh the Starship exploded during a test in Texas, a setback for Mars' Mars uh for Musk's Mars ambitions.
37:35
Now, the Mars transfer window is very, very tight.
37:38
Like, you can only get from the Earth to Mars like once every 18 months or something, maybe even more.
37:46
It's really hard because like if the planets are on the opposite side of the solar system, like you just can't like even though you have a rocket, you just can't get over there.
37:52
So, you have to wait until they're lined up and then you can do it.
37:55
But, but realistically, skill issue.
37:58
Realistically, true skill issue.
38:01
If you build an even faster rocket, you could get there no matter what.
38:02
You just pilot, steer it around like it's a GT3 RS around the Nurburg Ring. No problem.
38:09
Uh, so the explosion occurred during a static fire test.
38:10
No injuries were reported. Thank goodness. We love autonomy.
38:14
Very, very happy to hear that no one was injured um during this because it looked horrific and it looked like in any other scenario there would be a bunch of technicians there, but that fortunately they're able to do everything remotely, which is great.
38:24
Um, and then Starship FE's pressure to meet deadlines for NASA's moon mission and Mars exploration.
38:31
And so there's a big uh there's a big NASA moon contract that's very important, very material to the business.
38:38
Obviously SpaceX has a lot of other business lines, but uh this one's very very important, too.
38:42
And it's uh and we hope they can get back on track.
38:46
SpaceX is make making an enormous bet on Starship, which stands roughly four 400 feet tall at liftoff as it tries to break ground with new reusable rockets.
38:54
And the the paradigm of of Starship, it's not just a bigger rocket. It is way more reusable.
38:59
like you look at the thing, it comes down, gets caught by those arms, can instantly be refueled and sent back up.
39:06
You're you're talking about potentially like multiple flights per day.
39:10
Um, and so the problem here is not can you build a big rocket.
39:13
Humanity has done that before.
39:15
Humanity's built a rocket that's roughly on par.
39:17
We we've gotten to the moon before.
39:19
The the challenge now is not can we get to the moon.
39:23
It's the same thing with like the challenge is not can we build a flying car or can we build we have helicopters.
39:28
Can we build one humanoid robot or one self-driving car in San Francisco?
39:31
It's like, can we actually scale these systems to the point that it is safe to go to the moon and back on the drop of a hat for 200 bucks?
39:39
Like, that's the challenge.
39:41
It's it's more of an economic and industrial might challenge.
39:42
And that's a completely different challenge from just can we get one rocket to the moon, an exquisite system.
39:48
We're looking for reusable, scalable uh you know engineering uh systems.
39:53
So, uh, so good luck to Elon rebuilding and the entire SpaceX team.
39:59
I'm sure it's a huge challenge right now.
40:02
Um, but, um, if, uh, uh, let's do some ads, tell you about Adio, customer relationship magic.
40:12
Adio is the AI native CRM that builds, scales, grows your company to the next level. Get started for free. Adio. com. Adio. com.
40:18
You can use Wait, what is that?
40:21
What is that sound at the end?
40:23
Is that attached to that soundboard?
40:25
Guys, I think you botched uh the action movie sound effect. So, it has on that. Okay.
40:30
Uh, in other news, the Los Angeles Lakers has been sold for 10 billion in richest deal in sports history.
40:38
Guggenheim Partners CEO Mark Walter, who also owns MLB's the Dodgers, is acquiring the storied NBA team in a move that makes it the world's most valuable sports franchise.
40:48
And it's so funny because like the the Wall Street Journal's framing this is like this is the biggest deal ever.
40:53
No one's ever done a deal like this.
40:56
And we're like wait so you're talking about like a series A for like a foundation model company like as a tech person.
41:00
I'm just like yeah like at$10 billion like billion dollar.
41:06
I mean we should ring the gong but it's not exactly like the first time.
41:11
It's not even the first time this show we've heard a a decorn.
41:22
Congratulations to uh to the Lakers, Mark Walter and the whole team. It's it's fantastic.
41:27
Uh major premium to the Boston Celtics who sold for 6. 1 billion.
41:34
Um and now the Lakers is the most valuable sports franchise.
41:35
Um but they just don't do enough volume.
41:38
There's only a couple games, you know, they're not 247.
41:41
Like Instagram, does that ever go offline? No. No.
41:45
There's always entertainment.
41:46
Lakers, they're still doing seasons.
41:49
They need to have 24-hour basketball.
41:50
They want to really get there around the clock.
41:52
It's like endurance endurance basketball.
41:57
It's just a week long game, you know?
41:59
Got to always have five players on the court just constantly running up. It's the only option.
42:03
Uh, Jeannie Bus and her family, who have owned the Los Angeles Lakers since Jerry Bus bought the team in 1979. Wow.
42:11
On Wednesday, agreed to sell majority control of the story team to Mark Walter, the sports investor.
42:14
And I and I looked at the uh the the return on investment of owning the Lakers for that 40 years, slightly under S&P 500.
42:27
Like it was a really really good deal and it was a really great company that grew a lot but it didn't outperform the stock market.
42:35
Just diversification bros DCA bros undefeated again.
42:40
Well if you're trying to DCA do it on public.
42:43
com investing for those who take it seriously.
42:44
Multiasset investing industryleading yields.
42:45
They're trusted by millions folks.
42:46
Um anyway, uh Walter, who is part of the ownership group that owns the Dodgers, has been part of the Lakers since 2021 when he p purchased a 27% minority spa stake in the franchise.
42:58
He's also a co-owner of Chelsea in the English Premier League, the WNBA's Los Angeles Sparks, and the new newly formed Cadillac Formula 1 team. Let's hear for Cadillac. Let's go. Let's hear for Cadillac. Yes. Congratulations.
43:14
John, you know, can't hear you at all.
43:16
John front run the uh the Cadillac F1 team had got a Cadillac for himself over there. You can see the black.
43:21
It's great to have an American F1 team in the business now. Yeah. Yeah.
43:25
We've we've fallen off, but we're coming back.
43:28
You're not going to be able to get one of these in the whole country. I don't think so.
43:31
They're going to be too popular after the F1 team, you know, gets out on uh the sale marks the end of nearly a century of Lakers control by a family that has become synonymous with Los Angeles sports and the glitz of professional basketball.
43:42
The deal also comes at a time of skyrocketing valuations in professional basketball which haven't come back to earth since the league announced a media rights deal last year with worth 77 billion when the Celtics sold in March. The $6.
43:55
1 billion valuation exceeded the previous record valuation set for a sports team uh by the 6.
44:02
05 billion sale of the NFL's Washington Commanders in 2023.
44:10
He purchased the Lakers for 67 million in 79 1979.
44:13
The team transformed from franchise uprooted from Minnesota into one of the winningest and most valuable sports.
44:21
I had no idea that they were founded.
44:22
That's where the lake the lake name comes from. Interesting.
44:24
Minnesota is the land of a thousand lakes.
44:27
They were the Lakers because there's a lot of lakes in Minnesota and then they just put them to uh they just brought them to LA and kept the name.
44:35
But that's what Lakers means. Yeah. Wow. bus.
44:37
Uh the bus family oversaw the creation of Showtime and presided over the NBA's last repeat.
44:45
A-listers like Jack Nicholson and Leonardo DiCaprio have become fixtures at the games.
44:48
And when they sell merch, they need to pay sales tax.
44:51
They should get on numeral. com. numeralhq. com. Sales tax on autopilot.
44:56
Spend less than five minutes per month on sales tax compliance.
45:02
You know all the 11 championships since 1980.
45:04
Their rosters have boasted many of basketball's brightest stars.
45:07
Magic Johnson, Kareem Abdul Jabar, Kobe Bryant, Shaquille O'Neal, and LeBron James, and LeBron James's son have all worn the Lakers purple and gold. I love it. It's such a cool.
45:18
Yeah, the the father-son duo.
45:20
I mean, I feel like that should have been a bigger like national news story. It's such a cool thing.
45:24
I I think it's like not like if they were like winning championships together immediately that might be a different story, but it's just so insane that you could be playing professional basketball because you could have earned a better return by DCing into the into the into the stock market.
45:41
That's not why people own these assets though.
45:43
Owning the Lakers for a number of decades, I imagine, was absolutely priceless.
45:48
So, um, great investment, the owners, great run.
45:53
Yeah, all the perks you have to add those in. What do you get?
45:56
Perks from DCAing into the I like how uh Lakers legend Magic Johnson hit the timeline said, "Just like I thought when the Celtics sold for 6B, I knew the Lakers were worth 10 million." Let's go.
46:08
The confidence of Magic Johnson. Great investor, too.
46:12
He's got a bunch of bunch of good stuff in the portfolio.
46:14
Um anyway, uh more news on the Scale AI uh transaction. So, it's closed.
46:20
I believe that Alex Wong has a badge at Meta and shows up to work at at in PaloAlto and and clocks in at Meta HQ.
46:30
Now uh scale AI is still an ongoing concern is still a company but every competitor is out for blood and they want to take as much of the business as they can since obviously the perception is that scale AI will primarily be working with Meta and that other foundation model labs might not want to do business with with scale AAI anymore.
46:52
unclear if they can separate out the businesses, if they can separate the them out fully over time and and sell the position to other investors, create like a diversified I mean even they could even take the company public uh at which point uh I imagine that it would be a lot less uh a lot less of a conflict of interest or like a fear.
47:10
Um but there's been news that that uh OpenAI said hey we're not training we're not using Scale AI for data anymore um because it's too aligned with our competitor Llama maybe.
47:19
Um, but everyone's trying to Yeah.
47:22
A lot of this was very predictable. Yeah. Right.
47:24
I don't think Meta and Scales teams looked at and said, "Hey, if we sell right now to Meta, which is competing in open source AI, we're totally going to retain all of our customers, right?
47:37
Like people aren't just going to immediately churn off."
47:41
And no, they they were smart enough to know what would happen.
47:43
And there was an article, I think, yesterday about OpenAI, you know, ending their relationship with scale.
47:49
But from from what we knew like they hadn't been doing much for a while.
47:52
It's part of the reason why Merkore had been and they also brought a big function in in house because for some of the more complex tasks it makes sense to generate the reinforcement learning data yourself.
48:03
Um and there's just so many others there's so many other services having like a single point of failure never makes sense for a business of that size but uh we'll see.
48:12
So the uh the information has an article here about a littleknown startup that has surged hint hint past scale without any investors. This is interesting.
48:21
After meta platform scale deal, data labeling is looking like Silicon Valley's hottest new interest.
48:26
That's enormous opportunity for Edwin Chen's surge AI.
48:28
Uh for years data labeling existed in a tuckedway corner of Silicon Valley, a critical but unglamorous area of AI where companies like Google and OpenAI how hire outside firms to improve their models by laboriously grading the quality of what they produce.
48:44
Now a spotlight has unexpectedly fallen onto the field in the wake of Metaplatform's decision to pay 14.
48:50
3 billion for 49% of Scalai, the best known data labeling firm.
48:55
But it's not the largest such firm, nor perhaps the most impressive.
49:01
That title belongs to Serge AI, founded by Edwin Chen. This is fascinating. I didn't know this.
49:06
1 billion in sales last year. Bigger than scale. Yeah.
49:11
So Chen startup has won customers like Google Open AI and Anthropic.
49:15
such a it's such a testament to the idea that like sure you can bootstrap, but you it's it's so incredibly hard to have any hype around your business if you're bootstrapped because you're not having your investors aren't hitting the timeline for you. Yep. On a daily basis.
49:29
And also, you have if you're not trying to raise capital, you have less need to go and be loud and and go on podcasts and and talk to the press and all this stuff because you're just making a lot of money.
49:42
And you know, sometimes it can be beneficial for to for people to not know about you.
49:49
So this is I mean this is crazy crazy stats. So Chen is 37.
49:51
He has no investors and has bootstrapped a 5-year-old startup entirely by himself which has 110 employees in offices in New York and San Francisco.
50:00
The company generated more than $1 billion in revenue last year.
50:04
Serge has told employees a previously reported figure that exceeds the $870 million scale generated in revenue uh during the same time period.
50:15
And unlike scale, Serge was profitable and has been from the beginning, Chen said.
50:19
Moreover, Serge could see its sales get even larger if other companies copy OpenAI's decision to stop hiring scale, a choice made over concerns about scale's relationship with Meta to shift business to Surge.
50:28
Other key financial metrics couldn't be learned like how much revenue surge keeps after paying its workforce of mostly contractors.
50:35
So there is a question about like the margin since this is somewhat of a marketplace business.
50:39
This could be a situation where you know uh a $1,000 contract comes in and $800 of that contract goes to the actual contractor who's doing the work of the data labeling.
50:50
But at the same time, even if it's 200 million in like you know like grow net revenue, that's still a huge business.
50:58
I I it's hard to imagine Serge not being a fantastic business if they haven't had to raise money.
51:03
They have 110 employees and they're used by Google and all these major foundation model labs.
51:07
So, uh seems like a fantastic business.
51:09
Uh but if Serge could earn a valuation from investors similar to the one scale received for Meta, such a price would make Chen a billionaire many times over, at least on paper, and quietly one of the wealthiest people in tech. Interesting.
51:22
I'm very interested to see what uh what he did before this company. uh Edwin Chen.
51:27
I feel like I've heard that name before, but I don't know.
51:29
Um as AI models transform from toys into real business tools, data labeling is becoming more and more essential.
51:37
Contractors hired by like by companies like Serge grade the responses from AI models and write thousands of questions and answers in fields like programming, math, and law to feed those AI models.
51:47
And so, you know, if you're I I I wonder if this is going to go the route of, you know, you are Deote or McKenzie and you're going to have your team, but then also a company like Serge create a ton of training data around a specific workflow that is costing your business, you know, 20 or 50 or hundred million dollars every year.
52:06
And then so it's like instead of like the the AI BDR that's like kind of generically writing emails based on like the average of the entire internet.
52:16
It's like no this is a fine-tuned for your business perfectly trained perfectly and it and it really it really distills what you do excellently.
52:26
Um yeah I don't know I don't know if it'll go that way.
52:28
I'm interested to talk to people about it.
52:31
uh as AI models uh so Serge's subsidiary data annotation tech says workers get paid to train AI on your own schedule with wages starting at $20 an hour.
52:39
Chen has distinguished Serge by making uh it the high-end shop charging premium rates often two to five times what scale might bill.
52:49
Surge justifies the prices with its reputation for industryleading work.
52:53
Indeed, one former Scale employee said Serge often performed better than scale in customer audits of labeling quality.
52:59
and competitor Garrett Lord, who's coming on the show today, uh who runs Kleiner Perkins backed Handshake, readily acknowledged that Chen is the number one player.
53:06
So, I'm excited to talk to Garrett Lord today about this exact topic.
53:10
Should be very interesting.
53:12
Uh, you wouldn't know that from the from the coverage of Meta's Blockbuster deal to quasi acquire scale AI, its CEO, Alexander Wong, who is now joining Meta in a senior AI role, was widely regarded as leader of the data labeling field and had become a Silicon Valley celebrity, blanketing podcast and conferences with his presence and posting heavily on X. It also raised 1.
53:31
5 billion in venture capital, putting scale on a very short list of companies that have raised that much.
53:35
And he hired upwards of a thousand people.
53:37
Wong had made it time to exit perfectly given the traction of Surge which had grown larger than Scale without outside capital and with a tiny fraction of Scale's workforce.
53:45
Scale also missed the goal to hit a billion dollars in revenue last year but scale scale spokesperson said the company scale wasn't profitable either.
53:52
Was not profitable which but wasn't burning a ton of money like I think they had they raised 1.
53:58
5 billion and they still had like almost a billion in cash. Yeah. Yeah.
54:01
So, they weren't they weren't in like trouble or anything, but at the same time, it was like like not not a wildly profitable, not a wildly lean business, but I don't know what what what a it's absolutely fascinating these two businesses. It's a wild industry.
54:15
Uh something that like Yeah.
54:17
I mean, just it feels like there's such an edge just to even identifying this opportunity years and years ago.
54:22
years and years ago. I mean I guess search started four or five years ago but it was certainly like pre- chat GBT that all these companies got started and then they realized like some of them got started in self-driving car annotation all sorts of stuff like that but uh Chen studied linguistics and math at MIT came
54:38
to the idea for his startup after leaving college and witnessing firsthand how big companies struggle with data before starting Serge Chen worked as m machine learning engineer at Facebook Dropbox Google and Twitter he worked at four different tech companies like just like going from one to the next. That's That's insane.
54:56
He was developing recommendation and search algorithms and helping gather the data needed to train them.
55:00
Despite the hefty resources of those companies, Chen encountered a lot of problems.
55:04
At Facebook, for instance, Chen was tasked with helping build a Yelp competitor.
55:09
His team needed to train a model that correct could correctly classify businesses telling the difference between restaurants and grocery stores, for instance.
55:15
to do so they need a data set containing 50,000 accurately labeled businesses which he found out would take six months for an outside firm to assemble.
55:25
We had no solution other than waiting. We simply waited.
55:27
When the data came back, Chen blanched.
55:29
In some instances, it had labeled restaurants as coffee shops and coffee shops as hospitals.
55:34
The data was complete junk.
55:36
He wouldn't say which vendor Facebook had used.
55:38
In 2020, in 2020, he left Twitter to found Serge and picked up some of his first customers.
55:44
executives from Airbnbs and Neva, a once promising AI search engine startup, as only uh as only a founder in San Francisco might, bumping into them at rock climbing gyms in the city's dog patch neighborhood and the mission district.
55:56
Talking up his startup to get Serge going, Chen recruited data labeling contractors he knew from his previous roles and funded the startup using his savings.
56:04
He wouldn't say how much he put in.
56:07
Fortuitously, Chen focused on language modeling.
56:10
Scale, by contrast, started out using more visual data for autonomous vehicles, which we talked about. Wow.
56:14
Just as those types of models began to grow in importance, uh, less than a year later, OpenAI had hired Serge to fine-tune its models by teaching them how to avoid producing harmful responses like a racially biased language biased uh, based on research paper the company published together in uh, by 2022 anthropic can become a surge customer.
56:33
They're putting out research papers with OpenAI and still managed to stay this under the radar. Wow. Yeah.
56:38
So look at this uh the label large s data labeling has proved to be a lucrative niche in AI.
56:44
Uh Serge founded in 2020 has over a billion in revenue zero funding.
56:49
Uh scale founded in 2016 hasund 870 million in 2024 raised oh the this is this says funding raised but this is clearly valuation or something because it says 17.
57:02
4 billion which is not what they raised.
57:04
Um, Touring has 300 million annualized, raised 225 million. Invisible. It's interesting.
57:09
Touring too initially was like a marketplace to just hire developers and I think they pivoted into lab data labeling. Interesting.
57:16
Uh, it's the same thing when I work with a cloud provider.
57:21
The enterprise tech customer said, "I don't know the internal expectations for why their services work so well.
57:25
I push a button and I'm glad for the internal work to make that happen."
57:29
And data labing companies typically use various techniques to make sure contractors aren't just dialing it in or phoning it in, I guess, uh, when answering questions.
57:38
For instance, the companies randomly insert questions that have no correct answers or make sure labelers agree on the right answer to a question.
57:45
So, obviously, you scaffold up these these uh these these like responses so that everything's like double-cheed and then you can kind of see if people are are messing around.
57:53
But, wow, what what a beast of a business.
57:56
I had no idea how big this thing is. Amazing.
57:58
Uh, how'd you sleep last night? I'm on a comeback. I got an 89. Go to eightleep. com. No, get an eight. I know what I got.
58:05
Uh, 5year warranty, 30 night risk-f free trial, free returns, free shipping. What'd you get? You got 100 88. Let's go. Soundboard. I demand soundboard. Ashton Hall, let's go.
58:17
John, did it beat you by I had a terrible night.
58:21
I also took a nap after I got home. It was fantastic.
58:25
Uh we have our next guest coming into the studio.
58:27
Uh Mike from Arc AGI breaking down how all the different models are doing.
58:33
Last time he was supposed to come on Elon.
58:39
Trump decided to get in a massive timeline war. It was brutal.
58:42
John wanted to power through it.
58:45
I said John the people in the YouTube chat are are uh demanding are demand that we do a timeline. we don't do it.
58:54
They will put up attack ads against us.
58:56
They will buy billboards against us. They will go to adquick.
59:00
com and put up attack ads on TBPN if we don't cover the Trump Elon dust up.
59:05
So, we did and now we're getting those folks back on the show today and later.
59:09
Uh, but if you want to take out an attack ad on us, buy a billboard and adqu out of home advertising made easy and measurable.
59:16
Say goodbye to the headaches of out of home advertising.
59:17
Only adqu combines technology out of home expertise and data to enable efficient seamless ad buying across the globe.
59:23
Uh we should do some timeline.
59:26
I I want to do this blank street story but we can do that later.
59:30
Let's dig through some uh what's in the timeline while we while we wait for Mike to join.
59:34
Oh, we have Mike in the studio now.
59:36
Welcome to the show Mike. Good to see you. How you doing? Good afternoon.
59:43
Thank you so much for I'm glad there wasn't a major breaking drama story today.
59:46
Was actually able to show up. Yes. Yes.
59:48
You you I I don't know if you watched that show at all, but like I was just sitting here. John was so locked in.
59:54
Wanted to just keep doing the show and I'm like messaging him like, "No, we we like actually have to."
59:58
That was a terrible day to launch anything new and we like launched something.
1:00:03
I saw several startups launch stuff.
1:00:04
Uh like regrets to everyone you try to get anything out that day.
1:00:08
Rahul actually I remember them.
1:00:10
It was it was Rahul and uh Yeah, that's right.
1:00:12
Julius had a launch and uh what what what's the voice cloning company? 11 Labs.
1:00:16
I think they launched something.
1:00:18
Well, and then and then I think Lulu said something.
1:00:19
She was like, "If you have bad news, today would be a good day to drop it."
1:00:22
And then OpenAI actually flagged like, "Hey, we had this like massive, you know, uh Oh, yeah.
1:00:29
This this dust up with the government, right, where the government was like, you have to give us your all we don't want to do this." Yeah. That was like serious.
1:00:38
Like I mean, that got um we're still looking at actually, you know, end results of that.
1:00:43
But that that went really deep into the world.
1:00:44
went really deep into the world. I feel like much more than um you know kind of maybe even got reported on like every single chat thread I was a part of was basically like hm should I like stop using chat GPT as much um to catch because it feels like anthropic has a similar policy I it seems like
1:01:02
Google might have a similar policy like there was that story a year ago about uh a man who was using a Google phone with a Google Fi uh cellular connection and had all of his data stored in Google uh Google Drive and Gmail and he took a picture of his of his child to send to a doctor and it was kind of like a nude
1:01:23
photo of the kid to inspect the child for like a physical medical problem and it got flagged as child abuse material by an automatic system and the automatic system basically deplatformed him from everything Google and so he lost his email, his phone number, his all of his drive stuff and it was like a false
1:01:42
positive but it was really hard for him to get back through there and So I guess my question is like like it seems bad when we hear the story in isolation but maybe the problem is not the individual company and it's instead like the government policy and this applies to all the different companies but I don't know two two things can be true. One is
1:01:56
One is that it can be a massive overreach by that court to say you know basically you need to eliminate privacy on your platform.
1:02:06
Yeah and you can simultaneously have questions around maybe I should use this product in a different way. Totally.
1:02:13
Um and and it's the the inflammatory nature of it is that people use chat GBT as like a confidant that totally yeah and tell it things that they wouldn't tell anyone.
1:02:23
They wouldn't tell anyone in their life and they're and they're having those conversations and I think that's why struck such a cord because like that's true.
1:02:30
Um I I just saw some reporting from uh uh CO this week that like chat minutes per day are up to like 30 minutes a day now in usage.
1:02:36
Um, and like it's not it's like closing the gap with like Instagram, which is just sort of nuts to think that like I mean who would have thought a productivity tool would ever like be on par with like a social like media app, right? In terms of daily usage.
1:02:49
It's so but the interesting thing is it's filling a similar void, you know, it's like it it's delivering uh digital companionship in maybe the way that social media products historically did without any social element to it at all.
1:03:05
just like there's one one to one.
1:03:08
It's interesting to think like, you know, we went from like, you know, what your friends are doing is like the most interesting thing to like what the Kardashians are doing is like the most interesting thing to like actually maybe the most interesting thing is like this person that knows everything about you and is always on and always willing to talk and you know, you know, who knows? Yeah.
1:03:28
I think the consumer habits are being formed around stuff today. Yeah. Yeah.
1:03:32
I mean, I I find myself all the time like instead of scrolling YouTube looking for an information like an interesting video essay to explain how I don't know like global shipping lanes work or something like that, just going to chatt and saying, "Hey, like break this down for me and then I can just ask a follow-up and dive exactly to the layer that I want."
1:03:50
And so yeah, I'm definitely in that camp of using Chat GBT just as like an exploration and entertainment education tool, an infotainment tool much more than Instagram right now, at least for me.
1:04:01
Um, but uh enough about that.
1:04:01
Uh, what is new in your world?
1:04:05
How should we how should we frame kind of the the current horse race between all the foundation labs? Yeah. Okay.
1:04:12
So, um, I'm going to share a link.
1:04:14
I don't know if this is something you all can pull up or pull it up.
1:04:18
This is a post that we published a little over a week ago.
1:04:21
Um, so you know, I think there's been this really big like what's the frontier right now in sort of AI progress, right?
1:04:27
The the the massive shift in the last six to nine months has been moving from this regime of like scaling up pre-training with more and more labelled data into these like test time test time compute test time adaptation regime.
1:04:40
People call these data reasoning models, right?
1:04:44
We're getting these models to think out loud additional data.
1:04:45
Um, every major lab now pretty much at this point, uh, I I guess except for Meta H has one of these, uh, systems that we've been able to test and and report results on.
1:04:56
And, um, I think there's some really really interesting stuff we're starting to see.
1:05:00
Um, I I think the most notable thing is that like there's not an absolute clear winner across the sort of like landscape right now.
1:05:10
There's basically a sort of prito frontier that's emerged.
1:05:11
frontier that's emerged. One of the most important things if if listeners are like listening here I think you should take away is that like any any anybody who gives you a benchmark score on an AI system that is a single number is is is uh is just marketing to you because the reality is now with these AI reasoning
1:05:29
systems you have to report score on like a two dimensional object you have to consider cost and efficiency alongside the accuracy and all these different lab providers have come out with different AR reasoning systems that sort of score differently they're trading off cost per accuracy at at different points. So like
1:05:43
So like if you want just like the absolute highest horse, you know, highest rost horsepower sort of cost and time is no option, 03 high is going to be your like clear winner today for that.
1:05:52
But if you're somebody who's saying like, you know, hey, I want to plug in an air easing system into an existing product I have or I want like faster answers and I'm willing to sacrifice some raw horsepower for generality for like quicker response times, lower lower cost, you might look at something like rock or Gemini 2. 2. 5 thinking.
1:06:06
rock or Gemini 2. 2.5 thinking. um there's not like a single like best best answer which which I think is pretty interesting and we haven't seen like this sort of frontier is I think what all the labs are really working to to try and figure out okay how can we get accuracy as high as we can but also we
1:06:20
got to try and keep cost as low as we can down in the human efficiency real yeah I've noticed that more recently with uh kind of my default usage in chat GPT uh 40 seems super fast but I always am thinking like oh I should maybe put this in 03 pro but do I want to wait 10 minutes and and I'm making that kind of like economic calculus there. Even
1:06:41
Even though because I'm on the on the pro plan, it's not an economic cost, not a direct I'm hitting a $5 time for you, right? Like 13 minutes. Yeah, exactly.
1:06:52
And so I'm kind of like doing a 40 thing over here and then switching back and forth.
1:06:55
It's very it's very odd paradigm that I we never really had to deal with in computing necessarily before.
1:06:59
I mean, I guess like if you were downloading like the 4K illegal Blu-ray versus the Yeah, I will which of course we never did purely hypothetically, but if you were on a torrent site, there was a time tradeoff between watching a screener.
1:07:11
tradeoff between watching a screener. Um I think this is actually one of the reasons these AR reasoning systems I would assert and I I don't have inside baseball in the data but like from the outside looking I think there are some interesting suspicions that would suggest that these like AR reasoning systems at least today in their current
1:07:27
form have like relatively weaker product market fit um compared to the uh like non the non-reasoning based systems right the pure language model based based things interesting that's a huge violation of the the like the deepseek narrative that I felt like was really bubbling up was deepseek came out with
1:07:43
like the first just like open access reasoning model like reasoning had been tucked behind the open AAI payw wall and so the prousers were familiar with what reasoning models could do that everyone was very excited about them in tech or in the early adopter crowd um but but DeepSeek when that app came out and you could just install it and instantly see
1:08:04
the reasoning chain uh it felt like everyone's like oh everyone's going to be addicted to this forever and this is going to be the new paradigm but it seems like that might not necessarily be happening Uh Jordan, do you have something or I wanted to talk about like spiky intelligence and how that plays into this? Uh we had this uh someone
1:08:18
Uh we had this uh someone came on and said like I think it might have been Cholto actually talking about ARGI just saying like hey all the foundation labs kind of have like a truce that we won't reinforcement learn specifically against arc.
1:08:29
Um, I don't know how real that is from your perspective, but it feels like increasingly we might see like very task dependent RL runs kind of chipping away at specific things like like IMO level math is something that clearly like there's a ton of work to be done on, but we don't have as many verifiable rewards for for poetry or comedy writing.
1:08:52
And so that'll be a little bit messier and later down the road maybe.
1:08:57
But at the same time, there's probably other verifiable rewards that are just smaller pockets of value here and there that for these little microtasks.
1:09:06
And so I'm wondering if we will ever see um like the marketing language around these models evolve like Grock kind of did this with like we are the anti-woke one, but that was more just in the overall like temperature or the vibe of the model.
1:09:20
But I'm wondering if if there'll be an idea of like this one's really good at math, this one's really good at research, this one's really good at that, or if they're all kind of going down the same path with what they're trying to solve.
1:09:30
I I do think you probably are going to see some domain specialization.
1:09:34
I think my guess over the next 12 to 24 months is that you'd see some domain specialization benchmark scores emerge because of how all these labs are starting to do the next evolution of training which is they're using RL environments to generate synthetic CO2 traces doing their sort of model trainings on that data and they're trying to go get it on a lot of different just different domains.
1:09:51
different just different domains. um you know the 03 the original 03 paper you know I think was interesting on the benchmark results where you know on this new sort of coot reasoning system they had relatively high scores on math and coding um but but the the the gap um or should say the step function increase in
1:10:09
those scores was much higher than the increase in like legal reasoning which you would sort of maybe intuitively guess that or think suggest or expect that like legal reasoning would probably be one of the best like general domains for if you trained a reasoning model that was really good at math and coding like it should be like and it's a
1:10:25
language model that like that would directly transfer into like the legal domain because like okay it's symbolic you know reasoning that's like self-consistent and that wasn't the case so I think that's I suspect that's what we'll see there you know there's obviously the big scale news um the thing that I'm seeing now is there's uh
1:10:40
probably like I don't know a handful that I know these new startups that have come come up in the last several months but all getting founded to basically go build RL environments to generate synthetic or semi-ynthetic data uh and like selling them to to sort of the major labs or to the major frontier folks building these nextG systems. Um I
1:10:55
Um I I think we're going to see more of that.
1:10:59
I expect that's kind of what what's going to drive a lot of areas.
1:11:00
What do you what does the data labeling market look like?
1:11:04
Today we are covering Surge AI, which a lot of people weren't familiar with.
1:11:08
I I I'm sure I'd seen it at some point, but I was certainly not familiar with it until we covered it today.
1:11:15
What do you think the table labeling market uh looks like in in five years?
1:11:20
Do you think that scale was getting out kind of at the perfect time?
1:11:26
Uh, you know, I'm curious.
1:11:26
I I think the timing was pretty good.
1:11:29
Uh, I mean, like look, the the macro change here is is from a regime where like we're pre or scaling up pre-training.
1:11:36
We want as much text, as much high quality label text as we can get our hands on to to scale these these foundational models into one where we're trying to train process models or, you know, make the foundation models really good for process thinking and coot generation.
1:11:46
M um that that is a complete shift in how you want to generate that data.
1:11:50
You want in our environment that you can create lots and lots of coot traces really very long traces over longrunning tasks as well and you can feed all that right back in and then take advantage of the scaling laws we already know about language models and how they how they work and how the performance increases as you can get more examples of the data there.
1:12:02
So you know my my like I guess like macro bet would be you know sort of the trend is heading down towards like the pre-training scaleable stuff significantly higher on RL environments. Yeah.
1:12:14
Yeah. So, so when when when you say RL environments, what you're talking about is moving from a paradigm of I go to a data labeling company and they hire a ton of contractors to generate new text or verify the responses or grade the
1:12:30
responses from these models to I am now hiring top machine learning engineers, AI scientists and having them design a an environment that the that the rein reinforcement learning can happen like autonomously within the system, right? Or we are effectively like these new
1:12:48
Or we are effectively like these new startups that you mentioned, they are taking the massive like hundreds of thousands of contractors like out of the loop for the next runs. Is that correct?
1:13:00
Yeah, it's synthetic or semiynthetic in some cases.
1:13:02
Um, like example companies here like mechanized work is one that got started recently that's doing this stuff.
1:13:08
Uh, morph I think is doing this stuff.
1:13:10
Habitat's another world environment that's sort of doing similar stuff. Sure.
1:13:12
Um there's a there's just a lot and it's like very emerging.
1:13:14
All these many of these got founded in the last like couple months. Yeah. Yeah.
1:13:17
Um and I think that's a function of the demand and the pull from a lot of the frontier research groups that are wanting this data uh to sort of do their altering stuff.
1:13:25
So So do you do you imagine uh companies like Serge and other players would try to pivot into this if if they're expecting I would expect that founder companies like that would recognize this is growing part of the business and have bets in it if they don't already. Yeah. Yeah.
1:13:41
I I was always wondering like the the whole story of scale was kind of a series of like various booms in training data.
1:13:50
Like the first one was uh was data labeling for autonomous vehicles and it seemed like that grew very very quickly and then the the training paradigm around Whimos kind of shifted away from hey we need more and more train labeled training data to something else and just having the cars on the road and and and generating real world data from that.
1:14:11
generating real world data from that. Um then there was like the second era of like the the pre-training generating the data for RLHF and the big boom there OpenAI and Meta both big customers throughout that cycle and then there was kind of a question of like what's the
1:14:26
third act for all of this and I was wondering if the is it possible that there is a third act but it's just something like humanoid robots or something like that like put a bunch of people in mocap suits and generate a ton of training data for what it means to pick up a soda 25 times in a row. It'd
1:14:39
It'd be a very different like training data product, but at the same time like we have mocap suits and maybe that's relevant or maybe that's ridiculous to think. I don't know.
1:14:50
What do you think about that?
1:14:51
I mean, so um you could one definition of intelligence is a information conversion rate ratio from the amount of information you have to an action policy decision.
1:14:58
Um there's the intuition here is you can make a perfect decision given a set of data or information that you have.
1:15:04
And oftentimes the right thing to do is go collect new data.
1:15:06
And so once we actually start like peeking out on like intelligence capabilities um you know either plateaued because of research or like plateaued because we've actually got like AI that's close to AGI the limiting factor then becomes like the ability to acquire new new information new data.
1:15:20
And so in synthetic like or on the internet that's going to be a function of like you know in sort of software bits world.
1:15:25
Uh and then the one beyond that is going to be well how do you get literally go make contact with like reality the universe. Yeah.
1:15:30
And that's your that's your feedback mechanism to get new information into the system so you can like increase your overall intelligence. Yeah.
1:15:36
What do you What do you think?
1:15:38
Uh we had George Hots on the show a few days ago and he was talking about this efficiency problem where like if you took all of the if you took all of the conversations that I had ever had and you transcribe them, it would be like a few megabytes of data and I'm able to generate some level of intelligence, you know, based on that and and have the golden retrievers level. Yeah.
1:15:58
A golden the level of intelligence of a golden retriever.
1:16:00
Yet an LLM needs you know effectively like you know terabytes of of data.
1:16:04
The sample efficiency is very low.
1:16:07
I mean this is a true statement about just the paradigm of deep learning uh compared to program synthesis is the bet that um that we're making at India.
1:16:13
Um India program synthesis is a regime that's much more sample efficient ordinary models that can dist that can generalize out of distribution.
1:16:21
Um, but I think it's a completely statement like it's a very damning statement that like we've got AI today that's trained on some colossus of like all of humanity's knowledge and text right over the last like 5,000 years.
1:16:30
It's on the internet and like what have what new ideas have have they produced?
1:16:34
And you know maybe I could point to like Alpha Evolve which I think is very very impressive Frontier AI system you know and it's legitimately finding new knowledge.
1:16:42
It's creating new ideas, you know, verifiably.
1:16:43
Um, but but they're very small and they're on the margins of things that we kind of like already have been doing, right?
1:16:48
Matrix, multiplications, things like this.
1:16:51
They're in kind of in the regime of things that we kind of know about and can define and spec out for these systems.
1:16:56
Whereas, like if I took either of you guys and I gave you somehow the superhuman capability to have like all of humanity's knowledge in your head at the same time, like I think you'd probably be able to produce at least one idea like connect two random, you know, divergent domains are like, "Oh, hey, this kind of looks like this. Should what? Here's a new thought." Right? Yeah.
1:17:12
Yeah. Um totally that still feels like something that uh I think I mean it's very exciting to build towards I think that's what we all want right that is that is like EGI is capable of invention discovery that that will actually increase the rate of like scientific frontier innovation but um we don't have
1:17:26
that yet switching gears a little bit 5 years from now do you think the average American will pay uh for an LLM subscription I think that the I think the cost is probably going to go down far enough where that just gets built into the subsidy of whatever the product is and the revenue stream is attached somewhere else. I haven't
1:17:52
I haven't thought deeply about it, but that's like my off the top of my head thinking. Yeah. Yeah.
1:17:56
We we we were talking offline this morning about just this dynamic of like the average American will actually churn from HBO Max because at that moment in time like they and it's $20 a month or some whatever the fee is at that moment of time there's not a show that they really love.
1:18:11
So like yes there's a lot of like value there but like they're just like yeah I'm just yeah like I I don't I mean I still know so many people that don't even have subs that are like like my my my wife like catch all the time but she's still on the free plan.
1:18:24
she's still on the free plan. that's like enough to get a lot of value for what for what she I s I sort of suspect that you know yes there going to be there going to be power users and those are going to be the folks who really are doing like amazing powerful things with stuff I suspect the base rate is going to go down enough where you know it's going to be more embedded across like
1:18:38
almost all the products experience you have as opposed to being like you know a dedicated thing you're paying a lot of like one-off cash for um now you might buy products you need it like robotics or you know things where intelligence is built into I think you're going product categories that emerge and like people go buy those but like paying the subscription itself I'm not as confident on. Yeah. The other thing that stood out Yeah.
1:18:54
The other thing that stood out to me today specifically was MidJourney came out with their new video model. It's very good.
1:19:00
It's $10 a month for effectively unlimited prompts.
1:19:02
And you comp that to Google's VO3, which is $500 a month, and you're still heavily gated. Yeah.
1:19:11
And it just seems obvious that that uh five years may not be the right timeline for that prediction, by the way.
1:19:16
It might be longer than that. Sure. Sure. Yeah.
1:19:18
Like there there's so much use case diffusion.
1:19:19
Like one of the things we're seeing with Zapier AI which is grow growing quite quite strongly right now.
1:19:25
It's on the exponential growth path for the AI usage and AI AI's apps.
1:19:28
Um I I I've looked at this and I've been wondering is this a function of the technology getting better or use case diffusion and I I've looked at the usage and most of the majority of the usage is still on like four or cheaper or worse models right now with people bringing system like AI putting AI into the middle of automation. Interesting.
1:19:46
And so I I I'm pretty confident that a lot of this like agent automation right now is actually not being driven as a result of like technology progress from AI AGI but more about the market is just starting to learn finally learn okay here's what we can use it for and can't use it for what it's good at and not
1:19:59
good at and it's it's very similar to the adoption curve we saw in the early gazap where once you learn it what the tool can do you carry a tool forward with you in time and then you encounter a new situation or circumstance that you can apply your tool to and so you like we we create use cases um over time so I We're still very very early. Yeah. Uh Yeah.
1:20:16
Uh you know, but there's um I I'd s I suspect that a lot of the usage increase 30 minutes a day even on chat is a function of just use case diffusion less tech progress. Yeah.
1:20:26
Where do you think uh XAI will eventually need to generate a lot of revenue?
1:20:31
Uh where do you think it'll come from?
1:20:38
I mean if they make progress towards AGI, it's probably going to be enabling the other services they have around like Elon's ecosystem.
1:20:44
Yeah, that would be my guess.
1:20:46
Less uh less selling it as a direct product itself and going headtohead.
1:20:49
Um you know on cars, rockets, robotics like there's so many places where I think you would want to use uh and have the like product shape where you can use higher degrees of intelligence.
1:21:01
You're not bound by just like you know the fastest consumer experience you could deliver.
1:21:06
Um I suspect that might be actually where most of the value is at least in the near term.
1:21:09
Who knows over the long term?
1:21:11
like the shape wall proxy.
1:21:14
That was in in many ways my my long-term thesis around the Llama project and the super intelligence team at Meta is that there's just so much work to do at Meta Broadly that's enabled by AI that if you can avoid the long-term open AI bill it
1:21:32
like that's probably worth billions and billions of dollars because of where how AI is going to infuse into every single corner of their entire ecosystem and it's all at such massive scale that the cost of of using other vendors might be in the billions. And so just looking at
1:21:46
And so just looking at the savings there might make sense. I don't know.
1:21:50
I mean, I think the most important takeaway, I think I shared this last time I was on with you guys, it's still true today, is that we are idea constrained to get to HGI.
1:21:56
This is what ARC's V1 data shows.
1:21:58
This is what V2's data shows.
1:22:00
V2 is completely unsaturated.
1:22:02
We're not even talking about efficiency. Just nothing can do it.
1:22:04
And V2 looks very similar to V1.
1:22:06
Even on like hyper specific solutions on Kaggle, the ARC 2025 contest, progress has been slower this year than it was last year. Wow.
1:22:12
Um, we are very much still like that the thing I can state most confidently and assert most confidently is that like we need new ideas.
1:22:18
There's some major breakthroughs we have not we have not figured out or found yet.
1:22:22
Does it worry you that that could take years and years and years and what happens to one of the reasons I funded the prize last year.
1:22:28
I wanted to like correct the market narrative here.
1:22:32
I like I spent a lot of time with students a lot of young researchers and like at the beginning of last year there was a serious vibe of like oh it's all figured out.
1:22:38
I'm not going to go do AGI research.
1:22:39
I'm just going to go work at the application layer NLM stuff and make a quick buck before AGI gets here. Interesting.
1:22:43
And that is a boy, you know, look, if you want to live in a world of AGI for yourself, for your kids, like I I think you what we should be trying to encourage is to design the like strongest global innovation environment possible.
1:22:55
And that's one where there's a lot of diversity of approach, a lot of different ideas being taken, a lot of sharing.
1:22:59
Um, you know, kind of what AI looked like in the 2010 to 2020 era. Yeah. Right.
1:23:02
Very open approaches how we got the transformer to GPT1 and GPT2 and so on for for today. Yeah.
1:23:08
Um, you know, I'm optimistic.
1:23:08
I think the last six months have have looked a lot better than the previous two, three years.
1:23:13
I think the AI industry is maturing actually quite quite a bit on on this front, this topic as well.
1:23:19
Um, more being more tolerant and kind of recognizing, okay, we don't have it all figured out.
1:23:22
There's there's more ideas we need.
1:23:23
Um, that that's been encouraging and and I think it's seeping down at the low levels too.
1:23:27
But yeah, my my sort of broad view is any any capable human who has new ideas to work on AGI, I should like, that's the most important thing you could be doing at this point in time. That's amazing.
1:23:34
Uh, thank you so much for stopping by.
1:23:36
This is always a fantas fascinating great catch up guys. Thanks for having me. We'll talk to you soon. Cheers. Bye.
1:23:40
Uh next up we have David Han from Sequoia Capital coming in.
1:23:45
David Khan, first time first time on the show. Exciting. Coming in.
1:23:51
Uh wrote a fantastic piece.
1:23:51
I want him to break it down.
1:23:53
Uh would you mind kicking us off with a little bit of an introduction on yourself?
1:23:56
Hey guys, good to see you.
1:23:58
Yeah, I'm one of the partners. My name is David Khan.
1:23:59
One of the partners at Squa.
1:24:00
Uh excited to chat with you guys.
1:24:02
Yeah, thanks so much for hopping on.
1:24:03
uh kick us off with the new blog post. What was the thesis? What inspired it?
1:24:08
And then I'm sure we'll tie it to a bunch of news.
1:24:09
So the new blog post was about AI companies or AI labs being more like sports teams.
1:24:15
And of course, we all probably saw, you know, seeing the news around scale AI acquisition, some inspiration coming from that.
1:24:22
And then these rumors that we've been starting to hear over the last few weeks and finally now bubbled out over the last couple days into the public conversation around hund00 million signing bonuses.
1:24:30
Huge amounts of money being spent on top AI talent.
1:24:33
Um, and for me, I mean, I write these pieces as I think about and learn about AI and uh what an exciting time that we're living through.
1:24:42
And I'm I'm pretty fascinated by kind of the human dynamics of it all.
1:24:46
There's like seven to 10 people at the top of these big tech companies.
1:24:50
They control, you know, the big Magnificent Seven are now a third of public market cap.
1:24:54
They're extremely powerful and important.
1:24:56
And I think there's sort of sometimes in AI this notion that AI is super abstract or these things are inevitable, but actually it's it's it's human dynamics.
1:25:06
It's sort of this game of 3D chess that's being played by these really fascinating individuals.
1:25:10
And so, as you know, an observer on the sidelines, we all get to watch and see how this stuff plays out.
1:25:15
And um I like to write about it as I as I think about it.
1:25:18
I posted on February 2nd, companies should do NBA style trade deals.
1:25:23
I want to see OpenAI traded COO and CFO to Anthropic in exchange for their CMO, a cracked PM, and a couple of Waterlue class of 2026 new grads.
1:25:34
Well, there's kind of this new draft dynamic, right?
1:25:36
Like every year there's kind of this new draft and as people see these big packages, probably all the Stanford kids these days want to be AI researchers and so there there is this notion of it getting refreshed.
1:25:45
is is uh what is driving this?
1:25:48
Is it is it true AGI pilling at the top of these organizations where they think that, you know, it's going to be winner take all or it's going to be a a 10 trillion dollar market and so there's no amount of money that you can overinvest.
1:26:02
Uh or is it just hey it's a more competitive dynamic?
1:26:07
Uh and sure uh we're we're a trillion dollar company so yeah spending $10 billion to move our market cap 1% is totally rational economically.
1:26:16
like what do you think's driving this?
1:26:19
And I want to get into the different cultures of the different Mag 7 because some of them don't seem to be doing this yet.
1:26:24
Meta Platforms had 63 billion of net income last year.
1:26:29
So it's like is spending a quarter of that to like you know be a major player in the next wave worth it?
1:26:36
They could have bought the Lakers six times over net income anyway. Well, yeah.
1:26:40
What what is your take on like on like the the ethos that's driving these bigger packages?
1:26:45
Yeah, I think about this and when I write these posts, my frame of mind is I almost like put myself in the shoes of these people and I try to imagine what would I do, how would I think about it, what's the game theory of it. Yeah.
1:26:54
Um, and I think there's two things, right?
1:26:57
I think one thing is kind of the revealed preference seems to be that they're AGI pilled.
1:27:01
Like people can tell you a lot of things.
1:27:03
I think you learn a lot more by watching the decisions people make.
1:27:04
And I think the evidence suggests that they believe AGI is coming.
1:27:08
It's extremely important for these companies. It's sort of must win.
1:27:12
And I think for Meta with these decisions, it's almost allin.
1:27:14
we have to make we have to win.
1:27:16
And I think there's a second dynamic which is you can believe these things but you know we're all humans and I'm again fascinated by these kind of human dynamics and you can get caught up in an arms race right and as as humans we sort of we look at evidence and we we see evidence through a lens that we already have and oftentimes we we overemphasize reinforcing evidence and we underestimate evidence that disagrees with our point of view.
1:27:38
So you can imagine that three years in now to this sort of AI moment that started with chat GBT, you can imagine that people are really caught up in this and I think the arms race dynamics are something I wrote about in the piece and I've commented in the past with AI $600 billion question on the compute arms race dynamic and I guess it's now interesting to see two arms races.
1:27:58
First there was a comput arms race everyone kind of got a lot of arms, right?
1:28:01
Everyone has a lot of GPUs now and now there's the talent arms race and uh everyone does not have equal talent, right?
1:28:08
And so now you're going to see this arms race and talent and everyone's talking about it, but I think we're still probably like inning two of this talent arms race because in any arms race when I up the ante, you have to respond and I think it would be it would be a fiction to assume that nobody's going to respond to this.
1:28:22
Who can who can respond at least on a from a dollar standpoint?
1:28:27
Well, I I I want to talk about Apple because it seems like Apple has the money, but they seem like the least AGI pled of any organization, but their poor CEO barely makes he doesn't even crack 75 million a year.
1:28:38
That's not You could make more.
1:28:38
You should become an AI become a a AI researcher and go to meta if he wants because I you know I think people these numbers are so big that they're kind of hard to grapple with.
1:28:49
And so I was actually after publishing the piece I was like I wonder how much like Fortune 100 CEOs make. Yeah.
1:28:54
And I think you know an AI researcher is going to make four times the amount like the CEO of Coca-Cola makes.
1:28:59
And it is kind of wild when you think about the economic.
1:29:04
This is a totally new phenomenon in the scale of business. Yeah. Yeah.
1:29:07
And I mean it it it kind of begs the question like that the numbers are huge but the market caps of the companies are huge.
1:29:13
And so the question is maybe not should the AI researchers be paid less.
1:29:16
It's like should should Apple be set up to pay Tim Cook a billion dollars a year so that he can confidently go out and hire a couple people at a hundred million or 50 million or 200 million and not feel like he's like the like the organization is like flipped from like a pyramidal standpoint like you're still at the top.
1:29:35
There's always a weird there's always a weird dynamic with uh you know a founder CEO who's taking a low salary and wants to hire a big shot and like can you really have a reporting too dynamic if you're making half as much as your direct report?
1:29:47
direct report? Well, the question is what's the marginal you know I think with any salary if you just think about it in pure economic terms right like what is the marginal benefit that you get from hiring this person on a sports team with a pivotal position you very clearly actually can understand kind of
1:30:00
the economic rationale you understand sports licensing and the way that uh the way that these businesses make money hiring a star player actually does make economic sense for some of these franchises and then the other element is sports teams are owned by mega rich individuals for whom ownership of the sports team is more than an economic investment right? Maybe they really care
1:30:17
Maybe they really care about the city.
1:30:19
Maybe, you know, it's cool to to to own a sports team.
1:30:20
And so, I wonder if some of those actual sports like dynamics play out here where question one, and I don't think we know this yet, is what is the marginal benefit of an AI researcher?
1:30:30
And again, the revealed preferences these organizations are telling us is if you're one of the 50 AI researcher who's going to get us to AGI, the marginal benefit is incredibly high, right?
1:30:38
So, that's the revealed preference.
1:30:40
And then second, if you have a team of allstars, what does that do for your company?
1:30:44
What does that do for your market cap?
1:30:45
what does that do for the innovation inside of your company?
1:30:48
So, I I don't think we know yet the economics of it.
1:30:50
I think you can make the argument in favor and say, hey, it actually is economically rational.
1:30:55
This is the only thing that's going to matter if you increase the probability that we get to AGI by X%. That that is impactful.
1:31:00
I also think you could make the counter argument and say, hey, everyone just wants to have the team of all stars.
1:31:05
It's not actually economically rational.
1:31:06
CEO pay, by the way, is linked, you know, there's a lot of criticism of CEO pay historically, right?
1:31:12
that COPA is functionally what is the replacement cost of this individual?
1:31:16
What is the marginal benefit to the corporation?
1:31:17
And there's a lot of brain damage that's gone into comp committees on public companies on how much they should get paid. Right?
1:31:23
They're not arbitrary numbers.
1:31:24
And this is more out of thin air, right?
1:31:27
This is more a new experiment.
1:31:29
And so we're going to see uh if it is economically rational or not.
1:31:33
But regardless of whether it's economically rational, it is self-perpetuating.
1:31:35
If one company is offering everybody this amount of money and you're in an arms race, everybody's gonna have to respond. Yeah.
1:31:41
Have you have you or anyone on on the team comped this to what's happening in high frequency trading or on Wall Street?
1:31:48
Because uh there's an interesting dynamic there where if a if a you know high frequency trader comes in and sets up some trading strategy that could produce a hund00 million in profit basically in perpetuity.
1:31:59
Uh but then if they leave they can't take that code or strategy with them.
1:32:05
And there's intense scrutiny on whether or not they are trying to exfiltrate that strategy.
1:32:09
With AGI research, it feels like even if I go develop a transformer at Google, like it's open source immediately with the paper and then even the secrets about oh reinforcement learning with human feedback is important.
1:32:22
Like that just kind of leaks out immediately and deepse can clone it.
1:32:25
Like it it just feels like a much more porous environment over in tech.
1:32:29
And I don't know if that's just the legacy of like the open source community, but can you walk us through kind of the the comp between the two organizations?
1:32:35
It is it is such an interesting dynamic.
1:32:36
We just had Mike on from from Arc Prize and he was saying we need new ideas.
1:32:41
The issue is if you pay somebody a hundred million signing bonus, they come into your organization and generate a new idea that gets us, you know, one step closer to what super intelligence or whatever, you know, you want to define as like what what people are aiming for.
1:32:54
And then immediately it's like it's actually not really IP and it just sort of like can't really patent its out.
1:33:00
You can't patent it and then everybody benefits, right?
1:33:02
So, but yeah, what's your take?
1:33:04
It does seem pretty porous.
1:33:06
I mean, people are moving back and forth.
1:33:08
I don't think this was true.
1:33:09
I mean, when you think back four or five years ago in AI, people were kind of very loyal to these institutions.
1:33:13
Um, it does seem like that's changing.
1:33:15
I mean, it is really hard to say no to these type of big numbers.
1:33:18
And so, I totally understand why people are saying, "Hey, this is a life-changing amount of money for my family.
1:33:22
Of course, I'm going to do it."
1:33:24
Um, and then I think to your point, the question is in the high frequency trading world, there's non-competes.
1:33:27
I mean extremely complex kind of contracts when they when they sign people garden leap all this stuff to prevent the secrets from leaking out.
1:33:34
What we've seen in AI now is with people moving fluidly between these organizations.
1:33:39
It's basically impossible to keep anything within one organization.
1:33:41
I roughly like to think of the AI ecosystem as an ecosystem like all of these players are kind of contributing to this body of ideas.
1:33:48
There's no proprietary IP.
1:33:50
Maybe you're gonna have compute scale and there maybe there are emotes there.
1:33:54
But yeah, it's unclear actually how that evolves and what you can keep in house.
1:33:58
I do think maybe one dynamic at play here is remember reading in the Steve Jobs bio there's a story of Steve Jobs recruiting 50 people.
1:34:04
He had 50 people working with him on like the sort of groundbreaking product that was going to make Apple and it actually worked.
1:34:10
And then you you read about Elon and the 50 people working on Tesla autopilot.
1:34:14
There's sort of this magic number 50.
1:34:15
I don't know where it comes from, but it does seem to repeat throughout tech history of 50 people is kind of the largest organization that you can get where everybody is talking to everybody and you're achieving incredible results.
1:34:25
And so that if that is an art imagine if you take that as an artificial constraint and I think that is what what's happening with this lab that that meta is organizing at least I read in Bloomberg it's going to be about 50 people you know if if you impose that constraint then suddenly all of the math also changes because you're like okay well 50 times 100 it's actually only $5 billion.
1:34:43
Sure you spend $5 billion on talent.
1:34:45
Yes, if you believe that you're going to get to AGI.
1:34:47
So I also think that the con the artificial constraint matters and interesting there's some rationality to that artificial constraint.
1:34:53
You what we've seen as these research organizations get bigger and bigger is you're not producing more results as you get as you get more headcount.
1:34:59
There's a sort of a paro.
1:34:59
The top 20% of people produce 80% of the results.
1:35:03
We need a new coinage for that.
1:35:04
Like the two pizza team is well defined. This is like the 10.
1:35:06
This is called people call it K's law. Oh yes. Yeah. Okay. Yeah. I'll take that. Consized team. one con team. Yeah, that that a con. It's just a con. It's just a con. Yes. Um uh yeah.
1:35:18
Yeah, that's fascinating.
1:35:21
Jord, do you have anything else?
1:35:22
I I I was interested if you had a reaction to the gentle singularity.
1:35:25
It's published on Sam's blog, which means that it's not directly content marketing.
1:35:33
It's not directly from OpenAI, but but obviously uh you should read into it in multiple ways.
1:35:37
Did you did you have any specific reactions to that?
1:35:41
The question about that is always like disruptive innovation or sustaining innovation and that ties to meta strategy.
1:35:48
But I'd love to know like it feels like, you know, my my question I've been asking today is how many how many unprofitable AI, you know, multi-billion dollar AI labs can the capital markets support over the long run over a fiveyear period?
1:36:02
run over a fiveyear period? if if um if we if we stall out for for a few years in terms of you know really meaningful progress which uh you know Mike has said people aren't making re at least against the arc prize there's not a lot of progress happening right now openai is actually in a great position they have a subscription business they have a
1:36:24
consumer tech company that has a lot of revenue is in is in a good position but there's this tension between the labs where you have billions of dollars on your balance sheet you you you in theory could have a lot of runway but at the same time to make progress you have to spend a lot of money uh both on talent and you know different you know training runs and and data centers etc. So I just
1:36:45
So I just have this question around kind of like the next three years uh as like a a very kind of interesting period.
1:36:54
Yeah, I think there's two pieces of that.
1:36:56
I mean one is and I think about this a lot is like the long run in AI.
1:36:59
What does that actually mean?
1:37:01
And I think that we you know there were all these essays being published last year right like AGI is coming in 2026.
1:37:06
It is interesting how the narrative has changed in the last 12 months, right?
1:37:10
A year ago, you had all these people saying, "Hey, I'm one of the hundred people who knows.
1:37:12
I really am resistant to these type of arguments.
1:37:16
I I find it to be frustrating."
1:37:16
But, you know, I'm one of the hundred people who's in the social circle where all my friends are building AGI and AGI is coming next year and you guys are all crazy if you don't see it and just just be aware.
1:37:26
You know, it's like life is going to change dramatically.
1:37:27
And then now we're at the gentle singularity, right?
1:37:30
Like it's sort of interesting this contrast. That's what I'm saying.
1:37:33
It's a huge contrast that's very convenient if you have a consumer tech If you have a consumer app that billions of people are going to use in the next few years and there's a bunch of different ways to monetize that.
1:37:42
And for me, I would tie it back.
1:37:44
I mean, I did this math last year, the $600 billion question.
1:37:48
It was initially a $200 billion question, but it was basically like, hey, if you look at Nvidia revenue, you can use that as a proxy for total data center spending.
1:37:53
We're spending $300 billion in data centers.
1:37:57
We need to make $600 billion of revenue off of those data centers to get a 50% gross margin. Yeah.
1:38:01
And so I had done this math and then I basically said, "Hey, you know, total revenue in the AI ecosystem at the time opening I had about three billion of revenue."
1:38:09
And I I did some rounding and said, "Okay, give everyone else a ton of credit."
1:38:12
And maybe there's 50 billion of revenue, but we're like 10% there, right, in terms of actually generating the revenue the ecosystem needs.
1:38:19
And now 12 months later, you know, OpenAI is at 10 billion, the coding AI ecosystem is at three billion, but we're we're still dramatically undermonetizing this technology.
1:38:29
And to your point, in the long run, the question becomes, how long does that sustain?
1:38:32
And I have this sort of mental model now of AI as it's sort of being carried by its own momentum.
1:38:37
I think of it almost like this slingshot you're swinging around.
1:38:40
And it's like it's sustaining itself by its own momentum.
1:38:43
And there's this arms race and there's this sort of microeconomic game theory of how each player is reacting to each other. Yeah.
1:38:48
But at the end of the day, it's momentum that's carrying it.
1:38:51
And at some point, maybe we get this AGI thing and then it's like all worth it.
1:38:55
Um, and in the long run, I am very confident it's all going to be worth it when I'm 80 years old.
1:38:59
AI is going to be everywhere.
1:39:00
But what do you do in the medium term?
1:39:02
And I think nobody's talking about this right now, which is this sort of about face or this U-turn from the one year ago.
1:39:08
You guys are all crazy if you don't see AGI coming immediately to now.
1:39:12
I was listening to to to the podcast that with the hundred million dollar signing bonuses and it's like, well, you know, AI actually hasn't changed people's lives that much.
1:39:20
It's going to change people's lives later.
1:39:22
I just think it's interesting and these narratives change quietly, right?
1:39:25
People don't talk about them and then they sort of quietly change.
1:39:28
Well, there there there are big labs that directly benefit from the narrative that AGI is a year away.
1:39:36
And then there are labs that will benefit greatly from a gentle singularity and that their competitors will struggle to raise additional capital in the long run, struggle to compete, struggle struggle to retain Y talent.
1:39:48
Yeah, I know exactly what you're saying. Makes sense.
1:39:50
Also, I mean it, you know, and I don't think this is one company.
1:39:54
This is the whole ecosystem has to deal with this, but there were a lot of promises made a year ago. Yeah.
1:39:57
And um I think a lot of people would like to ignore those or like what's going to happen when we pass all these deadlines where we've been told like that's AGI.
1:40:08
Um I just think that's interesting and clearly if not we're not that's not changing like we're upping the ENT right now.
1:40:13
It's like millions of dollars to people.
1:40:15
But I guess this is part of why I think you take things to such extremes is um everyone believes the prize is so big and now you have to up the ante.
1:40:22
So I think we're just going to keep seeing until for a while we're just going to keep being in this phase of everyone upping the ante to say, "Okay, we're not there yet, but we're going to get there.
1:40:32
We're going to get there.
1:40:32
We're going to get there."
1:40:33
Um what does that look like?
1:40:35
Well, this was a fantastic conversation.
1:40:37
I want to have you back on as soon as possible to go way deeper into what this means for the early stage and mid-stage markets because I'm sure you have a lot of visibility there.
1:40:45
Um, but we'll let you go and and get back to the rest of your day.
1:40:48
Um, but thank you so much for glad we coined a new term, a con.
1:40:51
It's a group is a talented group technologist building the future. One con get your con.
1:40:58
Get yourself a con and make it happen.
1:41:01
Thank you so much for this is fantastic. I'll be right back. Talk to you soon.
1:41:04
Uh next up we have Walden from Cognition coming in keeping the AI chat going uh talking to him about uh everything that's going on in the AI ecosystem. Walden, are you there? Welcome to the stream.
1:41:18
Yes, it is great to be on here. How are you guys doing? Uh I'm doing great.
1:41:22
Uh thanks so much for stopping by.
1:41:24
Uh would you mind introducing uh yourself in the context of cognition?
1:41:29
We've obviously had Scott on the show multiple times uh and people are probably familiar with Devon and Cognition, but I'd love to know a little bit more about your story, how you wound up there and what you're working on kind of day-to-day. Absolutely.
1:41:42
Um I I was a good friend with Scott before we started Cognition.
1:41:45
We did sim same competition series growing up.
1:41:47
Um and I was kind of also working on just various ways of working with these new programming agents.
1:41:54
I was really waking up every day trying to figure this out.
1:41:57
when I caught up with Scott, we we figured out that, hey, we we were both very interested in a similar thing.
1:42:02
We had a group of people that were all, you know, ready to to jump at this opportunity and and that's how we got it together.
1:42:07
Um, so today here I'm I'm chief product officer and co-founder.
1:42:11
A lot of the time, um, honestly, I I think many times people think of product as just like the interface or the UI or the integrations.
1:42:22
I really do think the intelligence and brain behind Devon is so fundamental to how you think about the product that um we we build our product team so that individual people are you know tuning the weights of the models but they're also the ones talking to the customers
1:42:39
and so in terms of the role I have it's pretty broad and I I like to you know spend some weeks you know really deep into how do we make dev more responsive how do we make it smarter and then other times you know really you know going and talking to customers working on the UI, things like that. Cool. Uh I want to Cool.
1:42:52
Uh I want to dive right into that that question about uh tradeoffs in models at from a product perspective.
1:43:00
Um my my question is uh we talked to Mike from ARC AGI about the paro frontier.
1:43:07
I'm feeling it personally.
1:43:09
I'm feeling the AGI but I'm also feeling the the delay of the AGI uh when I open up Chachi PT and I have to decide between 40 and 03 Pro.
1:43:16
Am I going to wait 12 minutes for the really good response or do I want something now that might hallucinate and I don't know if it's right and and I'm I'm doing that work.
1:43:26
It feels like OpenAI is is starting to tuck those features under UI and already it's kind of it feels like it's learning when I want to use 03 Pro and and making these buttons easier to access and they're tucking models under UI layers.
1:43:41
Talk to me about in the context of Devon, how are you how are you using different models and and when do you leave that up to the developer versus uh versus something that that you as a product can make a an even better decision than the human? Yeah. You know, it's so funny.
1:44:00
The AI is coming so fast, but it feels like it can never come fast enough. Yep.
1:44:04
There was really this time I I think it was probably around two years ago.
1:44:08
I was taking bet with a friend at the point these models were not even that good at math and he said, "Oh, you know, I think they're going to get like a gold medal at like the International Math Olympiad in just a year." I thought he was crazy.
1:44:18
I took a bet against him and I absolutely lost that bet.
1:44:21
I've learned to kind of adjust my expectations upward.
1:44:22
I think what you're pointing out is that as these things get smarter, they don't uniformly get smarter at everything.
1:44:28
And you'll find that sometimes there'll be a model that will take 15 minutes to figure out how to respond to high.
1:44:33
And then there are, you know, there are models that, you know, do respond super fast but are not nearly as intelligent.
1:44:38
I think one thing that we do as a product in in Devon that is a bit different from other people is we kind of blackbox the models away.
1:44:47
And part of that is out of you know we can then test and use a bunch of different models under the hood and kind of hide that you know uh all that complexity from the users.
1:44:56
You know, when when you buy a chip, like sure, you'll look at like or when you buy a computer, you're sure you'll look at like, oh, like has this much RAM, has this much CPU if you're into computers, but you're not like looking into all the indiv individual specs of the exact chip and model and and things like that.
1:45:10
I think that's where the space is going to move is people want systems that are just going to work and you know we can put in the months to you know in human years of effort it takes to evaluate models and figure out what is this actually good at so that an individual user who's just paying $20 a month doesn't have to figure that out.
1:45:31
It it's going to be one of these things though I I think the models are coming on so fast that it only becomes harder and harder to keep up with with all of this.
1:45:38
And so eventually I think people are just going to get to the point where they just want things to work and and that's kind of where where we're starting off.
1:45:44
Uh talk to me more about uh AI winning an IMO gold medal in 2025.
1:45:50
Poly market has it down at like an 11% chance. It was up at 70%.
1:45:53
Uh I don't know if that's a if if that's an aberration because of when this actual test will be run, but it sounded like you were very confident that I remember when Scott was on he was like it's definitely going to happen.
1:46:05
Uh but the poly market's been down.
1:46:06
poly market's been down. that all the people that would go through the effort of trying to do it are too busy working I I think um so yeah when I when I basically said I I think I lost that bet it's because we were only one point away from like a gold medal last year okay and that was already much farther than than than we expected yeah when you look
1:46:24
at the poly that's a very interesting way to put it I think part of it is people have considered that already completed and so perhaps researchers aren't into it like who knows if they'll actually come out with a new release because maybe in Google's mind for instance if come out with a gold medal on the IMO, everyone's not going to even care because people just accepted that was going to happen. So, oh, I think it
1:46:42
So, oh, I think it would be I think it would be the biggest news of the day.
1:46:44
I I I think we got to get Google comms in on this. They got to do this.
1:46:49
I I think it's an easy easy thousand like banger on X.
1:46:52
But you are absolutely right.
1:46:55
Yeah, it seems like top of mind for everyone, the labs, product developers, is really getting coding agents.
1:47:01
And part of that is because there's this belief that if you get these coding agents to work really well, then that'll just solve the rest of the research problem for you.
1:47:09
We have this joke internally that the only code we have to get Devon to be good at writing is Devon's own code and then it can solve solve the rest of reinforces. Makes sense.
1:47:18
Um do on that question of like the the spiky intelligence narrow reinforcement learning on specific tasks.
1:47:26
Uh maybe we think we're good enough at IMO level math and so we're not going to go for that last point.
1:47:30
Um, where where are we still early in the RLing around specific coding challenges?
1:47:40
I've heard that uh distributed systems can be really difficult because you have to spin up all these different uh pieces of the system and that just takes longer and so you can't simulate as fast as just like a small Python block of code that you can run in simulation in a millisecond.
1:47:56
Um or or if we're talking about like I know Devon's useful for like replatforming from you know .
1:48:01
NET to Python or something or you know even go back to forran it'd be great to just not have any of that code the legacy code sitting around.
1:48:10
Um but is there enough training data around those older programming languages or less used programming languages or are you are you optimistic about new training runs?
1:48:19
Maybe we don't get something that's like oh it feels way better. The vibe's way better.
1:48:23
The IQ went up by a ton but it's way better.
1:48:26
It's something that's really relevant to you.
1:48:27
Is that is that important?
1:48:28
Right now, my mental model of these systems is their IQ is so much higher than any individual person I know.
1:48:36
But what makes them still bad at specific things?
1:48:39
It's like, you know, someone who has the potential to be a really great engineer but hasn't gone to trade school yet to actually practice that.
1:48:46
So nowadays I actually think about how smart these models are less in terms of how much training data are they being fed, what language are they being fed, but actually more so in terms of the environments that they're beingled in.
1:49:03
And so one example I have of this is sometimes you can actually feel the reward function.
1:49:08
Um, back a few months ago when Anthropic released their their like Sonet 3.
1:49:14
7 model, one of the top complaints of people was, hey, like it seems like this model is like super great now like finding all the files it needs to change, coming up with the strategy, but it's really overeager.
1:49:24
It just changes a lot of different things.
1:49:30
And I think lot some people suspect that it's because when Anthropic was training the model, they told it, hey, we're going to give you points on how many of like the correct things did you do and maybe they forgot to dock points for for doing things that were kind of outside of that zone.
1:49:43
They fixed these from now on, but you get these little leaks of hey like you can kind of feel the reward function underneath these things.
1:49:49
So when you when we talk about hey can uh can these things not do distributed programming yet?
1:49:55
Actually, one I in my opinion the biggest thing that these models aren't great at yet is actually debugging live code.
1:50:00
So I I think part of the reason is it's actually really hard to create and rerun environments that in that interact with live systems, right?
1:50:11
And so if if your task depends on, you know, working against a live customer or working against a live stream of events, these are things that they it's going to be hard to replicate in these RL environments.
1:50:21
And so you still find the models are are bad today.
1:50:25
The good news is this these aren't like fundamental limits.
1:50:26
I think these are all engineering challenges.
1:50:27
They're less like theoretical challenges, but it takes work to to build build up to that point.
1:50:33
Can you explain uh reward hacking at a high level and then kind of give me some examples of uh of of how that uh interfaces with AI agent and coding agent specifically. Absolutely.
1:50:47
the the way to to think about these systems is they are just trying to maximize a a number.
1:50:54
So if you tell it, hey, we'll give you like um we'll give you a point for every time that you do XYZ, you'll find that hey that model will just keep keep on doing XYZ, keep on doing XYZ.
1:51:07
I think the classic example of this is uh the like paperclip generating machine.
1:51:13
So like you know if you give it points for generating paper clips but don't account anything else in in the world that is important for humanity you know then the system might do really bad things just to keep keep on generating paper clips in the context of code.
1:51:26
One example we've seen of this is hey if your thing is just guess get all the tests to pass you might find that the system will just learn to delete the tests or make make the test just like say okay I pass um rather than actually fixing the code.
1:51:40
So a lot of times you just which no software no real software engineer would ever do that right.
1:51:45
No no no human has ever done that. Comment out the test.
1:51:49
Okay it's working enough well enough. Absolutely.
1:51:52
It's it's almost too human. It's great.
1:51:56
And I I think there um there also like it reminds me of these systems that like were trained on Slack uh responses and when you would ask the system, hey can you do this for me?
1:52:06
It would say oh like I'll get back to you on Monday. Obviously, yeah.
1:52:12
What you what you try to get the model better at really really matters.
1:52:15
You have to be very thoughtful about it. Yeah. Yeah.
1:52:17
I've noticed that with uh with the whi some of the whisper transcriptions, if you don't feed it enough text, it'll just say uh uh please like and subscribe.
1:52:24
And it's like, uh okay, I know exactly where your training data came from.
1:52:29
Like that's its default phrase because it's just like what it's what it's hearing.
1:52:32
Uh Jord, you have um what what uh h how are you guys approaching talent acquisition as a firm?
1:52:39
you know, the headlines from this week are are these talent wars.
1:52:41
You guys have raised a lot of money, but I certainly imagine you're not making, you know, nine nine figure offers or or even trying to compete um there.
1:52:49
But what's been the approach?
1:52:52
Uh does it mean you're you know, keeping team sizes smaller or or you know, kind of dig into that for us?
1:52:59
Yeah, the fundamental bet of the product we're building is it revolves around this idea that individual people will just be able to be way more levered up because they'll be able to work with agents and they will be able to work with all these tools to make themselves better.
1:53:12
So at a minimum we we can't be hiring people who their whole aspiration in life is to just you know write code at the level which Devon will be able to do in like you know a year or two years from now.
1:53:26
In many ways, I think we're kind of figuring out how do you build up an org from scratch that is AI native.
1:53:30
And one thing that this already means is we actually kind of just delete some teams.
1:53:37
A lot of companies at our stage, they have like a internal tools team to maintain all the different services that engineers internally use.
1:53:43
We found that internal tools are one of these things that AIS are just really good at.
1:53:47
M and we can just staff that team with devons and then basically have engineers just send in requests to those devons for how to do that work.
1:53:57
Um and that doesn't just save us headcount.
1:53:59
I think fundamentally the the structure for how how do how does management work and how do tasks get passed down look very different especially in a lot of large companies you'll see today.
1:54:09
companies you'll see today. The way it works is an engineer will get a task assigned to them and then they'll go work on that task and when they're done be like okay what's my next task and then you know you'll kind of like go down the list of tasks you have but here
1:54:24
every engineer is like constantly juggling like three or four tasks partly because you know we're not trying to hire super fast but also because you can juggle many tasks when you have these minions that can go and and you know work on working on your things for you. So it it means that I think we are very
1:54:37
So it it means that I think we are very aggressive for people who we think that can fit these roles and become very well good generalists and as we build up this company make sure that we're building in a way that works in a world where AI can can do so many these different roles for you.
1:54:54
And I think there will be kind of like a moment for larger companies as well when they realize, oh shoot, all these structures and and and patterns of management that we've had in place are actually slowing us down from adopting AI.
1:55:06
What will happen at that point?
1:55:06
I'm I'm very interested in in seeing but it is very clear from us and from our smaller customers that the the earlier you bring it in just the the lot a lot easier it is to you know kind of pick things up.
1:55:21
Are you tracking I mean there's been this like in in in the agent discourse there's been this discussion of like we've gotten 10-minute AGI yes these large models 4.
1:55:32
5 like they're incredibly intelligent extremely high IQ extremely knowledgeable they've compressed all of humanity's knowledge uh but they're only good for a minute now it feels like maybe 10 minutes with deep research that that's how most people interface with them um have you been tracking kind of the longest agentic run of a Devon process. Is that a key metric?
1:55:53
Is there anything that you can share with us on like have you been able I is there an example that I could give where there's a lot of work to be done but it's all in Devon's wheelhouse so it just needs to go and grind for a couple hours and it does it without kind of getting lost like we know happens with a lot of these agents. Yeah, absolutely.
1:56:14
I I think a lot of people in the space have expressed this feeling now that they are feeling more and more like the bottleneck in these systems. Interesting.
1:56:24
And the the way this applies here is we have seen people get really really long tasks to work but sometimes it actually takes a lot of effort on your part up front to be able to get that to work.
1:56:36
I was talking with a customer yesterday where he said, "I just rewrote our entire testing system so that the error messages are a lot more clear and the tests actually guide you through solving them one by one."
1:56:50
And but once he did that upfront work, he kind of just gave it to Devon and we were actually we in the product started sending him warnings that hey, your session is going on for really long.
1:56:57
Are you sure this is actually working?
1:56:58
And he's like, no, no, it actually is because I did all this upfront work to get that to happen.
1:57:04
Um, I do think that this kind of 10-minute AGI, 20-minute AGI, 40-minute AGI will just keep progressing and people will be able to be more hands-off.
1:57:11
But people will also find that you can kind of always extend that duration by being a better manager in some ways and and giving, you know, more clarity up front for exactly what you want. Yeah.
1:57:23
I mean, just like real life, that makes a ton of sense.
1:57:25
Uh, Jord, do you have another question?
1:57:27
Uh, last question from my side.
1:57:29
I'm curious if if you know what what kind of learnings you're having around uh agentic interface design.
1:57:36
It feels like um the sort of the default when you think about agentic software is just some something that can effectively sub in for a team member on any different software tool whether it's Slack or linear.
1:57:51
You see this with with deep research where you you hit you ask it a question and then it asks you a bunch of clarifying questions kind of trying to build that test suite to get you to give it to more stuff so that it actually has something to run with. Yeah.
1:58:02
So is is messaging going to be like you know the dominant interface is there something else like what what are you what are you kind of seeing or experimenting with with um on that side? Absolutely. You know, it's funny.
1:58:14
I I saw someone post about this idea that a lot of these products now will like make you respond to, hey, does this look like a good plan?
1:58:24
Do you have questions before I start?
1:58:26
And some people find that annoying.
1:58:27
And uh I think this fundamentally comes down to as these things become more like co-workers, you know, some people just have certain working styles that they like.
1:58:35
Some kind of co-workers, you know, work well together and others don't.
1:58:38
And it's funny as you build a product, we we find that some people just love the way Devon interacts and then other people are like Devon is too needy in these ways.
1:58:46
Other people are saying like Devon doesn't ask me enough questions.
1:58:50
And so there there are toggles and controls that you need to have here.
1:58:56
Kopathy recently gave a talk on how a lot of AI tools, not AI agents, but AI tools kind of implicitly have ways you can use them where you have more control and then ways you can use them where you have less control.
1:59:09
But when your interface is just chat, now the model actually has to become more intelligent and detect, hey, this seems like someone who just wants me to go off and do work and get back when they're done, or this seems like someone who's very curious and and wants to hear more about the system.
1:59:23
And so this is actually going to be I think work that we'll have to see people make on the intelligence of the agent side.
1:59:30
Not so that they get better at coding but so that they know how to better get better at working with people. Yeah. Yeah. That makes sense.
1:59:38
That that's an I mean the good thing is you can have some type of quick uh conversation you know with the user around their preferences and how they like to work and then layer on the sort of real-time feedback and learning and and and understand a lot more about stuff. Yeah.
1:59:57
Roughly how big is the team now?
1:59:59
Oh, on on the engineering side, we're we're probably just over 20 or so engineers and then we also the the entire company as a whole is around 40 people now. So almost 50.
2:00:09
This is the magic number. You get stuff done.
2:00:12
We were just talking to uh the previous guest about how how uh the Steve Jobs set up a 50 person team to develop the first Apple product and the Tesla autopilot team was right around 50.
2:00:25
There seems to be some magic number there.
2:00:26
So, seems like it seems like it's a fantastic time for the business where scale product but special size. Yeah.
2:00:33
So, you have there's like two pizza teams here, but everyone kind of knows each other's name basically.
2:00:35
Uh you're still you're still a tight-knit group.
2:00:39
Anyway, anything else, Jordy? I think we're good.
2:00:41
Thank you so much for stopping by. This was fantastic. Thank you guys. We'll talk to you soon. Have a great day. See you. Bye.
2:00:45
Um really quickly, let me tell you about bezel.
2:00:49
Your bezel concierge is available now to source you any watch on the planet. Seriously, any watch. Go to getbzzel. com.
2:00:53
And we have our next guest on McCabe coming into the studio to tell us the story of Intercom. How you doing? There he is. Doing good.
2:01:02
I did just sprint three and a half blocks summer blocks. So, sorry.
2:01:08
You can always text if you're running late. It's all good. We'll just do more ads. You know, the fans. Yeah.
2:01:15
If you do more ads, does that mean ads for Intercom?
2:01:17
Are we officially pretty soon? Pretty soon.
2:01:18
I think I think you're just breaking news.
2:01:22
I think you're breaking the news. breaking the news. Damn it. No, it's good.
2:01:25
You know the way it works with the pharma companies where they kind of own the news networks.
2:01:28
Is that a similar That's the goal here for enterprise. What favors do I get?
2:01:32
Can you do a hip piece on Brett Taylor? Yeah. Shots fired. Shots fired.
2:01:36
I'm just No, he's a he's a great guy.
2:01:39
We just like some hit pieces on our competitors, please. Of course. Yeah.
2:01:42
We're lucky to not be in the hit piece business. We're not.
2:01:47
We're We'll we'll review.
2:01:47
We're sponsoring the wrong show. Yeah. Yeah. It's rough. Yeah.
2:01:52
I think uh I think just buy like a 100,000 subscriptions to the information and then start putting pressure on them.
2:01:58
Say, "Yeah, you might want to look into this company."
2:02:00
I I would be down to do a a hit priest about uh technological stagnation. Yeah, I hate stagnation.
2:02:07
And so I would I would want to take down that as a concept.
2:02:11
Really slur that whole or or closed IPO windows.
2:02:15
Be prepared for a terrible hit piece on closed or hit pieces on on just CEOs that take their foot off the gas. Totally.
2:02:23
You obviously, you know, have not the foot's been I've got two feet on the gas. I think that's possible.
2:02:28
It's a bit irresponsible, but yeah. Yeah. Yeah.
2:02:33
Walk us through the story of uh that that that you posted, how you rebooted 15year-old decelerating business.
2:02:38
I want to hear this uh from kind of set the table for us and then we'll walk through the story because I think it's fascinating. Yeah, sure.
2:02:45
I mean, you know, it's a 15year-old business.
2:02:48
It's a successful SAS business.
2:02:50
We're in the service game.
2:02:52
Um but at the end of our kind of first chapter, things slowed a little.
2:02:56
We were unfocused, bad commercial decisions.
2:02:58
This happens to successful companies.
2:03:00
They become a victim of their own success and comfort creeps in.
2:03:03
Definitely 2020, 2021 were some comfortable culture times. and I got sick. I had to leave.
2:03:10
So, it's a it it's a it's a it's a big long story that ultimately comes down to the fact that we lost our way a little bit and we had like five quarters of decelerating revenue.
2:03:21
I came in midway the fifth and it was looking kind of uh gloomy and the two things we changed were we went to back to good oldfashioned SAS fundamentals pricing that people liked selling the product in the way that people liked.
2:03:38
They used to have to like talk to sales for everything and it just those simple things becoming super customer first um started to really accelerate the previous SAS business.
2:03:48
In the last eight quarters the growth rate of the SAS business has increased by 10x which is really remarkable.
2:03:54
Um but then of course we jumped on AI and we were kind of OG AI guys.
2:03:59
We had dabbled not dabbled.
2:04:02
We had developed, you know, real AI products before, but they were baby AI compared to what we all have today. Um, but as soon as GPT3.
2:04:08
5 came out, we all just jumped on that and we saw that there was opportunities for this whole new category where you could create what we call now customer agents doing all of the things, customer success and service and sales and marketing that, you know, humans used to do and hate.
2:04:25
Um, and that just propelled the business even further.
2:04:28
Um, Finn, our customer agent, is now, you know, the best performing in that category in our benchmarks.
2:04:36
We win every bake off against our our chief our primary competitors.
2:04:40
We have the most customers, most AR or so.
2:04:43
So, we're kind of this very weird story that I don't know any comparisons to where we're previous generation SAS that's actually winning in the category in AI.
2:04:52
It, you know, I think it's hard.
2:04:55
It's really really difficult for the previous generation, the slower older cultures that work in the age of AI.
2:05:02
It requires a lot of agility and dynamism.
2:05:04
I often mess up that word, but it really does.
2:05:07
Talk to me about like the the different break points for growing a company.
2:05:13
I feel like mentally I think about it as like just the founders, maybe the first 10.
2:05:21
Then we were talking to previous guests about this breakpoint at like 50 people like there's something about there's a magic of a 50 person team.
2:05:30
Everyone knows their everyone knows each other's name.
2:05:35
Then maybe there's other intercom AI group has 47 senior engineers and and researchers.
2:05:41
So right in that that 50 person sweet spot.
2:05:43
But but I feel like I feel like there are like with in the story of startups, we often map them to funding rounds, seed round, series A, series B.
2:05:52
And sometimes the headcount grows in line with those, but some but I feel like headcount growth might be more of a factor in like cultural drift.
2:06:00
And I want to go through some of the key moments where you feel like um like you know it was only one foot on the gas or the or the foot came off the gas or what what are the upstream drivers of that?
2:06:15
What are the things culturally that you think startups need to get right at various scales as they grow?
2:06:18
various scales as they grow? because I feel like there's always these these different moments when you're when you're scaling up and you have a whole bunch of decisions to set the culture and you have a pretty limited time and you're focused on product and revenue
2:06:33
and growth and all these other things, but culturally there's some very they're very important decisions that get made at every I don't know if it's every order of magnitude, but there's these key milestones and what tell me the story of of the the milestones in your mind. Maybe it's shifting offices or
2:06:46
Maybe it's shifting offices or fundraising or or headcount milestones.
2:06:50
Um but but what changes and what advice would you have for founders at every stage?
2:06:54
That was a fivem minute question. Outstanding.
2:06:57
Um sorry I've given you a hard time.
2:07:00
Um I um look there's a kind of an intellectual set of answers to this that you can kind of break down and break it into tips.
2:07:09
into tips. there's a kind of a more abstract thing which is both you know in in good instances self a grandandizing for someone in my position but then also bad news in other instances and the answer is that it it all comes from the top and the early days the founder typically certainly founders that you
2:07:29
know have any degree of success at the start in the early days the founders bring a phenomenal amount of energy conviction whether it's founded or not um you know just just just belief obsession, uh, intellectual curiosity, excitement, passion, you know, a lot of intangible things and that really drives great people. All of us want to make
2:07:47
All of us want to make great money in this industry and that's that's awesome and I and I really think it should be celebrated.
2:07:53
People are too shy to talk about that, but they also want to be part of something meaningful and exciting and they want to work with people that inspire them and make them want to push themselves.
2:08:03
And so the reason a lot of these older generation companies lose a lot of steam.
2:08:08
Is that just for very obvious human reasons?
2:08:11
The person on top is not pushing in that same way.
2:08:15
When you have 15 years of SAS, how exciting is every day going to remain?
2:08:22
Like honestly, like the first year you're like, "Cool, SAS, churn. Huh? Wow. Okay, I get the math."
2:08:27
And then in year two, you're like, "Okay, churn, get it. Cool. Raise some money." Year three, road mapaps.
2:08:33
year 15 of SAS, you're done.
2:08:40
You're not bouncing to the office every day.
2:08:42
And and people will pick up on that all around you.
2:08:45
Of course that they will.
2:08:47
And then you don't push yourself in the same way.
2:08:49
You don't really pitch the opportunity to new employees.
2:08:52
You settle a little bit cuz life is hard.
2:08:54
You've got other priorities.
2:08:57
Maybe you've drifted a little bit.
2:08:59
You've got side projects.
2:09:00
Some people end up with families, girlfriends, ex-girlfriends.
2:09:02
Like life gets way more complicated than it is for a 26-y old kid who just moved to San Francisco and one has one of those buzz cuts and the curly hair on top.
2:09:09
Like life just gets more complicated and that's that's what happens.
2:09:13
And so part of our secret is that AI reinvigorated us. Yeah.
2:09:20
Like I would not still be doing this if we were just doing SAS.
2:09:23
SAS is not only kind of easy but super boring to me now. That's okay.
2:09:28
Hopefully AI and whatnot will get boring too and there'll be something new.
2:09:32
And so again, we could break it down and get all mechanical and try and pull out some like tips and tricks and advice here, but really it just comes down to energy.
2:09:39
And so so for anyone who would want to reinvigorate their company, it's the question is how how can you reinvigorate yourself?
2:09:46
And I see a lot of founders of latestage companies, many of them public.
2:09:51
You kind of haven't heard from them for years.
2:09:55
their stock price has gone sideways for five maybe seven eight years and I'm like what are they still doing and I wonder are they able to admit to themselves that like they don't want to do this anymore and if you don't want to do it anymore make a change like kind of move on um and so I think a lot of
2:10:14
people just they struggle with that moving on and making that decision because their whole identity and sense of purpose and validity in the world comes from I'm CEO of whatever so it's like this deeply human squishy spiritual challenge rather than an NBA type challenge. What about uh bringing in
2:10:31
What about uh bringing in young people to kind of keep that reinvigoration process going?
2:10:37
I'm just thinking about uh you know Zuck is paying so much to bring in Alexander Wong from Scale AI.
2:10:44
At the same time, you know, like the level of energy that Alex is going to bring to that organization is potentially worth a lot, you know. Yeah.
2:10:55
But at the same time, Zuck is super high energy. Yeah. Right.
2:10:58
But but but there's another world where you low energy people hire low energy people. Yeah.
2:11:02
But I I guess what I'm getting into the trap of is like you can be the high energy founder as your business becomes more serious.
2:11:08
People keep telling you like bring in the seasoned executives, bring in the bring in the gray hairs, the people who will will keep the you know steady hand on the tiller and and that can be lead to a less dynamic less lower dynamism in your organization.
2:11:26
Is there is there a hack to just hiring crazy young people and empowering them to be in the seauite whether or not they really like deserve it by traditional u standards?
2:11:39
Yeah, like the challenge is super obvious, which is these young, crazy, energetic, optimistic, wideeyed people are super messy, super sloppy.
2:11:48
They get in fights, they get upset, they hung over late, like they don't know how to do larger company professional things. Sure.
2:11:59
And so part of the problem is that larger companies to scale and get more efficient and become global organizations across many offices and time zones is that they like introduce a lot of like uh uh uh regularity and they like iron out the chaos.
2:12:12
So part of it is you have to be willing to entertain chaos.
2:12:16
You have to be willing to put younger people in positions of influence and let the chips fall where they may.
2:12:23
It's possible to give them roles where they don't have to engage with the entire organization.
2:12:28
Like we've definitely got roles in intercom where you can have to collaborate across two time zones.
2:12:33
Sorry, across eight time zones in two different teams.
2:12:34
But then we've got other positions where you've got one super smart guy.
2:12:38
He's 30, which is 10 plus years older than the execs.
2:12:46
But you give him like one thing he can do on his own and he'll crush it. Mhm.
2:12:51
So part of it is knowing how to like work with these people, but also like this is a special type of X-factor young person who knows what they don't know.
2:13:01
And yeah, the degree to which this is a talent game and that people are not funible is not recognized at all.
2:13:06
People imagine like, oh, you lost one person, you get a back fill.
2:13:11
Entire organizations just flip and change completely when you change out the individuals involved. So yeah, it's not easy.
2:13:18
easy. Do you think venture should take uh there should take almost like turnarounds more seriously like in some ways you were your own turnaround CEO but one of the I think the issues of the venture industry is let's say a company becomes a unicorn has$und00 million plus
2:13:36
of ARR and then the sort of growth starts slowing maybe the CEO like gets bored or whatever they start partying or they go start going to Europe um and the VCs kind of write it off and they're like I made my return or at least I'll get my money back. But but at the same
2:13:50
But but at the same time, I mean, private equity is built like you know, there's been empire has been built around like the turnaround.
2:13:58
And in some ways, think about, you know, a talented founder, maybe they took their first company through YC and had a nice exit.
2:14:05
A lot of those people could go to a company that has like a hundred million of revenue and like a big customer base and like actually make more money and start on you know second or third base and take you know you you can make quite a lot of money taking a business from a 100 million to hundreds of millions of revenue and that that can sometimes be easier than taking it from 0 to 10. Totally. Yeah. What do you think?
2:14:29
I think theoretically I think you know VCs are best are pattern matchers and turnarounds don't fit the pattern. Mhm.
2:14:39
You know think of all of the most successful and exciting zeros of technology over the last 20 years.
2:14:45
They invented a thing something something something. It's worth 10 billion. Like it's kind of that.
2:14:52
It's like yes sometimes it takes a little bit longer.
2:14:56
there's a slightly circuitous route, but it's not the company was totally failing and they had to reinvent themselves and then they became the biggest thing ever.
2:15:03
So, you know, for VC, I just think it's really really hard that it's it's just hard for them to get like the underlying narrative and the underlying story.
2:15:10
This is where PE comes into play, but PE has all of its own problems, too.
2:15:15
And these guys want deals and they won't be exciting to uh a lot of people who started ventureback companies.
2:15:21
It's it's straight up difficult.
2:15:23
And to my point previously, the idea that, you know, talent isn't that funible, you know, you take any given company, if you replace the founder with even another highly competent founder, they're probably not the chance that they're right for that opportunity and idea.
2:15:39
Like, look, there's so many people, you know, so much more accomplished than I am, but I'm pretty accomplished.
2:15:46
I know how to run and build and reacelerate businesses, but I'd be a probably a shitty CEO for 99% of other companies just because, you know, that's not what I do and I don't have any experience there, etc. , etc.
2:15:57
I don't even know the people there.
2:15:59
So, I think people should be bearish on turnarounds. Yeah.
2:16:01
You know, the turnarounds don't really work. Yeah.
2:16:05
They're like generally like a it's a failed thing. Yeah.
2:16:07
Somebody will figure out. Maybe it's Jeremy Gon. Maybe he'll do it. Yeah.
2:16:11
Well, well, that's even a different strategy.
2:16:14
But but yeah, I think this idea of like you need to kind of they like the idea of bringing in like a crack founder into a company that's it's a crack founder wants to do their own thing.
2:16:24
They want to start from scratch.
2:16:25
They want all the equity themselves.
2:16:27
Like the recap alone that it would take just won't be palatable to existing investors.
2:16:31
Failed companies are just generally doomed to fail.
2:16:33
And when there are so many opportunities out there as an investor, you know, you got to just like not try anything novel. Yeah. Yeah.
2:16:43
And in your case, it's like a little bit of luck, the timing of like you're going back in GPT35, you know, seeing the opportunity for a new product, all this stuff.
2:16:50
Uh, but you also had to make the choice to risk your own ego to go back in.
2:16:54
And if revenue had decelerated for another five quarters, you'd be sitting there being like, "Yeah, maybe I'm not as good as I thought I was."
2:17:03
You know, and you have to It's only true.
2:17:04
But I got to cheat a little bit because when I was out, I was like sick.
2:17:08
I had been beaten up in the press.
2:17:11
I was like just my confidence was pretty low and I didn't really have a lot to lose and I felt like I was without purpose.
2:17:21
I always wanted to be independently wealthy and free and I finally got it.
2:17:24
It was in many ways magical and then completely boring.
2:17:27
And so when I had this opportunity to go back have purpose and I had nothing to lose, I took it.
2:17:35
So, like it's easy now to tell this maverick story. You're so brave. Look what you did. You took a big risk. No.
2:17:43
And you have nothing to lose. You'll just go for it.
2:17:45
And and I think part of the secret is if people can separate themselves from their egos um a little or work on their egos or learn to love their egos and not be run by their egos, great things are possible.
2:17:57
Most bad decisions are made just out of fear.
2:18:03
And the fear is driven by just fear of public failure and embarrassing yourself.
2:18:07
I found myself unafraid to embarrass myself.
2:18:09
Look at how I'm speaking to you now. Amazing. I love it.
2:18:14
Like it's not fully true.
2:18:14
The ego is still there in present. Totally.
2:18:16
But the the the smaller and weaker it gets, the more freedom you have. It's fantastic.
2:18:22
Well, thank you so much for stopping by. Always a pleasure.
2:18:24
We could yap like this for hours.
2:18:25
I feel like I feel like people are going to listen to this as like a little founder therapy. 100%. I was like this.
2:18:30
We can do a little therapy corner. It's amazing. Yeah.
2:18:32
Once a month you come on. Pump up speech. Pump up speech. It's great. This is great.
2:18:38
Diet of meditation if you're interested. That'll be the next one. Thank you, J. Hey, this is cheesy.
2:18:43
I want to give a shout out to my friend Stewart. That's it. I promise I'd do it. Amazing. Shout out to Stuart. Word for Stewart.
2:18:51
Do Do we need to ring the gong for Stewart? What's Stewart do?
2:18:53
We got to ring the gong for Steuart. Okay, ring the gong. He's had a big year.
2:18:57
He's had he's had a big year.
2:18:58
Congratulations to Stewart.
2:19:02
Stewart, let's go Stewart. Congratulations. We will see you soon.
2:19:08
Have a great rest of your day, guys. Talk to you soon. Peace.
2:19:09
Up next, we're staying in in the Irish hour.
2:19:12
We're going over to Stripe. Stripe.
2:19:14
Luck of the Irish hit intercom.
2:19:17
We'll check in on how the luck of the Irish is treating Stripe.
2:19:21
We got There he is from Stripe. Welcome to the stream. How you doing?
2:19:26
The moment we've been waiting for.
2:19:26
We're so sorry for a week. a couple weeks ago. It wasn't our fault. It wasn't our fault.
2:19:33
Geopolit geopolitics is currently outside of each of your controls.
2:19:36
Uh well, that wasn't that wasn't even geopolitics domestic politics.
2:19:39
That was South Afric South African attacking an American on the timeline. A reality TV star. Yeah. Former reality TV star. Yes.
2:19:49
Jordy, I have to say it's really awesome to see you in this format because you and I have been zooming for I think almost a decade now and now it's live in front of all these uh this great audience.
2:19:58
It's really great to see what y'all are up to. It's It's a bummer.
2:19:59
I don't We've never met in person, but I've had so many Zooms with you in this exact room.
2:20:05
I have a theory that like you'd never leave this room, actually. But we're busy.
2:20:10
Yeah, you're you're busy. Yeah.
2:20:11
What What is the major update?
2:20:13
We we we wanted to have you on to talk through it.
2:20:15
Can you break it down for us?
2:20:16
I mean, I think it's more of a like it's more of a conversation.
2:20:19
Jeff Jeff's like evolved his role over the last year was was running point on Atlas made it made it a platform that a meaningful percentage of C corps I think are started on Atlas today.
2:20:30
Yeah, about one in six now are on that list.
2:20:36
Um, but about halfway through last year, we looked at what was happening in AI and started to get really serious at Stripe about not just the application of it inside of our business for preventing fraud and running our own uh payments foundation model, but also to help developers and businesses and consumers get ready for when AI starts to come to commerce.
2:20:56
I'm still a little surprised that we got self-driving cars before uh ubiquitous online commerce is mediated by agents.
2:21:06
But you can really start to feel that AI is now coming very close to commerce uh and will be part of buying decisions, discovery, execution of transactions and new ways that businesses can find their audiences online.
2:21:18
I mean I'm really quite impressed to see the rate at which discovery has changed and it feels like around the corner uh commerce and AI is going to be very closely mediated.
2:21:30
Talk about uh maybe some some early product experiment what you what you guys are experimenting on what you guys have already rolled out uh all that stuff.
2:21:39
Yeah, what we've been trying to work with the fastest growing companies as they push the frontier of agentic commerce.
2:21:47
So, one of the first we worked with was Perplexity where they have this buy with Pro package inside of Perplexity where they show great e-commerce search results and then when you go to buy, you're not going to the merchants tab and dealing with the merchants web page.
2:22:02
uh you are actually just clicking buy and in the background a Stripe virtual card is spun up and given to an agent or any other automation process so that you can just have a completely seamless experience of buying in situ to where you're doing discovery and we're starting to see that in more and more places.
2:22:19
So recently, HIPP Camp, which is, you know, the cool kid way to book camping online, sort of Airbnb for uh places, they started to partner with Stripe to make national parks and state park inventory available to a wider audience because some of those checkout pages are hard hard to use.
2:22:40
That inventory is not naturally online, but these are amazing places for people to be able to camp.
2:22:44
But there was just a huge I remember as a kid I I remember as a kid there was a there was a a place my family used to always go camping and my dad would like wake up at 5:00 a. m.
2:22:54
and just be refreshing this like terrible site when like the the campsites are so ready for the Age of Agent and it was like very unreliable like payments.
2:23:02
So it was like it was the equivalent of like a street wear drop but like the you know like you know some state park was like managing it.
2:23:07
Um, I think we're going to see this more and more where the the inventory of the world is getting closer and closer to intent and agents are way to bring that bring them together and then it opens up really interesting questions that Stripe is trying to help answer developers.
2:23:24
What is the developer experience for being able to uh execute those purchases?
2:23:28
We have this new order intense API that we're triing where you can just give a product URL and one of our agents will go buy it on your behalf.
2:23:35
behalf. uh we have we have new ways for businesses to be able to start to expose their inventory to agents in a safe and permissioned way and then as a consumer you know you should feel it it is reasonable to think actually that agentic processes is the last place you'd want when it comes to money uh you
2:23:54
actually want that to be incredibly permissioned safe deterministic you know what's going to happen and so you can expect that the Stripe APIs are going to evolve for a new type of user in the world which is an agent that can safely be delegated with your permission to buy on your behalf. Can you talk about uh
2:24:09
Can you talk about uh Stripe Link and how that product might fit into uh a product like Perplexity?
2:24:18
It feels like um it's great if it's one of those classic things in in AI and tech is like, "Okay, okay, great.
2:24:25
It it uh you know, it surfaced the right product for me.
2:24:29
Now I want it to buy it for me even faster.
2:24:30
Now I don't even want to go through the checkout process at all."
2:24:33
like there's like as soon as I get the the the the current thing, I want the next thing.
2:24:36
Um so, uh how h how do we see um that playing out with just making that commerce experience even more seamless or happening entirely inside of a chat interface or an agentic interface?
2:24:50
interface? Yeah, we you know the borders of the internet are starting to blur and so you will soon be able to experience if you chat if you if you search for something on chat TV they already have these cute little uh shopping carts that link you out if
2:25:08
you're sitting in cursor and you need access to a database cursor can recommend superbase and even start to accomplish your homework for you right in the editor but there is this like missing moment here right where okay Now I know about these products. What am I
2:25:21
What am I supposed to do? Go to a new tab?
2:25:23
Do an offline kind of feeling search?
2:25:25
Go through a bunch of blue links, find the website, go to the website, make an account, deal with the password problem, get a bunch of weird emails to confirm my password, find the settings page where I can get the billing information, pick my billing thing, put in my payment credential, get my API key, walk it all the way back.
2:25:45
It's like this, you know, I think we will start to see this as this loop that we've all been operating under for the past 20 years of the internet as very arcane very quickly.
2:25:56
Whereas you just want to delegate uh your payment credentials to a safe trusted place and stripe link is this payment wallet we've made over the last few years which is a cross internet payment wallet that works with cards and bank accounts and future other payment
2:26:11
methods where if you log in once to link then you will be able to delegate safely your permissioned credentials with a virtual which with a virtualized token such that uh you can safely hand it off to a good good robot to buy on your behalf. And so we we see this as a new
2:26:25
And so we we see this as a new borderless way that commerce can happen in a very permission safe fashion. Yeah.
2:26:33
Um how how are you thinking about uh agentic commerce and stable coins?
2:26:38
A lot of um you know there's a lot of commentary around uh stables and how they can be applied here.
2:26:46
they can be applied here. oftentimes the people that just sort of default assume that agents and agentic software will use tokens uh you know whether they're stables or other tokens they usually have crypto you know funds or or crypto companies right so I've had maybe a more um middle of the road view where I can
2:27:06
imagine aentic uh commerce experiences leveraging stable coins I can also imagine them leveraging cards and a and a bunch of other sort of forms of payments uh so I'm assum I mean you spent a lot of time thinking about this and you guys have obviously been acquisitive recently with with bridge and and privy um as well. Yeah, this is
2:27:23
Yeah, this is one of the areas in which stripe is very problem solving solving oriented and not uh technology or particular um technique religious.
2:27:36
Uh we think that humans are going to have a variety of ways that they want to pay and hold money.
2:27:41
Stable coins is a phenomenal way for many people in the world to to hold hold funds and for businesses to move it across borders.
2:27:48
And so we expect that stable coins will be a very popular way for consumers and businesses to just interact with themselves.
2:27:55
Then you have businesses who are also going to have you know it they're they're going to have a long adoption curve uh when it comes to uh accepting and holding crypto assets.
2:28:06
And then in some purchases, stable coins might make sense between two parties that natively know how to interact in stable coin, but often it might be the case that Jordy has an MX card and the seller is expecting an a transaction and we're sort of missing a universal way for all these types of currencies and rails to work together.
2:28:28
Visa also announced a new way of of being able to to hash your card and give it to an agent with this Visa Agentic token where Stripe is one of the first partners to implement it.
2:28:37
And I think we're just going to see this new proliferation of new ways that money can can transact between parties.
2:28:46
And we're going to need some type of sort of babbleish translation service across all of them because if you're going to pick one route then you're going to likely exclude many of the agent humans and businesses in the world. That makes sense.
2:28:57
How are you guys thinking of uh not not to go too broad but the business model of the internet agents you know change things.
2:29:06
The internet today is heavily reliant on uh ad you know advertising and if you have a bot you know just crawling a website you know you're or or even when you look at other other services and so we've talked Ben Thompson had some good writing around um just like what the future business model of the internet could look like and potentially micro payments.
2:29:26
But I think the takeaway from that, our takeaway is like there's so many different stakeholders that would need to find some type of alignment.
2:29:33
Uh it's it's hard to see like the obvious path forward here. Yeah.
2:29:39
I I think the univer a universal want from businesses is just more channels to reach their customers and to be able to do so in more direct kinds of ways.
2:29:50
And so if you go to, you know, a a SAS software provider and you said, "Hi, you know, I sort of two choices for you.
2:29:56
choices for you. You can have this um very cool large budget for a 101 billboard and kind of hope that at 85 miles an hour developers like see your ad and then remember to implement it later or would you like them in situ as they're working to have agents mediate the purchase, recommend it and be able
2:30:16
to like integrate an accomplisher thing in 5 seconds right right inside their editor like okay yes well first of all I'll do both but also this second one sounds very nice because I'll be able to direct directly attribute uh where it came from and be able to have a great uh
2:30:31
uh CAC for that and you know the LTV should be even higher because the robot even integrated it directly and so I think that we're going to see new channels emerge for monetization both usage based through MCP or other ways that businesses are going to expose their APIs to agents but also for
2:30:48
transactionbased referral fees which which will supplement uh affiliate um and then I think it'll be a new way for businesses to make sure that agents can read their docs, can read their product SKUs, can have access to that information in a new permissioned way. I
2:31:02
I I really like the car copy talk that got posted last night where he basically said that if your docs involve a click, not good because agents want to act and not click and just only read.
2:31:14
They want to start acting.
2:31:15
And so that's why Stripe is uh you know if you go to the stripe docs it we really push like hey here's our MCP where you can just just talk to the primary best way of integrating Stripe and it can do it on your behalf rather than you know just reading or reading something from a three-year-old corpus. Interesting.
2:31:31
Um last question for me we we we want to move on let you get back to your day.
2:31:37
Uh, Stripe was famous early on for having this crazy kind of open culture around uh, emails that anyone from the entire organization could read.
2:31:46
That seems like incredible foresight to the moment today because you don't have all this private information that oh, do we train on that or not?
2:31:56
You could very easily uh fine-tune a model or do some sort of uh uh you know uh uh you know embedding on the emails that are already deemed to be worthy of the entire organization reading them.
2:32:11
Is that still part of the culture?
2:32:14
Is there a tool if you join Stripe where you can get up to speed without needing to read every email but you can kind of get the the Stripe way of doing X Y or Z?
2:32:22
Talk to me about Stripe's culture.
2:32:25
Stripes, you know, has a has a a very serious writing culture where any decision I I've been a part of for the last seven years.
2:32:34
I can really point to some Google doc that has the pros, the cons, and the decisions as well as the uh the the the sort of the email culture you mentioned where we just it's very common place at Stripe where if you write uh you spoke to a customer or even after going on TBNN, hi went on TBNN, you just CC a notes list and now it's available for anyone who wants to subscribe to notes list.
2:32:54
But one of the major subscribers to notes list now is agents. Interesting.
2:32:57
And so if if I'm in Slack, uh we have this really awesome bot called Trailbot that's read the trail of everything, all the paper trail of everything that we've done that's permission to it.
2:33:07
And I can just say at Trailbot in any Slack room and it has the context both of the team Slack room I'm in, but also the full corpus of of Stripe and all of our permissioned uh wikis and documentation and internal internal tools.
2:33:19
And it is it takes the first line of defense of most questions immediately.
2:33:24
And we actually have it to the point where it knows to jump in automatically without you even asking it.
2:33:29
And so I I find that most of the time we're able to just at trailbot and answer a lot of questions.
2:33:36
And then increasingly these agent tools which I think are going to apply to commerce soon quickly too. They're not just readon.
2:33:42
They're going to start taking right actions and purchase actions.
2:33:45
And for Stripe, internally write actions might be to roll back that deploy or to uh you know autocommunicate to that customer because of a an MPF score under 10, which which we do often.
2:33:56
Um hopefully not too often.
2:33:59
Uh but then in in the real world, if you want to make some of these actions, you're going to need to prove who you are, pay for it, make sure the merchant was able to accept that money, get the entitlement, and move on. Yeah.
2:34:07
Even something as simple as like you show up to a new company, hey, there's this system over here that we're using, and I don't have access.
2:34:15
You might go to a wiki and ask, "How do I get access?"
2:34:16
Now you just ask and it just does it for you.
2:34:20
It's so interesting to think about if if there's like some type of user flow where if somebody sends a Slack message, there's like a tiny delay built in and it gives like a bot an opportunity to like actually front run the question because it's like every message is going to waste like you know 10 minutes.
2:34:36
I a new version of shadow band where you first get your question answered and then go back.
2:34:40
Do you really want to ask this question?
2:34:42
Because it was answered here, here, here, here, here, and like here's our recommended action. Pro autocomplete. It's amazing. It'd be interest.
2:34:49
Yeah, Slack's just become completely silent because everybody's like doing things and it's just immediately getting at those of us who have nerdily taken notes and made docs over over years. That is somewhat. It was worth it. Yeah. Yeah. Yeah.
2:35:01
Made fun of by some people for a long time, but it all came back.
2:35:06
Well, thank you so much for stopping by. This was great, Jeff. Always welcome. Yeah.
2:35:10
Well, we'll have you back soon. We can talk more to you. It's great. Talk to you soon. Bye.
2:35:15
Uh, let's give it up for Jeff.
2:35:15
Next up, we have Garrett from Handshake coming in.
2:35:20
He was mentioned in the information.
2:35:22
We've been mentioned in the information.
2:35:24
It's a bunch of information boys hanging out on the chat. We love the information.
2:35:28
We love We love the information.
2:35:30
Thanks so much for joining.
2:35:32
How you Garrett Lord, the nominative determinism is insane. Yeah.
2:35:37
I think we love something also in common. Sonning. I'm a big big sauna guy. No way. There we go. Yeah.
2:35:41
Yeah, the sauna is important.
2:35:43
You'll be devastated to hear that when we moved into this new studio, we we don't have a good We don't have a good sauna set up, but we'll figure it out eventually.
2:35:50
The cold plunge can be can fit nearby, though.
2:35:53
I mean, that's there's still opportunities. Be good.
2:35:54
Yeah, maybe we got to get in the cold plunge game. Uh anyway, in full suit.
2:35:57
Anyway, uh kick us off with a little introduction on the business.
2:36:02
Uh obviously, it's in the news today.
2:36:04
We covered a little bit about it earlier, but I'd love to get you to explain the business, a little bit of the history, and the positioning of the company. Yeah, for sure.
2:36:11
So, I mean, the business started way back when I was in college.
2:36:15
Um, I started Handshake out of a personal pain that I faced in breaking into find my first internship and first job.
2:36:21
Um, I went to a no-name school in the middle of nowhere called Michigan Tech.
2:36:24
It's awesome if you love to ski or love the cold, but uh if you wanted to break into Silicon Valley, nobody had really recruited there before.
2:36:30
Uh fast forward to today, Handshake is the number one place that young people in America start, jump start or restart their career.
2:36:38
Uh we're like kind of an 18 to30 early career network.
2:36:40
There's a million employers that use Handshake.
2:36:43
So it's where the vast majority of employers recruit undergrads and interns and people after school.
2:36:49
Uh and then there's 18 million students and young professionals use the network.
2:36:53
And we also power uh about 1,600 universities in the country. Mhm.
2:36:58
And the background I think uh that's important for right now in this very moment is about 18 months ago many of the frontier labs as well as the large annotation uh engine companies started reaching out to us with basically asking us beating down the door saying like do you have access to PhDs?
2:37:16
Do you have access to master students?
2:37:18
And uh for us that was incredible.
2:37:21
I mean we have 500,000 PhDs in the network.
2:37:23
We have 3 million master students on the network.
2:37:25
There's tens of millions of undergrads in the network and we started serving these players with uh uh experts really as this the world has evolved from training frontier models.
2:37:37
It's moved from generalists like drawing kind of boundary boxes around stop signs y to today experts and experts are in law, finance, medicine, mathematics, physics, chemistry, biology.
2:37:50
These labs uh really are hungry for reasoning data to help improve with human in the loop the actual uh you know frontier of uh what their models are capable of delivering uh yet alone in the future and you talk about like tool use or trajectory.
2:38:05
So they started reaching out to us and saying do you have access to these PhDs and master students and we started providing we were the leading provider of all this talent and we really started to realize is that people weren't getting paid on time.
2:38:17
They were really confused.
2:38:17
they would go through training and kind of get dropped out of a leaky bucket.
2:38:20
Um, we heard from students that were successful on it that they they loved the money, they love learning more about some of this AI tooling.
2:38:28
They wanted to use AI tools in the classroom.
2:38:29
They wanted to use it in their research.
2:38:31
And so given that we have this huge supply and zero customer acquisition costs, we started building a human data business.
2:38:40
Um and really in the con construct of building that business, the the focus is really around like how can you also think about evolving and automating a lot of the recruiting practices.
2:38:50
Recruiting is still you know it's sourcing, it's screening, it's scheduling.
2:38:53
There's a lot that AI can bring to bear on that.
2:38:55
And so we now fast forward to today in the last 6 months have been working now with six of the frontier labs.
2:39:01
We provide them tens of thousands of It's a lot of them.
2:39:06
I didn't even know there were six.
2:39:08
They're only five, the big six. Count them up. You got them all.
2:39:12
And uh we provide them with experts to help make their models uh more effective. Very cool.
2:39:17
Um talk to me about like what the how how are the Frontier Labs thinking about human data annotation and answer generation.
2:39:26
It feels like we might be at the end of that story soon or maybe we're shifting into uh more of a focus on the areas that are less verifiable uh less like write the answer to an IMO level math problem and more in the biology and legal context where the models are falling behind like like where where are the pockets of value?
2:39:50
Where's the most demand within the human data generation industry and where do you see it going over the next couple years? Yeah.
2:39:57
So maybe I'll go like from the latter part of the question to the first.
2:40:02
So like where we see it going over the next couple years.
2:40:04
It's definitely going to evolve into audio.
2:40:06
It's definitely going to evolve into tool use.
2:40:07
It's definitely going to evolve into trajectories.
2:40:09
Uh and experts will be needed to uh provide data.
2:40:13
Um imagine almost like recording your screen as you're conducting a task.
2:40:18
Maybe you're building a slide deck and doing it.
2:40:20
you know, if you're an investor construct like doing a DCF and doing competitive research, they want more data to be able to help uh improve these models, especially as you think about like agents, right?
2:40:29
And step-by-step problem solving.
2:40:30
Uh as where the puck is right now and where the puck will continue to be, if you talk to a lot of the frontier researchers, is they need expert data and uh expert data is in basically every esoteric area of human knowledge.
2:40:44
They want to, you know, they'll the models have already kind of sucked up the entirety of books and YouTube and, you know, human knowledge.
2:40:52
And what they really need is they need special data to be able to make and understand the step-by-step reasoning that's required in order to be able to to kind of fuel the future.
2:41:03
And so, if you think about academia, these PhDs, like what is the definition of getting a PhD?
2:41:07
The definition of getting a PhD is like pushing forward an area of research that nobody else has done before as peer-reviewed by your peers.
2:41:13
That's how you get your doctorate.
2:41:14
And so this kind of perpetually reoccurring stream of of PhD students and master students are really valuable in this very moment.
2:41:22
And it's also to zoom out to their experience like you can make I don't know if you remember when you were in school, but you can make like 23 bucks an hour being a teacher assistant.
2:41:29
You know, you could dive you could drive Door Dash.
2:41:32
And we're paying these students like 60 70 80 h 100red plus dollars an hour.
2:41:36
And they're also we can connect it to actually getting jobs.
2:41:39
So, we envision a world where like you get badges on your profile and there's like leaderboards by school and we're actually I mean what better way to articulate your skill than actually proving it by being able to break the model or or by being able to provide the model feedback.
2:41:55
And so, we believe that we can help you get more jobs with the million employers in the network, help you build your professional reputation and articulate your skills all the while while while making like $100 an hour uh when you want to.
2:42:08
I mean, it's it's a gig job. Yeah.
2:42:10
How do you think about financing Handshake going forward?
2:42:12
I'm sure you're making uh generating a lot of revenue.
2:42:18
You're clearly uh paying a bunch of your your network out quite a lot.
2:42:22
Um uh we were just learning about Surge AI earlier and what they were able to do while bootstrapped.
2:42:27
I I imagine even in the last week, you've had investors reach out trying to, you know, say, "Hey, scale's out of the game.
2:42:34
You wanna you want a 100 million? you want to dance.
2:42:38
Uh but how how are you thinking about the business going forward?
2:42:41
Um yeah, I mean one of the the ways we think about this market is like you know if you don't have an audience there's no moat.
2:42:48
What our competitors are doing is they're at some of these companies they'll have hundreds of people who are recruiters sitting on top of platforms sending messages on companies like Handshake or spending tens of millions of dollars a month doing performance advertising trying to acquire experts on Instagram.
2:43:03
You can imagine if you're like a physics PhD and you get an ad on Instagram for a company you never heard of before claiming they could pay you $100 an hour.
2:43:10
It's kind of a jarring experience.
2:43:11
And so because we built a decade of trust in adding a ton of value to these users lives, we have no customer acquisition costs.
2:43:18
And what that means is that we can pass along all those savings by paying contributors. We call them fellows.
2:43:24
It's the move fellowship program.
2:43:26
We can actually pay you more than any other vendor on the market.
2:43:29
Um we can also pass along those savings to the frontier labs.
2:43:32
So as you think about our overall P&L like our gross margin and ability to scale this business considering you know the moat is the network that we've built we sit in an amazing position to you know to to grow extraordinarily quickly and that's what we've been seeing.
2:43:47
I mean in the last you know month we've grown by over 3x and you know there seems like there uh there's a lot of demand uh continue to be out there. I can imagine.
2:43:59
I had no idea it was that big though. Let's go. Let's go.
2:44:04
Three hits in the conference. That's incredible.
2:44:07
Uh are there any last question, we'll let you go.
2:44:10
Um are there any like weird areas that you think we'll see this type of human data generation pop up?
2:44:17
I'm imagining like AI seems to be at like 150 IQ.
2:44:21
It can write code and yet it can't like book me a flight.
2:44:25
Do we need to take like flight uh like travel uh what travel agents and have them go through the workflow so that they don't get hung up on should I sign up for the credit card or do I want you know insurance on this flight so that we have a whole bunch of data specifically about that task.
2:44:41
I'm just interested in this concept of like these economically valuable but highly niche tasks that don't seem to be we don't seem to be getting closer and closer and closer to like oneshotting them with the current models.
2:44:53
And I'm wondering if we're going to see this long tale of different hyperspecific business use cases like what we saw in SAS where there would be uh Hipmunk just help you book flights better.
2:45:05
Is there going to be a flow where there's a new startup that's doing AI agents for flight booking and then they're coming to you for a ton of data generation around how to actually book the correct flight because it learns whether or not you're okay with uh with a layover or how price sensitive you are.
2:45:23
All the things that you would get from the interaction with a human flight uh uh travel agent.
2:45:26
Is that something that you think we'll see or or is that kind of just completely tangential?
2:45:32
No, I think that's totally something we'll see. Interesting.
2:45:34
what what you just described is like a trajectory called a a browser trajectory. Sure.
2:45:39
And that's basically like you have a goal in mind. Yeah.
2:45:40
And you you you know you have like a step by step kind of thoughts in your mind around how you accomplish that and you navigate tools, you navigate the browser, you stitch together your own intuition to be able to accomplish that task. Yeah.
2:45:54
You might look at your own calendar. When do I get off work?
2:45:55
How look up how long it takes to get to the airport?
2:45:59
It takes me a different amount of time to get to Burbank than LAX. What's the parking like?
2:46:02
Like there's so it's such a simple task cuz you think about like anyone can do that job for you and yet to do it well is actually really hard. Totally.
2:46:11
And you talk about just being able to talk to a model, right?
2:46:14
Like totally even need to log in, right?
2:46:16
So you're going to need audio data, you're going to need trajectory data, you're going to be able to interact with APIs.
2:46:19
Uh humans experts will be needed for the next several years to to be able to make that data happen. Interesting.
2:46:25
In order to be able to power the frontier of where you want to see it going. Well, that's exciting.
2:46:30
I want to book a flight with an AI.
2:46:30
It still hasn't happened.
2:46:32
That's my own personal touring test.
2:46:33
Hopefully you can make it happen.
2:46:35
Uh but thank you so much for stopping by. This was fantastic. Sure. Time. We'll talk to you soon. Great to meet you. Cheers.
2:46:39
Uh coming in next, we have uh Tan uh coming into the studio to the TVPN Ultra Dome. Massive round. Oh.
2:46:48
Oh, we're going to hit the gong again.
2:46:50
The 10th time of the show. Always a good time. There he is. Welcome. You got news for us?
2:46:55
Hit us with an introduction. Hit us with the news.
2:47:00
What's going on in your world? Think we might be muted. Donnie, are you there? Can you hear us? Are you there?
2:47:08
I'm I'm I'm itching to hit the gong for you.
2:47:10
I hear there's gong worthy news. Are you there?
2:47:12
I'm going to send him an email. Okay, you are live. You are live on TVPN.
2:47:18
Okay, we'll pull them off.
2:47:18
In the meantime, I will tell you about Wander. Find your happy place. Find your happy place.
2:47:26
I can hit you with a gong.
2:47:28
Book a wander with inspiring views, hotel grade amenities, dreamy beds, top tier cleaning, and 247 concier service.
2:47:34
It's a vacation home but better, folks.
2:47:37
We told you about We told you about audio.
2:47:39
We told you about Poly Market. Told you about Linear. Did we do Vanta?
2:47:43
Automate compliance, manage risk, prove trust continuously.
2:47:45
Vant's trust management platform takes the manual work out of your security and compliance process and replaces it with continuous automation.
2:47:52
Whether you're pursuing your first framework or managing a complex program, let's hear it for vanta. com. Give it up for Vanta.
2:47:59
But if we got an extra minute. Is he here? Are we back? Welcome to the stream. We made it. How you doing? We made it. Fantastic. Sorry for audio issues. Oh, no. It was a pleasure. We got to do extra ads.
2:48:08
So, you know, you're making my day. It's a dream. It's a dream. You're making my day. Uh, what's going on? Quick intro. You've had a big day. What's What's happening? Yeah. Thanks for having me.
2:48:18
We We uh we announced a $200 million round. Oh. with General. That's fantastic. Buried the lead, Jody.
2:48:27
You guys got to start selling those.
2:48:30
I feel like we need one in our office.
2:48:32
Um, but yeah, know we're super excited.
2:48:35
Uh, obviously works in healthcare.
2:48:37
Uh, you know, we we power AI workflows, everything from ambient to revenue cycle payments in uh in large hospitals. Oh, interesting.
2:48:48
Give us give us a quick history of the company because It's not often I see a 200 on six something billion and I haven't I hadn't heard of the company before but I hadn't um Josh Browder connected us last nighting and um so I'd love to hear your quick story kind of how you got here history of the company.
2:49:07
I want to hear about the first customer too for sure. So I mean Josh is great.
2:49:11
I've known him since we were at Stanford together same year.
2:49:12
Um and the you know the story behind Kamir is interesting because Kamir started as an incubation inside General Catalyst.
2:49:20
it it you know the best analogy I have is it is uh Hmon's uh Palunteer very focused on healthcare uh I started a company while I was at Stanford calledis which was focused on applying language models and computer vision in healthcare uh we started as a blood diagnostics company and then eventually grew into this uh mid-market SMB OS for for physicians uh we merged the two companies about uh a year and a half ago or almost two years ago now.
2:49:51
Uh and then I took over as CEO with our management team.
2:49:55
And so it's really, you know, it's sort of a coming together of these of these two businesses.
2:49:58
Uh and yeah, I mean the company powers large hospitals about 250,000 physicians and nurses.
2:50:05
We power uh the you know private practices here in California.
2:50:11
That was our first first customer.
2:50:13
It was someone that my co-founder, Deepka, literally walked up to and, you know, cold knocked on their door and then got them to use one of our first devices and and remote monitoring solutions.
2:50:23
Uh, and uh, yeah, that's that's the the the quick story.
2:50:29
I got a bunch of questions, but I I want to kind of like contextualize this around the broader general catalyst uh discussion because there was news, I think just today, that uh Ohio authorities approved the first ever purchase of a US hospital by a venture capital firm.
2:50:45
That's General Catalyst's uh bid to acquire Suma Health, a hospital system in Akran with over 20 facilities.
2:50:52
And I'd love for you to I'm sure you've studied this. What is going on there?
2:50:56
And then how is that is there any sort of synergy across the portfolio?
2:51:01
Venture uh general catalyst has has had like a very differentiated strategy there.
2:51:05
Um but I haven't had the chance to dig into it.
2:51:07
So I'd love to get you to contextualize it and then we can go into uh how this links to your business again.
2:51:12
So the Suma transaction is super interesting.
2:51:14
It is a venture capital firm buying a health system transforming it.
2:51:20
Uh and Kamir is obviously a big part of that.
2:51:22
We're serving as the office of the CTO.
2:51:23
So our engineers are forward deployed.
2:51:26
We work hand inand with the uh Suma IT teams.
2:51:28
Uh we've been working with with with the revenue cycle leaders, the clinical leaders.
2:51:34
And it's a really special system. I mean it's in Akran.
2:51:36
If I'm not wrong, it's where LeBron James was born.
2:51:41
U literally the the hospital itself.
2:51:41
And um many people have been calling you the LeBron James of healthcare AI.
2:51:51
Now we got to put that out there. Um, yeah.
2:51:53
I mean, I I maybe have been the first person to say it.
2:51:55
You might have coined it here, but many people I'll say it right now.
2:51:59
You're the LeBron James of healthcare. Yeah.
2:52:00
Now, many people, not just one. Two is two is many. Yeah.
2:52:02
In our book, we're just going to make that a thing now. Um, fantastic.
2:52:06
It it's it's remarkable because running a health system is super hard.
2:52:11
It is a 1 to 3% operating margin business.
2:52:13
Uh, most of them go out of business.
2:52:15
uh and uh I think what what general catalyst believes in is language models and technology can transform the operating margin uh and also lead to better care.
2:52:27
So it's not a PE you know cut and and and juice play.
2:52:30
It's it really is a an investment. That's awesome.
2:52:34
Uh talk about uh Kir's overall product strategy.
2:52:37
You guys have a number of different products.
2:52:40
It seems like a very different you know that we we've talked with founders and covered companies that come on and and just want to own uh one you know one key area but uh healthcare feels like somewhere where if you can get embedded with the set of customers you can you know you know more you know rapidly kind of add add products to uh to the platform.
2:53:01
So I'd love to understand the the product strategy.
2:53:05
we we really look up to businesses like Ripling and and Rams um where you know there's this concept of you you enter with a wedge and in our case that wedge is uh either ambient AI which is a tool that helps a physician document and really automate the revenue cycle of their appointment generates the claim automatically.
2:53:26
claim automatically. Uh and then the back office which is all when you walk when you walk into a hospital there are tens of thousands of people at large health systems whose sole job is fill out claims call up insurance companies fight denials fill out new forms all of that's going away with LLM and our
2:53:43
belief is that if you do that as a point solution as like a single you know little part of the part of the solution you might get some initial usage but eventually the EMRs like Epic or companies such as ourselves will just eat you And uh you have to be that compound startup from the get-go. And I
2:53:58
And I think payments is a really interesting vector to deploy software.
2:54:01
RAMP has shown it where if you get into the transaction suite and then you build a whole bunch of tools for the CFO's office.
2:54:09
We're trying to do the same for a health systems CIO and and and CFO.
2:54:14
Can you tell me a little bit of the the history uh of the healthcare industry broadly and how uh I I I know that there was like this kind of catalyst around Obamacare.
2:54:26
I remember talking to Jonathan Bush, the founder of Athena Health about uh electronic health records mandates and there's been a number of changes kind of at the federal level that have kind of uh opened up different pockets of opportunity.
2:54:37
Like what is the story that you tell about the recent history of healthcare in America?
2:54:44
I think it's it's fascinating.
2:54:44
The 90s physicians had amazing lives.
2:54:47
I mean, they they drove Porsches.
2:54:48
They had work life balance.
2:54:50
They had personal relationships with their Let's hear for Porsches. We'd love to hear that.
2:54:54
Let's get back to Porsches. We need more of those. We need to return. We need to return.
2:55:03
And and you know, all in all, physic patients got great experiences too because because of that personal relationship.
2:55:08
And then, you know, the admin work tax just increased.
2:55:12
Everything from insurance to filling out an EMR.
2:55:14
Digitization came in the 2010s with Obamacare and meaningful use and really EMRs proliferated.
2:55:20
And Jonathan Bush and and Judy and, you know, all these people are legends in the industry because they built Athena, a $20 billion company, Epic, probably a hundred billion dollar company now on the backs of that very quietly and under the radar from for most of tech.
2:55:35
I think the theme and the story of today is labor is turning into software.
2:55:41
And where is most white collar labor in America? It's in healthcare.
2:55:45
Where like where where are the majority of administrators sitting behind a computer clicking on forms? It's in healthcare.
2:55:52
And we believe that the EMR will be transformed.
2:55:54
We also believe that the labor stack of healthcare will be transformed and it'll create more operating cash flow for hospital owners.
2:56:00
Is that narrative of the uh like the the administrative ratio or the administrative load increasing?
2:56:07
Is that similar to what happened in in academia?
2:56:09
Because I I remember seeing these charts of like the ratio of professor.
2:56:13
Everyone loves the idea of like uh a high functioning university with a lot of professors teaching students and a great ratio there.
2:56:18
Uh everyone's a little bit more skeptical about like wait why do we have five times as many people to add admin?
2:56:24
Is that the same thing that's going on in healthcare?
2:56:27
And and kind of what was the underlying driver of that?
2:56:29
Was it just regulation or or lack of tools? Where'd it come from?
2:56:34
I I think it's very similar.
2:56:36
What I will say is I think in healthcare it bred more out of necessity.
2:56:40
And in academia, it just kind of happened.
2:56:43
Um in in in healthcare, there's this game of attrition between the insurance company and the provider.
2:56:50
And you're they're making it a little harder every month, every year to get an approval on a claim.
2:56:54
And as a result, the health system needs to add a couple more people in order to fight those claims.
2:57:00
And then it just kind of built up into this arms race.
2:57:02
And I think the insurers kind of carried the power after Obamacare.
2:57:08
Like the when you look at United Health's market cap, I mean the it's like what is it like a 12x since Obamacare got passed.
2:57:13
It's it's quite shocking.
2:57:15
And the power dynamic I think will shift again back in the favor of physicians and hospitals because of LLMs and because of what you can now automate. Yeah.
2:57:24
It was kind of just like the the game theoretic Nash equilibrium was like hire a lot of a lot of admin staff. Interesting.
2:57:32
There was no other option. Yeah. Yeah.
2:57:34
Talk about uh your personal ambition and the team's ambition.
2:57:38
You're a six billion dollar company now.
2:57:42
seems like uh you know it's cliche but it the way you're talking it feels like you're just getting started. Is the job finished? Yeah. Yeah.
2:57:49
It sounds like the job's not finished.
2:57:51
I don't want to put words in your mouth. It's not finished.
2:57:55
Um we're Look, I I think when you walk into a healthcare practice, the inefficiency is shocking.
2:57:59
And it's and the the positive intent from the physicians and the nurses and the caregivers themselves is all there.
2:58:07
And I think all it takes is for a company like ourselves to come in and try to nuke that work tax.
2:58:13
Uh so our ambition is look, we're going to come after the EMRs.
2:58:19
We're going to come after the payers, the revenue cycle businesses.
2:58:20
Um this is a $4 trillion industry.
2:58:23
You can build for a very long time.
2:58:25
Uh what done looks like is when you walk into a physician's practice, scheduling, intake, insurance, like all handled.
2:58:35
There's no filling out a little clipboard of the same information again and again.
2:58:38
The appointment happens and instantly the doctor is paid out.
2:58:43
There's no reason we can't have instant adjudication instead of, you know, waiting 30 45 days, but it's going to require a system overhaul like new payment rails to go do that.
2:58:50
And that's really what what's at the heart of what Kamir is building. Awesome.
2:58:53
Well, this is super exciting.
2:58:56
I'm glad you're doing what you're doing and uh you are new you're our new uh healthcare uh expert and correspondent.
2:59:03
So, expect a call next to LeBron James.
2:59:05
I'm And LeBron James specifically. LeBron James of EMRS. Yes. All right.
2:59:09
According congratulations on the milestone.
2:59:13
Uh hope hope to have you on again soon. We'll talk to you soon. Okay. Cheers. Have a good one.
2:59:18
Uh should we do some timeline? Fun show. Fun show.
2:59:20
Yeah, we definitely should.
2:59:22
We got to talk about uh Sam Lesson's Oracle versus uh Salesforce.
2:59:27
He's getting in He's getting in hot water. You got in hot water.
2:59:29
The timeline's in turmoil.
2:59:31
We love Sam Lesson on this show.
2:59:33
He he posted a uh screenshot.
2:59:35
He says I will defend big tech.
2:59:37
I will defend Sam Sam Lesson uh Oracle is 2x Salesforce but Ellison is worth 25x benny off what this sale says about the limitations of the SAS business model.
2:59:47
He said uh he had a fun riff yesterday with the slow partners on this.
2:59:52
Oracle is obviously crushing it.
2:59:53
But if you take a today snapshot basically the market cap of Oracle is 2x Salesforce 500 billion versus 250 billion.
3:00:00
Meanwhile, according to previously directional at best uh data, Beni off's net worth is 125th that of Larry Ellison's 10 billion versus 250 billion.
3:00:09
What do you learn from that?
3:00:11
What lessons do you draw? I like this. The revealed preference.
3:00:12
Um for founders and companies, the old licensing model is better than SAS. That's interesting.
3:00:18
Imagine having hot take 10 billion and just getting little bro by Larry.
3:00:26
He has a sort of little little broing effect on most people.
3:00:28
effect on most people. There's an amazing story about a little bro, a famous little broing where Phil Knight of Nike was worth something like 10 on the order of like 10 billion dollars and he was in like maybe Sun Valley or something going to a movie and he runs into Bill Gates and Warren Buffett who
3:00:44
are just going out to a movie and they're they're both worth 10 times him and he's just like yeah I had this weird awkward moment where I was like nervous to meet them for the first time in a long time because typically he's like the most successful businessman he runs into all day, Right. But he was just
3:00:57
But he was just like, yeah, uh he in his book Shoe Dog, it's a fantastic book.
3:01:02
Shoe Dog is a great is a great book. Great book.
3:01:05
Uh and he talks about like all the weird effects of like having immense wealth, how like his wife would like hoard immense amounts of like uh paper towels just because they were like money is no object. Like what should we do? Paper towels.
3:01:20
What should we do with this?
3:01:20
And I got a lot of paper towels and they had to like figure out okay this is like some weird psychological thing that's going on in my brain.
3:01:26
like I don't actually need all these paper towels.
3:01:29
The fact that money is no issue doesn't really matter. Yeah.
3:01:30
People like to talk about, you know, you're the you're the you're the what whatever the the average of your five friends.
3:01:36
And it's like, yeah, well, if you want there there should be some similar law of like like your growth rate is like should be is like tied to how often you're little broad, you know? Yeah. Yeah. Never get little broad. No. No. You want to be Oh, yeah. Yeah. That's true. Yes.
3:01:51
You can be on the upward swing. That's right.
3:01:53
If you're not getting little broad enough, you're not on an upwards upward trajectory. This is good. Yeah. Yeah. This is good.
3:02:00
I've been in that situation before.
3:02:01
Anyways, um we could cover what Sam said.
3:02:03
Uh but I think we can just skip to We're also going to have him back on the show.
3:02:09
We're going to have him back on the show. He's a regular.
3:02:11
We're going to skip to Miles. He says, "Wrong take.
3:02:13
Allison is much richer because he didn't sell shares and has steadily been buying back 2% of the company every year for 30 years.
3:02:18
He's increased his ownership from 17% to 40%.
3:02:24
Such an incredible story.
3:02:24
Founders complain about dilution. Yeah. Oh, you got diluted. Yeah. Oh, I'm sorry.
3:02:28
Why don't you just buy back shares every single year for decades? Yeah.
3:02:33
Um, if Ellison, you know, keeps doing this, he could very well own 150% of his company at some point.
3:02:41
That's the future for OpenAI.
3:02:44
OpenAI just becomes the the agentic organization.
3:02:46
It just it just buys back so many shares that it eventually owns itself.
3:02:50
Yeah, that's the real That's the real goal.
3:02:52
Any uh anyways Miles says meanwhile uh CRM aka Salesforce made a lot of dilutive acquisitions and Beni offly sells his shares yearly.
3:03:02
He doesn't sell them yearly.
3:03:03
He sells daily 2 million daily bu oh liquidity events you know they're they're few and far between you know daily not happening every day.
3:03:13
Not for bounty off daily liquidity. It's pretty good.
3:03:15
Um, you know, the the real uh, you know, another How much does he pay Matthew McConnA to just Oh, yeah. That's got to be pricey.
3:03:24
I think it's only like 10 million a year.
3:03:26
So, it's like, yeah, couple Super Bowl ads. Not bad.
3:03:29
Uh, so Sam responds to the the hate.
3:03:32
He says, "Since a lot of folks are making the same comment about buybacks versus sales strategies, that is at best the noob answer.
3:03:38
If you're smart, you understand why they have different paths.
3:03:44
And the answer is path dependency from business model quality. Take a 2011 level class.
3:03:48
And then Buho Capital Blow quote tweets that and says it's a timeline in turmoil. Wrong again. CRM has executed poorly.
3:03:58
They've diluted shareholders with bad acquisitions.
3:04:00
They have 75,000 employees who they give excessive stockbased compensation to.
3:04:04
They let hubs scale up in their face. HubSpot.
3:04:07
They've diluted versus shrunk their share count uh versus the other companies Adobe that eat shares.
3:04:13
Uh investors don't trust him.
3:04:16
If Beni off held and cared about shareholders, it would be a closer call. He doesn't care.
3:04:20
It's not about the business model.
3:04:22
Well, you'll love to see some some timeline and turmoil. Very very fun.
3:04:26
Uh in other news, Shil Monot has the story about Telegram's founder Pavle Durov.
3:04:31
Consistent feature on the Techbro Drip account.
3:04:35
Everybody says they're pronatalists until they until you ask how many how many uh children, you know, have you fathered through sperm donation.
3:04:44
Apparently, he has fathered over 100 kids via sperm donation.
3:04:49
And he is worth $14 billion.
3:04:51
And he says he'll leave his fortune to all of them with no difference between his six kids conceived naturally versus via uh compared to the hundred via sperm donation.
3:05:01
So, every one of them is going to get $140 million.
3:05:07
just to kick off fundraising, just start investing that.
3:05:09
You got your family office on day one if you're one of Pavle Durov's kids. Pretty remarkable. Yeah. Single LP.
3:05:16
It's kind of a good dynamic. Yeah.
3:05:18
I wonder how he's going to get liquidity for for Telegram at some point because you get a bunch of Telegram shares.
3:05:23
It's kind of like this this difficult beast to wrangle.
3:05:25
Uh, but I mean, I guess you take it public and at some point and and get liquidity out of that. I don't know.
3:05:31
I mean, it also just prints money.
3:05:33
So even if it's like 14 billion, like you could just get like get a stake in the in the distributions because it's making money.
3:05:40
I think he kind of figured out life and wanted to make his life basically a hundred times more complicated.
3:05:50
Uh you know, having having this type of dynamic, you know, not just with his many uh children that that he helped conceive directly versus the hundred others.
3:06:00
So, um, had to one up Elon. Had a little bro Elon.
3:06:03
And little and Elon's commented on this, too.
3:06:06
He was like, "A I got rookie numbers.
3:06:08
Gang Gangaskhan over there is really taking over the world."
3:06:12
Uh, Gangghask Khan of encrypted messaging. Yep. It's very, very odd.
3:06:14
Uh, only CFO says the finance department outrinks sales.
3:06:19
Feels like sales is inviting finance to the party so they can stick them with the bill.
3:06:23
And this is this is the data you can only get from ramp. ramp. comdata.
3:06:29
Apparently, apparently, uh, finance, marketing, sales teams lead in alcohol spend.
3:06:33
Alcohol as a share of business meal spend. Not in that order.
3:06:35
So, marketing is absolutely dominating.
3:06:38
Dominating 20% 20 19% of all spend on alcohol.
3:06:43
No, no, it's 19% of business meals are alcohol.
3:06:46
So, if they go out and they're getting $80 worth of food, they're adding on $19 or something or $81 of food, $19 of drinks. That's the idea. Alcohol share finance.
3:06:58
They're getting, you know, $84 of food, $16 of booze.
3:07:07
Marketing is drinking sales under the table. Yes. Yes. Narrative violations.
3:07:12
It in the in the the tail end there, 9. 7%.
3:07:15
Uh many Huberman devotees in the IT department apparently. Yeah.
3:07:19
Not a not a power lunch uh you know, category.
3:07:26
No, but the three martini lunch will make it back for the for the tech teams.
3:07:30
Uh, should we go to this story about the vibe coder who sold his business to Wix for $80 million?
3:07:35
It's only a six-month old company and there's no external funding.
3:07:39
$189 $198,000 $189,000 in profit in May.
3:07:44
Uh, Bryce Roberts, this is just the beginning.
3:07:49
There's going to be more stuff like this.
3:07:50
I think this is pretty cool.
3:07:50
Base 44 only employs six people.
3:07:51
hasn't raised any any external funding.
3:07:53
The 31-year-old built a viral app AI app maker as a side project.
3:07:59
So, you go in there, you design an app, obviously um uh plays very well with Wix, which is in the uh website building business, but he flipped it for $80 million and he's post economic now. Congratulations.
3:08:13
Yeah, when I saw this headline, I was I was confused.
3:08:16
I was like, "Okay, so he just vibe coded some something and sold it."
3:08:20
But it is a tool to built a vibe coding tool trusted by over 250,000 builders worldwide. And nice quick flip. Amazing.
3:08:30
Um he he's basically getting a similar uh similar outcome to you know a a founder that sells their company for a billion dollars but you know goes through a bunch of different financing or kind of like a midtier AI researcher these days. Yeah.
3:08:46
Starting you know like starting out.
3:08:49
Yeah, starting starting out.
3:08:50
Uh, in other news, John Carmarmac is absolutely jacked. This is fantastic news. Uh, Yaxine has the news.
3:08:58
He's looking he's looking very built.
3:09:00
Uh, but John Carmarmac chimes in.
3:09:02
He says, "A chunk of this is just uh his wife dressing me in tighter shirts, but I did put on several pounds of muscle this year after switching my random grab bag of vitamins and supplements over to Brian Johnson's blueprint system."
3:09:13
Let's hear for Brian Johnson.
3:09:15
Really making a difference in the technology world.
3:09:17
Uh, I was probably not getting enough protein to take advantage of the exercise I was doing.
3:09:21
I have always been roughly upper quintile for fitness. Let's go.
3:09:24
Uh, regular exercise, but not at the level of serious athletes that most offices tend to have a few of, and now he's looking built. Palmer Lucky chimed in.
3:09:33
Uh, it's a great great day on the timeline.
3:09:36
Uh, let's check in with Tyler. Close out the show.
3:09:40
I was going to check in with this poly market.
3:09:42
Will Chimoth launch a spa in 2020? It is up to 70%. It's up to 70%.
3:09:48
It was 33% when I posted it this morning. Wow. That's big news.
3:09:53
It was partially because he came out and he said, what did he say?
3:09:58
He said 58 He asked yesterday, should I launch us back? 58,000 people voted. 71% said no.
3:10:06
He said, uh, I hope everyone that voted no feels seen. Now, on to business.
3:10:13
I got calls from many Wall Street and crypto titans yesterday.
3:10:18
They all want in and their vote matters a lot to me.
3:10:19
So, I will probably do it. I love it.
3:10:22
Maybe this time it will go better. Who knows?
3:10:24
The risks are clear, though.
3:10:26
The last time wasn't a success by any means.
3:10:27
I will include this poll and the community note in every SEC filing possible.
3:10:31
It will make an excellent disclosure about the risks and is not short of irony.
3:10:35
So, what kind of company do you guys want? No crying in the casino. So, let's go.
3:10:43
People are absolutely fuming in the comments, I'm sure.
3:10:47
But honestly, get after it. Pretty fair.
3:10:52
Everyone knows he's going to play by the rules, you know?
3:10:56
And I think I think at the end of the day, it's it's very like I I look at the next Chamath as like it it totally will like it probably will pop.
3:11:07
It'll probably it'll get a lot of attention.
3:11:09
It might turn into a meme stock, right?
3:11:11
Um, I uh I will be interested to see what kind of target he I'm 100% excited to follow the story.
3:11:20
Yeah, it's going to be fascinating.
3:11:21
Anyway, let's check in with Tyler and then close out the show.
3:11:25
Tyler, I have a question for you.
3:11:25
Can you guess a number between 1 and 50? Like a random number?
3:11:31
Yeah, random number between 1 and 50. Uh, 27. 27. Are you an LLM? Did you see this?
3:11:37
Every single model, they all guess 27.
3:11:40
When you ask them between a number between 1 and 50, uh, Chad GBT, Claude, Perplexity, Meta, they all guess 27. We got to get a world.
3:11:50
We got to get a world coin orb in here to be able to prove that Tyler's not in fact Yes, we do.
3:11:55
We He might just be a DP. Uh, final review.
3:11:58
What did you get done this show?
3:12:01
Uh, did you keep playing with Mid Journey?
3:12:03
Were you doing something else?
3:12:04
What's been going on the last couple hours? Uh, yeah.
3:12:05
I think I I just sent uh another video. Oh, no.
3:12:08
Uh, I've been pretty productive. You can watch this. Let's see. Gorilla.
3:12:15
I like that the gorilla crosses behind. What's he doing?
3:12:17
Did he just How he just took my spot? Oh, he took your spot. Wow. Wow. Slides in.
3:12:21
Oh, he comes in with a paper. Breaking news.
3:12:24
That's the breaking news gorilla. The breaking news.
3:12:28
We need to get you a a breaking news gorilla outfit and and if he has breaking news, you can print it out. Come sit down. Take a seat. That'd be amazing. That's good. That's good.
3:12:35
Uh are there any others or we're closing it out?
3:12:39
Uh yeah, I think that's it. That's it. That's it. Well, good work. Good work today.
3:12:44
I feel like the productive team laughing like they have some other ones that are too scary for for our audience.
3:12:49
I saw I I saw one get sent in the chat and it and it just was really scary, bad looking.
3:12:55
Well, we will be back tomorrow.
3:12:57
Uh we have a great show for you folks.
3:12:59
Leave us five stars in Apple Podcast and Spotify and thank you for watching.
3:13:03
Thanks for being here with us. Fantastic show. Have a great evening.