Anthropic, Glean & OpenRouter: How AI Moats Are Built with Deedy Das of Menlo Ventures

I v- entered Venture and I'm like, "That is the company I wanna build." From 2019, I remember going to parties in the Bay Area, and I would say and just shutting down the conversation right there. Anthropic is the fastest-growing software company of all time.

When we invested in the company, uh, it had no revenue. In India, academics holds the, the same sort of prominence as sport would hold in America. On average, people are quite poor, so education is seen as the means to social mobility.

The way it works is similar to countries like China or some other countries where you take a big exam, you get ranked. Um, a million people take the core engineering exam, and the top 10,000 get in, and the top 200 get into computer science.

That's how hard it is. Those top 10,000 get into IIT. Everyone's heard of that. M- That's, like, where a lot of the great, you know, Silicon Valley people from Sundar to, to many other f- people come from. You look at a guy like Rahul Patel, who's become the CTO of Anthropic, and he's not from a top university in India, and he sort of worked his way up to a position of such prominence.

It's testament to the fact that even though you didn't have the opportunities early, and even though you might not believe you could do it, if you work hard enough for a long time on things you care about, anything can happen.

Hey, everyone, welcome to the Latent Space Podcast. This is Alessio, founder of Kernel Labs, and I'm joined by Swyx, editor of Latent Space.

Hello, hello, and today we're finally joined by the epic return of Deedy Das. Uh, welcome back.

Thank you for having me, guys, again.

Yeah.

I'm so glad to see you.

Uh-

All of us have different jobs now, I think.

All, all, all different jobs. All different jobs. Classic Bay Area, you know. It's been two years, right? So last time it was April 2023 you joined us, uh, remote, and you were still at Glean back then.

Mm-hmm.

I was actually even also looking at the Claude timeline. Uh, so Claude 1 was March 2023, and Claude 2 was July 2023. It just feels like so long ago.

Man, I remember the time when, I don't know when, what your first experience using Claude was, but, but mine was, I remember early Glean there was, uh, somebody from the company was like, "Hey, there's this interesting new LLM that's not OpenAI, and the only way you can talk to it is by tagging Claude in a Slack channel."

Mm.

And I'm like, "That's a bizarre interaction model for, uh, a whole new product."

It's the best model.

And, uh, and now fast-forward to now, and I'm like, "Okay."

Yeah.

It's-- we've come, we've come quite a way.

Yeah. I think actually Cl- they only recently introduced Claude in Slack, right? Or, like, publicly.

Comeback, The comeback.

Yeah, yeah, yeah.

It's like how it started, and now Claude is back in Slack.

Yeah, ChatGPT in Slack, Claude in Slack. And so since then, uh, I wanted to start with Glean, obviously, 'cause of, you know, we, uh, we're gonna cover a lot of startups in this episode. So Glean has-- Glean was like a billion dollars, I think, as based on my research, and now it's at $7 billion, so your, your, your options are good.

What's your take on, like, how Glean's going and the market in general?

I would say that now being on venture side, I have a, a bit of a, a different take than I would've had at Glean. But broadly, one of the things that I love about Glean is it's such a boring, unsexy company that became sexy later.

Mm.

So from 2019, I remember going to parties in the Bay Area, and I would say enterprise search, and it's shutting down the conversation right there. You know, like, nobody would ever ask a counter question if you said enterprise search.

They're like, "Oh, that sounds boring as hell. Leave me alone." Like, um, and, and fast-forward to 2022, enterprise search gets more, um, got more conversations. It was like, "Interesting. Tell me how you're doing this-

Yeah

... this search." I think what, what was nice about that observation is in those three years, we did a lot of work and not, didn't take shortcuts on a lot of things that ended up generating a lot of value for us now.

And I can go into what, what all of those things are, but if you look at Glean from a high-level business, it is top-down enterprise sales. It's very hard to rip and replace. We have-- We expand contracts very easily because the TAM is so large.

It's every knowledge worker could use a version of enterprise search, and then the AI on top, I still call it search, but information retrieval in the enterprise. And we've, we've, we solved a lot of critical problems, I, I can go into that too, in order to get there.

Then comes, you know, December t- 2022, the ChatGPT moment, everything that's happened since, and now when I look at Glean, you know, it's a different world. We were very quick and correctly prioritized LLMs earlier on. It did a lot of good for our business and the company.

But now the, there's fire from a lot of angles. Like, everyone wants to be a part of the enterprise search story, and it makes sense. I mean, it's a large, unconstrained TAM. LLMs are particularly useful for gathering information.

Obviously, consumers are interesting, and enterprises therefore are interesting. How do you do this in an enterprise? Well, gather all the knowledge and then put an LLM on top. So that being said, I'm still very happy with the Glean stock.

You know, Glean's also valued at $7 billion, not $100 billion. So I'm-- I, I think the company has a lot of growth. I think it's done a lot of the hard work that nobody's willing to do. And I also think, you know, VCs have a tendency, including myself now, to, to trivialize a problem into a one-sentence sort of narrative.

And with Glean, that narrative was often, "Oh, well, you guys built this enterprise search thing which never worked, and then AI came along, and it started becoming a thing," which I think is not the, the story at all.

I, I really think we did all the hard work to build search, and AI happened to accelerate our go-to-market motion at the right time. And now I see companies trying to tack on search. It's not easy. I know the kind of, like, last mile stuff we did for some of our customers, and I just know that when I think about other companies, I'm like, "Would you really go all that distance?"

It's not a moat. The moat is just we did the hard work. And so I'm pretty happy. I mean, things can go any direction, but I'm pretty happy with the, the way Glean's going right now.

And just to spell out the two main challenges. So one is obviously Claude, I think today, launched enterprise search, um-

I was gonna say-

... connectors

... I have the screenshot.

Um-

Did you see, like, "Hey, we're introducing enterprise search"? I'm like, yep.

Son of a gun. Um, and then the other side, you have the data providers adding these rate limits, kinda like Salesforce has done with Slack. It feels like that part is more challenging than, like, the competition-

Right

... from other companies. Like-

Right

... yeah, any-- How do you think about that?

Two questions, I guess, competition and the rate limits. Uh, on the rate limiting side, it's happened for, for several, uh, of the, of the SaaS tools. I think one advantage that Glean has is... Well, the first thing, let me address the premise of the argument.

When I think about why SaaS tools would limit API access, inherently it never made sense to me. I can see why you do it for business reasons. Maybe you wanna launch a competing product. But Glean doesn't eat into your revenue.

If you are Slack and you've sold, call it 100 seats at a company, and g- you have Glean at that company, if anything, Glean is-- only shows Slack results to the 100 seats that you've sold. So we aren't eating into your business.

So from primary first principle as business logic, I don't see why you'd do it. If Glean is on Slack and more people are searching through Slack, it actually lets you, uh, sell more seats, not less, 'cause we don't reveal permissions to people who don't have access.

If we were to do that, then I could see maybe a business case. Like, oh, you're taking the Slack data that I've only sold one license for, and you're showing it to 1,000 people. That's problematic. But we're only showing it to the licenses that you've sold.

So firstly, that's my first point. The second thing is, uh, we do have thousands of integrations, and in a lot of enterprise customers, Slack is, it's, is really important, and that's, that's, that's a critical data source. But we also have many-

Teams

... many more. And so team, you know, um, it's just, you know, the law of large numbers. So maybe if everyone decides to shut it down, it could be more problematic, but if one person does, then, um, you know, less.

And the third thing I'll say is, if you talk to the customers, they're also super unhappy about this 'cause they're like, "Look, we bought your product. We own the data. You don't own the data. And so if we wanna to buy another product to use our data in Slack, why can't we do that?

Why are you blocking the API?" So those are the three prongs of the argument. I can't-- I don't know how this will all-

Yeah, yeah, of course

... end up, but, uh, I don't think it's that sensible that it is, uh, like this. And, and I'm still optimistic that we'll clear out some of those issues.

Yeah. Um, anything else you wanna say? So like, like, you know, obviously we're about to move to Anthropic, and Anthropic just launched enterprise search. And so what would you say, as a veteran of enterprise search, that Anthropic should like, you know, take, uh, take note?

The question of, of the labs competing with Glean has always been a thing since-

Yeah

... 2022. Sam Altman, like we were just discussing earlier, like Sam Altman once came out and said, "If you're an investor in OpenAI and one of these five companies, including Glean, we don't want you as an investor" or something to that tune.

Yeah, and it's, that's just facts, yeah.

And, um, but I-- yet, here's what I see. Look at the revenue of Anthropic and OpenAI right now. These are billion-dollar revenue scale businesses. Glean is several hundred million dollar revenue scale business. So the way I think about, and this can even allude to, like, how I think about startups' right to compete and right to win, is for Anthropic and OpenAI to build a deep enterprise search system, it doesn't make them that much money.

They have to put all this effort to make, what, an incremental 100K sell, 200K sell maybe, even a seven-figure sell. Is that moving the needle on your, you know, five-plus billion dollars in revenue, or 10-plus in, to the end, the end of the year for OpenAI?

Not really. And the amount of effort it takes to get there is big sales teams, huge FDE teams, tons and tons of customization. And my question is like, in the long, long run, you could build a semi-reasonable enterprise search tool.

If you really wanna go deep, I don't think you will ever dedicate the people to do it. And the last thing I'll say is, you think about it from an Anthropic engineer's perspective. You joined a-

Right

... big AI lab to work on models, not to build Google Drive connectors, right? Like-

There's a meme like, you know, like- ... build the fucking integrations.

Build the integrations. I think, I think I'm, I'm still very bullish. Um, but yeah, competition happens, so.

Yeah, yeah. Uh, it, it's, actually, I wasn't asking about competition. It was just more about, uh, what are the hard problems that people don't appreciate.

Oh, okay. Um, we can talk about all the-

That, that was probably safer category for you. You know, it's just-- but basically, like, uh, you know, I-- and I'm in this boat as well. I've joined an enterprise AI company that has to worry about and, and build for these issues, and I'll just give you one very example.

Until this point, we never had to deal with two Slacks. Like, and like-

Ah

... and, and enterprise has like, you know, when you, when you acquire another company, you have different systems.

Yep.

And they all duplicate, and they all overlap.

Yep. Oh, man. I have some great stories about like-

So De- Devin had, you know, like I, I'm sure, like, there's like some pro user version of this, but I still haven't figured out how to use Devin properly with two Slacks.

Wow.

Because Devin's optimized for one Slack.

That's funny. That's funny. So Slack workspaces w- like, that reminds me, that was a thing that we had to address at Glean. I think like every enterprise company has, like, the same sort of hurdles-

No, no, no, yeah

... but when do they get there?

We, we looked at each other, we're like, "Oh yeah, we're a real enterprise now. We have two, two of everything." Like

That's funny. Okay, Glean, bunch of interesting problems. I'll, I'll talk about some of them. If you wanna prod, feel free. I think number one most interesting to me when I joined the company was consumer search was largely regarded to be a solved problem.

Not really, but largely. The way most consumer search systems work is by aggregating feedback data on how users use search, whether they click, hover, how long they stay on a website, and that's what powers ranking systems to get better over time.

Very, very powerful, critical way of how, like, Google, Bing, and all the above work. In enterprise, if you take a 10,000-person company Even if every user issues two search queries a day, which is quite, quite a lot, say even five, I don't know, that's just not enough volume to have any meaningful quantity of feedback for this to be relevant.

On top of that, add to the fact that freshness is way more critical in the enterprise in certain ways than it is in-- There are more freshness-seeking queries in enterprise than there are in consumer. And then number two is the distribution of queries in consumer is very head heavy.

It's not in enterprise. In enterprise, maybe the query that everyone wants to search for is benefits or payroll. It's just, like, not that useful. Really, like, it's every person's doing a job, and they have different needs, and they have different things they wanna look up.

So given all of that, the techniques behind the hood, under the hood that work for consumer, they don't translate to enterprise. You have to invent a whole new set of signals that actually makes enterprise search work. And evaluation becomes very, very difficult, too.

On consumers, you have tons of data to pick and choose how you wanna evaluate what's the right result to show for this query. In enterprise, and I have this story a lot, like, we look at some of our customers' data, and we would look at each other and go like, "We don't really understand what this query means.

We don't really understand what these results are. We don't know what is the right ranking or not. We have actually no idea what we're doing here." So... And which happens. Like, it's so out of domain for, for even us.

Some of our customers are working on very, very specific problems. And so all of those, that's one huge, huge challenge. How do you make ranking work in enterprise, you know, in a, in a great way? There's many. I'll touch on the more second interesting one.

Second interesting one is selling productivity tools to enterprises are challenging because as no matter what ROI argument you make, people aren't actually buying tools for ROI. People buy productivity tools because their users like using them.

Mm.

So for example, when people buy Slack, I don't think any buyer is going like, "Let's measure how much faster our, how much more productive our, our team is getting by using Slack." It's probably not even getting that much more productive.

That's not what they're looking at. They're kind of saying, "Everyone uses Slack. It's pretty useful. I'm gonna keep Slack. I don't think we're gonna change that one." If you take that analogy to search and, and search systems, the issue is search systems aren't inherently viral or growthy.

Slack has a very clear virality moment. Like, everyone's talking to everybody else, and so that's just how you have to speak. In search, it's kind of a one-player game. You're not really sharing things. You're not really talking to everybody else.

So the challenge for us was, like, how do you get... sell a productivity tool by getting everyone to love this on day one? For a product like Search, it's not easy. If you look at how Google did it, they had Chrome.

So great, like, have a great source of sense of distribution, get everyone to, like, query, and then they'll learn to love it, hopefully. So we had to figure out what that meant in the enterprise as well, and how to get everyone to, like, adopt and embrace and love this new tool.

Yeah.

So two-

Makes sense

... two of the many.

Good pointers.

Yeah, just a question on that.

Yeah.

Was there any-- Because, you know, oh, you have a new search tool, it's like, go search. And it's like, what am I searching, you know? Like, what was that blank canvas onboarding for people? Anything good?

Um, several different things worked well for us. Uh, I can think of two at the moment, but I'm sure there were many, many more. Um, I'll say one of them was, say, for, for a handful of companies, like many companies actually, we would say, "We wanna take over your new tab page."

Mm.

And then the critical part was, "Tell us what we need to do to earn the right to do that." No one wants to give away their new tab page. So, so, so we went the last mile, and there were companies who were like, "Well, we have a new tab page.

We're pretty happy with it."

Mm.

So we'd ask, "Do you have a search bar on it?" They'd be like, "Well, yes." I'm like, "Okay, what is, what is that using?" And they'd be like, "Well, it's using our internal thing." I'm like, "Do you like it?

Clearly not. That's why you're-"

Right

... "talking to us, so let's just rip and replace that." But doing that extra mile was pretty important. So that's one, new tab. The second one that we liked was, uh, Chrome extension, and then doing the-- I forget what we called this, but when you were on your native product and you were issuing a search query, we'd ran a lot of evals, and we thought we were better at every product at their own search.

So if you were searching on Google Drive, we will do a Glean-

Mm

... replace of the search bar and the page pretty natively, and, uh, it would teach people, it would teach people to use Glean, and be like, "Okay, this is pretty useful. I think these results are great." And it automatically filters to Google Drive anyway, so you know, functionality is not lost.

And, uh, we would slowly get people to be into the ecosystem that way.

Yeah, superset adoption. Uh, something that OpenRouter also does. Okay, so Anthropic, we have to obviously address the elephant in the room. You guys are huge, huge Anthropic investors. I think right after you maybe got promoted or you became a partner, you, you guys led the D?

What, what's the chronology that-

I think we did part of the C, and then the D, and then every single round we had more than Perata.

Yeah.

Yeah.

Obviously, one of the greatest companies in, in AI. I honestly had no idea that it would, like, we would be sitting here... Like, Anthropic has 10X'd in the time that you've been at Menlo. And it, like, I just wo- what's it like being an Anthropic investor?

What, what do you think about, what, what are the considerations back then versus now?

Anthropic is the fastest growing software company of all time. I think I can say that fairly. Um, I haven't been disproven yet. So I think the-

People say that, but, like, everyone says that. Like, you know, we're, like, first to, like, one billion, first to 100 million. I don't know. It's, it's, it's hard to tell, but-

I do believe the, the numbers are zero to 100 in one year, 100 to a billion in one year, and this year it would be one to the projection that is public is nine. But even, even to this point, like, I know a lot of people, we've seen the graphs on Twitter.

I, I, and a lot of some of that is bullshit, some of that is GMV, some, all this other stuff. But in Anthropic's case, I think it's, like, fairly legit revenue, and I do think it makes it the fastest.

Definitely at, like, the one billion plus scale, I can't think of too many examples. So clearly it has outdone itself. I would say that when we invested in the company, uh, it had no revenue. I mean, that, that, that's just fact.

So when we wrote our first investment, it had no revenue. It was a-

$18 billion Or f- 4 billion. 4 billion

Right. It's been fascinating to see this company succeed. I, I j- I couldn't have predicted it. We-- all of us, this was beyond our wildest expectations. I think whether or not it continues to perform at, at, at this rate, I believe it will, but it is already somewhat of a generational company in, in, in many ways.

And so it's kudos to the team to deliver, like, this, these, these, these awesome results. You know, one of the risks, I would say, like kind of taking a tangent, one of the risks with a company like Anthropic is you essentially had a team of extremely idealistic researchers.

And very often, you know, the, the standard deviation of outcomes when you have teams like that or similar to that are... is, is quite large. There was a world where maybe they would've not worked at all and would've absolutely fizzled to the ground.

But I think it is the same qualities that would make them have a high propensity to fail, made them had a high propensity to, to succeed. And if you look at, there's many other things they did right, but if you just look at a product like Claude Code, there's not many inno- product innovations in AI that I can think of that are so critical s- as something like that.

'Cause we had the whole chat era of RAG systems and ChatGPT. That was a critical innovation. But since then, there was a lot of followers, a lot of d- deep research, which is kind of, I would say, an addendum.

Couple of other things happening here and there, agents, cool. But you know, if you think about agents that actual c- end consumers use and gain value from, in, in my mind at least, Claude Code was the first time I saw that in a terminal, in a weird interface.

It was just weird. Like, it was like every PM's nightmare. No PM would've thought of that. And so it's such a-

Except for Kat Wu.

Yeah, yes, except for Kat Wu. And so, you know, it kind of gives... I, I... it goes to show how, um, Anthropic is able to function as a company to be able to innovate like that, which, uh, is, is, is quite rare, especially for that scale.

To some extent, I think you just, like, hire good talent and then, like, let them loose with a lot of tokens, see what they come up with. They, they, they tend to build good stuff.

Well, like, it's interesting to talk about, right? Like, 'cause take OpenAI and DeepMind as a comparison point. Like, I think we'd all agree they all have great talent.

Mm.

But they all don't innovate the same way.

Mm.

And it's always been interesting, like, just as a academic exercise to, to think about, like, different leadership styles. And maybe from the outside looking in, you'd be surprised how little I actually know from an investor standpoint about how Anthropic actually operates.

But it seems like it's a company that has, you know, such high retention numbers on employees because they are very free-spirited in how they let the employees guide the direction of the product, versus other companies which are much more either top-down or prescriptive or like, "Hey, we need to go after this, and we need to go after that."

It's like, "Hey, let's see. Let's see what happens."

Yeah.

"Try."

Yeah, I think, um, at my last conference, uh, SignalFire had some stats. They track all the l- uh-

Yeah

... LinkedIn pages of everyone, and, like, Anthropic has, like, the best retention, and, like, the-- it's like an, a net gainer, whereas every- everyone else is like a net donor of employees to, to Anthropic or something like that.

I'm referring to the exact same, uh, article, where I think their retention, one-year retention on employees is 80%, which in AI world is, is quite wild.

Yeah. I mean, Anthropic does not have image generation. They do not have a IMO goal-winning model. I feel like they don't-- they just do their own thing.

Yeah.

They do it great.

They have nice hats.

Yeah.

They have thinking caps.

They sell out-

Thinking caps.

So a- actually, I really wanna discuss this, but I don't know how to... I think I, I need to get, like, some, like, marketing PR agency person because, uh, people actually forget 2024 they had out-of-home advertising campaigns, which sucked.

Everyone was, like, dog piling on them. And then this year it's, it's, like, slightly changed. It's still Anthropic butthole, but, like, slight- slightly, like-- and but they just, they decided to focus on thinking, and, like, suddenly everyone loves them.

Uh, and they have, like, the cafes and all that. Like, it's, it's a very interesting public image rebrand, and I don't know if it's because the models are just better or it was actually, like, PR. Like, which one comes first, like, chicken or egg?

Like, models or PR?

It's a good question.

Yeah.

It's a good question. I would say, though, like, i- ignoring the model side, like, w- like, I do think this one is, like, aesthetically better.

Yeah. Yeah, yeah.

Purely. Like, purely on aesthetics-

Just, just it looks nicer

... uh, yeah, and, and the vibes, and I don't know.

It, it's very-

It's hard to-

Like, I have sat in those meetings, and it's like, like, someone's pitching you an idea, and you're like, "I don't... It looks good. Okay." And then, like, it becomes the, one of the most hated campaigns of all time.

And then one year later, someone else comes with, like, a slightly different looking idea, and it's like four... the words are, like, different in, like, four ways. Like, they, they, they chose, like, slightly different words, but it's not that many words, and suddenly that one is the one that works.

Yeah.

It's, it's It's-

Well, as, as somebody who, like, writes online a lot, I can relate to, like, a couple of things different can be the difference between something people care about and not. Yeah, early, like, in Glean, we had, I had such run-ins with marketing 'cause the first campaign we d- we actually did this campaign.

I was just like, "Really? AI for work that works." I'm like, "Okay," like-

Was that a hit?

No, I mean-

Oh, okay. I didn't know

... in enterprise, like, how does one even measure what is a hit, what is not?

Right.

I mean, no one really cares enough, I feel one way or the other. Um, but yeah, you-- we've all seen, like, really cringe AI ads. If you've seen the Cisco ad in the airport-

Yeah

... I hated that one for a while.

Yeah.

All kind of generic. So I like the... Anyway, I like the Anthropic one, but-

Ah, okay. I, I'm gonna sprinkle in some of your tweets. So you had, you had one ad about the, the billboard where, uh, the, the Reddit guy was like, "My boss really wants you to know that we're an AI company."

I thought that was the single most honest billboard I've seen in San Francisco.

It absolutely. I think the... it's, like, the, the testament to all the comments of people going like, "Yeah, I relate." Um, I mean, we've all heard it. Like, everyone, it feels like even on the technical side, people are struggling to catch up Gain a sense of meaning again.

I've had developers go like, "Fuck, man," like, "Is this it?" Like, "What do I do anymore?" And even that's happening on the technical side of people who semi understand what's going on. On the non-technical side, people are like, "So there's this new thing, it's AI, and generally my boss literally just wants me to do something in it, and I don't-"

Something's wrong with AI.

"... really understand."

Yeah.

Other than ChatGPT is quite helpful.

Yeah. I have some charts. Uh, I don't know if you like, uh, have any of these like in, in mind, but I, I'm just gonna sort of bring up some of the Anthropic charts. So I think it's just...

I wanna just put it on the record for people who are not paying attention to, to understand. In 2023, according to... These are Menlo numbers, right? Uh, 2023 market share for OpenAI was 50%, and when, uh, mid-2025 you guys have OpenAI at 20% market share.

Anthropic was at 12, now at 32.

So-

That's like API, enterprise API market share.

Correct.

Just to-

So I, I should clarify that that is enterprise LLM API-

Right

... spend.

The, the market that Anthropic hap- happens to focus on. Yeah.

Um, and, and cr- critically it's also spend numbers, not token numbers. So I think those clarifications are, are important, and also the methodology is, uh, going and surveying, you know, vast amounts of, um, e- enterprise users on how they are, are doing their spend.

Yeah.

But that being said, yes, the point, the point-

The point is-

... remains

... uh, market share of OpenAI has gone down. It's not a negative. Obviously, OpenAI has done super well. It's just that diversity has gone up. Like it used to be there was basically only one choice, and now there's like three or four like legit fron- frontier labs, maybe more than that if, if you count like all the open models as well.

But, uh, I think it's just super interesting and, uh, u- under-discussed still that you can actually build like a sustainable, uh, advantage as a, as a frontier lab.

You know, I'm, I'm sure you guys remember, like there was a lot of conversation at some point about the commod- commoditization of models, and, uh, to an extent maybe it's happened. I mean, like models, a lot of the frontier models are neck and neck on a lot of things.

Um, but in practice, and this, this data was in that market map of that market survey as well, is that once people like something and they get used to it, they don't really churn off it once it fits their needs.

And so we've seen a lot of that. So there's a lot of like churn in hobbyist developer type category. But in terms of enterprises, often what'll happen is they'll buy up s- uh, large chunks of long-term compute and dedicated instances, in which case you just don't churn, right?

Like this is, this is what you use. So I think that's part of the effect. And, and, you know, to commend OpenAI, like OpenAI was just focused on something else, which is, you know, they have- they've launched the most incredible consumer product that we've seen since God knows when.

So, you know, so they were probably not focused on enterprise until- ... now again.

Yeah.

How do you re-underwrite the company internally as you invest? So I mean, even since we're talking about Claude Code, right? It's like I think that was like a pivotal moment in like the trajectory of Anthropic. What are the things that matter to you when you're like looking at a company like Anthropic?

Like does this market share number matter? Like how do you evaluate both the opportunity and like what are the numbers that you really care about versus like sure, higher market share, but like that's not what we cared about.

I don't think the market share number is-- The market share number is more imp- is, is more critical to understanding the TAM. At that stage, to be very honest with you, at the stage that we, we invested in Anthropic now-

Mm-hmm

... like the only things that would really move the needle on the decision is, uh, here's the revenue, here's the margin, and here's the trajectory, and here's the other markets we may be able to underwrite that they wanna go into, that they may be early in or planning on, on going into.

I, I think it's really difficult to underwrite on, on market share other than knowing what like the potential cap of the TAM might look like. So the pie will also expand potentially, but other than that, I don't think it's a...

It's, it's just like a, it's a nice vanity metric more than, uh, more than anything else.

Yeah. In your mind, is it kinda like, you know, people in crypto are always about the flippening of like Ethereum and Bitcoin.

Yeah.

Like is this something that matters? Like Anthropic can go to 50%, or is it OpenAI was only at 50% in a moment in time, which was a new market. Like yeah, I'm curious how you think about the-

I don't wanna color like the way Anthropic probably or the way all of us think about this, but I just don't think it matters that much. In my view, I'm a very paranoid person with startups and companies and technology, and so in my view I'm like, "Great, now let's make it last," or like, "Great, but what's next?"

And so to me, it's like nice to have. It's really not, um... I, I mean, look, if we're investing in a round right now, which is like north of 170 billion, sure, it matters. Some of the numbers matter.

But the future of the company is, is all the value is really in what we underwrite as the future, and the future means that I'm more concerned about what's happening next. What are the new models? How do you gain market share?

What has to be done? What are the new products that, that are going to be built? I'm less concerned about like where it's at right now in terms of market share. But that's just me. I don't wanna speak for others.

Yeah. I think the new models are, are really good. I mean, uh, Opus 4.1, Sonnet 4.5, Haiku 4.5, all, all released in the last few months. Uh, and, uh, it's re- it's, it really interesting. I think OpenAI and Gemini are in this sort of price war a little bit with the, the Pareto frontier that I, I track, uh, in terms of like LMSys versus, uh, the pricing.

And Claude can still charge a premium, but still like have a lot of market share obviously, and I think like that's just because they have a better model, and like people just n- naturally gravitate to it, especially for coding, but also other things.

And, um, I, I just think like articulating what makes a model good is just very, very difficult. Obviously, this is benchmarks and evals, and everyone has like, "Okay, today it's your turn to be best at Suitebench, and then like tomorrow's my turn."

Uh, but like it's, it's really stupid. Like we're, we're just like talking about like, you know, 0.12 differences in, in like Suitebench. But I wonder, you know, if you're talking about like, okay, I am investing $13 billion in Anthropic for s- series F to underwrite Claude 5, right?

What, what does it have to do? Like I-- what kind of, what kind of conversation does that look like? I, I have no idea. I'm not saying that you know, but I'm just like...

I would say that- Despite what you said about the premium, I think it's, everything you said is true. Um, I still do worry. I think cost is, is a concern for a lot of people, and so the Pareto, the Pareto frontier does still matter.

I'm glad Anthropic's where it's, where it's at right now, but who knows where that changes. When it comes to, like, Claude 5 and thinking about the future, one thing I think about actually that's really nice is I think we can take for granted right now that furthering the intelligence of models in ChatGPT, a consumer product, does not lead to more users or more retention.

It only is really applicable to a s- this thin slice of users who care about very smart type queries, right? And I would say maybe, like, under 10 million, right? Maybe that's just a random estimate, but most of the 800 million users on ChatGPT are asking, like, "How do I fix my dishwasher?"

Mm-hmm.

"How do I, like, like, rephrase this email that I've sent to somebody?" And that's done. Like, we know how to kinda do that. So what's interesting there is now that means we're at a point in consumer where m- maybe this is too early to say, but OpenAI has kind of won, right?

Like, how do you catch up to something where model quality is not gonna be a differentiator? You already have the users, you already have the retention, you already have great product and people are paying. But the, the interesting about Anthropic is if you look at coding, that's probably never gonna be the case.

Like, there is always an increasing frontier of how you g- good you could be at a task like that.

Yeah.

And we're nowhere close to that frontier, so it's more possible to underwrite the quality of the future models versus, like, an OpenAI where it wouldn't be as much of a revenue driver on their consumer business than as it would be for Anthropic.

Yeah. Talking about coding, let's, let's just, like, talk about it because I think, like, this is also a very fu- fun discussion. One, there is, like, the, the what are the margins of Claude Code, uh, which there are some numbers I, I, I, I don't want you to, to get yourself in trouble.

But then there's also, like, how do you think about the Claude wrappers, right? And, uh, there's-- we've, we've talked to Bolt and Lovable, but then also, like, I'll put Cognition and, and Cursor in there as well, right? Like, how do you think about this market of...

Like, basically there's a whole ecosystem of startups. They have all done really well built on top of Claude.

I think it's great. I mean, there's-

Is it sustainable? Is it-

I, I don't see why not. I mean, I don't... I, I kind of will allude to the margin question, which is, like, can, can Anthropic continue to do this strategy? Which, you know, I'm not gonna comment on the margins but, like, if you are trying to build out a enterprise-friendly business, there's, like, two broad approaches, right?

Like, high customization and high price, which is usually less scalable, uh, and then you have low customization, low price, which is very, very scalable. So and, I mean, in a SaaS world, I guess it's a Slack-Palantir continuum. And so this is kinda different, but generally Anthropic wants to play here, where scale fast, keep it cheap, get everybody on it.

If we trust that most people or a significant number of people will stay on Claude if they continue to build products on top of it, then I think that's a win for the ecosystem and it's a win for Anthropic.

I don't see why they would care. I think the interesting thing, and again, I don't know what Anthropic's future plans are, but, uh, like, you know, Ben Thompson obviously talks about this, is classic strategy, which is every time you own the, I guess, the means of production, you will end up getting into the markets that your users-

Mm.

-use you for. And so the classic Amazon example, which is, like, first you are the market where people sell. You find all the places that you can sell things that are commodity at high volume, and then you start creating batteries and Amazon-branded batteries, and then you push out a bunch of people who sell batteries.

So that, that's a risk, I think, for those companies that use Claude heavily and rely on Claude to think about. But at this point of time, we're too early. Like, I don't think Anthropic is anywhere near thinking about that 'cause you're still very much competing with other models on, on that layer.

Yeah, playing a different game.

Yeah.

Yeah. It's interesting, like, would you rather be an investor... This is basically model layer versus app layer. So far, model layer has won, and, uh, I think there's been-- there was a s- there was a kind of a app layer summer, and then now, now it's, like, very back to models again.

I mean- ... I, I like, I like the discussion. I like the discussion- ... because, uh, I, I was, I ha- I was at a dinner where we, where, like, somebody was talking about this kind of question, and I was thinking about it more at, just at that dinner.

And maybe this is a, this is an ill-formed thought, so, like, feel free to push back.

Yeah, we're riffing. Yeah.

But when I think about, like, moats to classic, like, VC startup banter, in my mind, I think the moat is what is the hardest to do in any part of the stack. And so when I think about people that w- like, tend to dismiss, there's other cons- like, aspects to it too, but people tend to dismiss like, "Oh, you know, the app layers will capture all the value."

Well, if the app layer is easier to build, I think the model layer is, is harder and therefore will naturally capture all the value net of competition from other model providers. So said a different way, it is far easier for Anthropic to try to go into one of the spaces of the apps than an app to try to go into the space of Anthropic, which makes me feel like one is more defensible than the other, um, all else equal.

So I think both can thrive, and that's ideally what everybody wants. But, um, yeah.

Yeah. I think very brutally as an investor and as a human with my own, like, ti- limited time on Earth, uh, you know, if, if Anthropic can go from three, uh, $4 billion to $183 billion in two years, uh, then e-everything, everything else is a waste of time.

You know what I mean? Like, so, like, uh, uh, I, I... You, you kind of, like, do want to, like, really get this right. You, you can't, you can't just be like, "Oh," like, "every-everyone's great," then, like, you, you know, and, and sort of hedge your bets.

Like, sometimes you have to go all in on the right thing, and you spend a lot of time and effort identifying the right thing. And so yeah, that's, that's where I'm... what I'm trying to do more of these days.

I think the means of production thing is interesting because- Cloud code only makes sense to be built if it's, like, the best thing, right? Because if cloud code is, like, mid, they're better off promoting Devin and cognition to sell more tokens.

Mm.

So I'm curious, like, as the market gets more competitive, on one way it's like, well, we don't want to use Devin because Devin supports all the models, and so we end up losing some of the revenue. But I think there's...

Right now, cloud code is obviously the best way to use the cloud models, so it drives the most usage. But I'm curious, in the future, there's gonna be more pressure on like, "Hey, this product actually needs to be great to make sense for us, again, to invest our resources into building it."

Yeah, so, so going from model lab to model lab plus product company, right, which is what, uh, OpenAI has done.

I would push back on, well, A, I don't think everyone would agree that Claude Code is the best way to use Claude. I've heard multiple people even in the last few months say that, "I would... I- I'm a Cursor guy," like, or, "I'm a Devin guy."

Like, people have their, their preferences, so I don't think it's set in stone.

Mm.

However, Claude Code is a great way to use Claude also, and there are nice flywheel effects, obviously, because once you capture the way people are using Claude Code, you also get so much data to then make Claude Code better over time.

So I think those are the two main reasons, but at this point of time, maybe this is the, this is, this is oversimplifying, but I can't think of too many apps that have a very meaty layer on top of the model that's, like, very impressive yet.

There are somewhat meaty layers, and it's getting there. It's a time thing as well, right? Most of these companies haven't existed for more than two years. So, um, I think it gets there, but I don't think we're at a point where, you know, we're like, "Holy shit, that app has so much stuff, interesting things and technology built on top of the model," where it becomes so difficult for the model company to go and try to compete.

I think tomorrow if Anthropic decided to or OpenAI decided to take on another app, g- given their distribution and their engineering and the fact that these are still not as thick as you'd like them to be technically, they could.

Uh, whether they should or not is different, but they could. And, and, and that's something I, I do think about.

Thank you for, uh, engaging in all these, like, very meaty-

Yeah

... discussions. Like it's-

Yeah, you don't even work at Anthropic-

I, I-

... so I know we put you on the spot, but-

Yeah, no

... that's okay.

But, like, this is what I wanna get on the podcast because-

Yeah

... like, a lot of people don't get the chance to, like, talk about this, but, and this is, like, a normal SF dinner. The last hit on Anthropic I'll point out, which is more fun, which is, uh, there was a new CTO joining Anthropic from Pesit.

Yeah.

And, uh, you know, you're, like, the ki- king of Indian posting. What is the significance of this for you? You know, last time you were on the podcast you talked a lot about, like, the Indian, um, the university system and all that, and, uh, to see this guy rise up and...

In India, largely academics holds the, the same sort of prominence as sport would hold in America. Everyone talks about it. It's Asian culture, right? Everyone talks about it. It is top of everybody's mind. It is something a lot of people want to be good at, and it's extremely competitive society with a very large population.

The way y- And, and, and everyone, on average, people are quite poor, so education is seen as the means to social mobility by a large amount of people in India. The way it works is similar to countries like China or some other countries where you take a big exam, you get ranked.

A million people take the core engineering exam, and the top 10,000 get in, and the top 200 get into computer science. That's how hard it is. That's pretty hard. And those top 10,000 get into IIT. Everyone's heard of that.

M- That's, like, where a lot of the great, you know, Silicon Valley people from Sundar to, to many other f- people come from, from IIT. And in India often what I have seen, and this is something that I'm generally very curious about, is, like, what is the motivation of humans and what is the dictator of outcomes in their life and their career?

And one thing I've noticed a lot is, A, there are some societies that are inherently, I think, less meritocratic, where you get so judged for what you have in the past that you're not allowed to prosper later. And I think largely many work environments in India and, and, and other places i- in Asia can be like that, uh, number one, so you're not judged on the merits of your work.

You're judged on the merits of what you've done. And number two, there's a very strong self-fulfilling prophecy effect of I've seen people who underrate themselves because they think they, they couldn't be number one at something.

Mm.

You know?

It's like your own mental k- prison.

It's your own mental block where, like-

Yeah

... I couldn't get into, like, I don't know. And, you know, people in the Bay Area also like this. Bay Area is kinda like Asia. Um, in the Bay Area I know people who grew up who are like, "I couldn't get into a, a good college, therefore I am stupid, and therefore I should not work that hard."

Right? Like, it's, it's inherent that they could be smart. They just believe they're not, and that also has an effect, psychological effect on your long-term prospects. You look at a guy like Rahul Patel who's become the CTO of Anthropic, and he's not from a top university in India.

Some people would obviously debate that. But in, in, in general, I don't think it's, it's a really well-known university in, in India. And, uh, and he's come to a society that is quite meritocratic, and he sort of worked his way up to a position of such prominence.

I don't know him. I don't know what everything else he's done. But it's testament to the fact that, you know, I think this is why it resonated with so many people, is even though you didn't have the opportunities early, and even though you might not believe you could do it, if you work hard enough in certain environments for a long time on things you care about, anything can happen, and I think that's why I wanted to share it.

Yeah.

I thought it was just a very-

And, and you choose to work at Stripe and Glean and, you know, do well. But I think choosing the right company is, is also a v- a very... Like, okay, if you're not gonna do the, the credentials path, you have to be lucky and selective at working with, at, at good places, and a lot of people make that mistake and, and I, I definitely did.

I, I had good credentials, and I worked at bad places, and, uh, yeah, it, it's, it's very interesting that, that, that kind of path in life.

You work at a pretty good place right now.

Yeah, but I, I took, I took a long time to get there.

I mean, just the, you know, this is funny. I have this, like, automated pockets research and when it send me the email about you and it's like, you know, "Deedy has a strong presence in AI and immigration"- ...

were the, the top two topics that it talked about. Yeah, let's talk about the Anthology Fund.

Um, so it's a $100 million fund and close partnership with Anthropic. Like talk a bit about that. I think people are really curious about how close that actually is.

Yeah. So, you know, the Anthology Fund we set up when we invested in Anthropic around the beginning of last year, and the sort of idea was, okay, Anthropic again, it's so hard to think about Anthropic was a very different company back then.

It was a much smaller company. And they were like, "Look, we-- there's incentive for us to run our own fund." OpenAI runs their own fund. There's a developer ecosystem that we wanna create around this. It's really nice to have great startups or that are u-using Anthropic, close to Anthropic, building around Anthropic.

And we said, "Okay," but we had a discussion about, do you wanna have it inside Anthropic or do you wanna have it outside Anthropic? 'Cause inside Anthropic would mean something, would mean a corporate venture fund. You'd have to hire for that.

You have to have a whole role. And typically, if you look at corporate venture funds in history, obviously besides OpenAI as a notable exception, they tend to not be very good because all they prioritize is who uses my stuff the most.

And, uh, and that's not a good way to invest in companies, so we thought this would be better. And the incentives in corporate venture funds are a little bit not misaligned. So we, we did that, and now we look back at this fund.

Obviously, Anthropic's in a very different place. We've, we've funded about 40 companies. The rate, it's kind of a hard thing to calculate, but the rate at which companies graduate from when we invested in them to the next round is significantly higher on Anthology Fund companies, and we write both small and, and lead checks.

I mean, the thing-- the two-- several notable companies from the Anthology program have been OpenRouter, Goodfire. There's a company called Endia, Prime Intellect, Whisperflow. So there's a quite a handful of, of pretty interesting things here. And, um, yeah, I think what the other really nice thing about it is it really allows us to move fast on, on companies that, you know, where we may not feel immediately comfortable or ready to write like the full check, so we can like participate in a round and then get closer and hopefully go and build a relationship and, and, and lead that in the future, lead that, the next round in the company in the future.

It also lets them get really close to the Anthropic ecosystem, so we have all these events with like the founders and all the execs-

Yes

... and things like that, and people really enjoy like getting-

I've been to some of them, yeah

... getting it from, from hearing it from the horse's mouth. Now I think, you know, I would say like Anthropic is in such a different place. It's no longer a unknown entity, so, um, the program is, is gets, gets a lot of demand, but you know, people kinda know what they need to know, and so we're still working on like how do we make this program more useful and more beneficial for founders and Anthropic-like.

Yeah. Well, so, you know, I, I congrats on all this. I think it's, it's pretty successful. One thing I'm-- one reason I'm trying to highlight this for, for Lean Space is also like how does AI change venture, right?

And, and something-- that's something that, uh, Alessio is exploring as, as well. And th-that's why like I don't really know how to categorize Anthology Fund because it looks like a kind of like a what Conviction's doing, what, uh, but what YC is doing maybe, but like later stage, right?

Like, um, some of these already have their C. Some of these already have their A. Abacus is in there. Is that, is that, is that our Abacus?

No.

No, no, that's a different Abacus. But, um, what's the model? Like what, what is, what are the predecessors that you draw inspiration from for, for like setting up this fund or do you just not? It's like, it's like a corporate venture fund man-managed by Menlo, somewhat funded by Anthropic.

I would say like you can think of the companies that go into Anthology in three categories. One is strategically important to Anthropic, and those could typically be somewhat later round, somewhat bigger companies. Two are companies that are using Claude heavily and are just great companies to be, to be in.

And three is just very, very early stage founders with-- that are h-very high potential that may potentially be, uh, using Claude models and Anthropic and, and so on. We don't require people to use a certain model or the other, so we, we keep it pretty open and we do everything from, uh, like 100K check to a $20 million check.

So like I think the-- it's, it's really broad in terms of what we can do, and we wanted to intentionally keep it that way. When it comes to where we draw, so there's some old, old examples, but I don't think it's really relevant.

There was a fund called the iFund that, that Kleiner did with Apple way back in the day-

Oh

... which was kinda similar.

How did that turn out?

Um, I don't remember. I, I, I don't actually have enough data on that, but that's one example.

Well, then, then you know the answer.

Um, no, I, I mean, I'm sure there are some great companies that came out of it. I just don't know who-

Yeah

... like the details about who or what was in it. So yeah, I mean, I think so that, that's kind of, uh, how it's been for us, and I think it's been a really great program and we've had...

I mean, uh, we were excited about the companies that we could lead the rounds in as well.

Yeah. I wanted to get quick hits for people who maybe never heard of Goodfire and like I, I know, I know them 'cause I've been, I've invited, um, Mark, uh, to, to my conference, and I've been to a bunch of their events.

Actually, I'll just give you, I'll just give you that list, right? Goodfire and Prime Intellect are in your research category.

Right.

There's others with like diffusion-based language generation, novel architecture. Uh, it's all over the place. Research is like the most Wild West of this. How do you view like sort of research investing?

Re- I can talk about any of those companies for-

Yeah

... briefly as well, but the way I view research investing is it is extremely hard to pull off, but when you pull it off, the results could be very remarkable. One of the hard parts is the tension between do you keep investing in research hoping for something that yields a better result that leads to a better product or do you try to monetize and scale what you have already?

That's tough. It's a, it's a really tough thing to do. It's a really tough decision to make when you're, you know, working with those founders. You're on that board. It's like somewhat anxiety-inducing when you're thinking about this even from an investor standpoint.

Like, do I just get to like a couple million AVR? Do I like start do-doing something or do I like keep the research bet strong? The way I think about research investing overall and is, I mean, honestly follow where- The talented people have the most competence, and then have an idea around w- how this could be useful in what I call a top-down way.

It's not really top-down, but the way I frame it is if f- if I fast-forward 10 years from the future, what do I think is very likely to exist, and what are the ways I can get there? If I do believe strongly that there is something like that, and I believe there's this team very strongly headed towards that direction, I can sort of draw a dotted line and go like, "Okay, maybe we can see something here."

So that's how I broadly think about it.

So concrete example, Goodfire is, like, the most, the most, uh, interesting one. Mechanistic interpretability, I didn't even think that was a market that was worth investing in, but obviously Anthropic does.

Mm-hmm.

Uh, and uh, they seem like they have good vibes. Uh, what, what's the, I guess, the, the summary of y- of, like, your take on the company?

The way I think about the company is right now almost all frontier and some many non-frontier m- AI models are complete black boxes. We don't understand why they produce the outputs they produce. All of the eval and studies on them are empirical studies, not intrinsic to the model.

So it's like, "Hey, here's the outputs we saw, and therefore this is the benchmark score," or, "This is how we think it did." If we believe as a society that five and 10 years later in the future these models are going to be critically important for making pretty heavy decisions, whether it's, you know, I call it ev- anything from whether somebody should get a loan or insurance or a legal decision, then I don't think that the black box appro- approach is long-term scalable.

It's just not how society can function, where it's you say, you throw your hands up and say, "Well, this is what the model said," and then I asked it, "Explain yourself," and it said this other stuff. Great.

Mm-hmm. Mm-hmm, mm-hmm.

Like that's kind of what we have today. That's the best thing that we have. Mechanistic interpretability is really going into the weights of the model and trying to figure out why did the model do what it did. And one of the more concrete and relatable examples of this that, you know, you m- guys may be aware of is GPT-4o had this, uh, phase of sycophancy that, um, a lot of users really liked, but it's kind of one of those things that's not as easily detectable in an eval.

Unless you know you're specifically maybe testing for it, even then it's quite hard. It's very personalized. It's not like any keywords might arise, obviously. But it is something that is quite easy to tell in even current interpretability methods.

You can tell when a model is being sycophantic. You can tell when a model is trying to lie. You can tell when a model is trying to, uh, steal or persuade you of something. And so I think the, i- if we further that research direction two, three years in the future, we will be able to understand why models say what they'd say.

It's brain surgery for LLMs is my cat- catchphrase.

Yeah.

Um, but doesn't apply to LLMs only, all, all models, and that is a pretty important insight into deploying AI at scale.

Yeah. And you don't know the business model yet. Don't ha- don't need to, as long as we fi-

There are some ideas-

We'll figure it out

... that we have-

Yeah

... um, but not ready to talk about publicly, and some that are working also it's not right to be public.

Does it feel worthwhile to do this on such small models? Because I think most of the work is done on the open source releases. Like, how much of a gap is there between what they're able to do and then translate that into doing it for-

There's no gap-

... Opus 4 format

... for scale. Like, w- they've shown that even for the biggest open source models, you, like, even for, like, DeepSeek's big models, you c- they can do it. And, and in general, like, scaling is not the bottleneck. Obviously, access to the weights would be a bottleneck-

Yeah

... but not, not-

But they're in the Anthology Fund, so they can work with Anthropic.

I wish.

Infinite money.

They can work with Anthropic, um, but they, they don't have Claude access.

Yeah.

Uh, Claude weight access, so.

Uh, for listeners who wanna l- hear more about MechInterp, we did a podcast with the MechInterp team, Emmanuel from, uh, Anthropic, so, uh, that's your, like, 101 there, and we'll do something with Goodfire at some point. Prime Intellect, another very hypey company.

Uh, you don't have to say it, but I, I know it's very much in the water that they have raised a very large round. So I ignored distributed AI for a long time. It's usually crypto people coming over saying like, "Hey, we have these GPUs all over, all over the place.

We will somehow ignore the speed of light," and, like, and just like, "You can use our GPUs to train models." That's why I ignored Prime Intellect. I was wrong. Tell me why I was wrong.

You may not be wrong. I mean, look, like, I could be the kind of person who goes and shills all of their companies and says- ... "This is the best thing ever, and if you don't think it's gonna be a $10 billion company, you're wrong."

Every company has risks at this stage, and Prime Intellect has their fair share of risks, and whatever went through your mind went through my mind when I was looking at that company. I do strongly believe in, like, I'm sure you've seen this quote too, is, and the quote of, "Pessimists are probably right often, but they rarely change things."

And it's an easy thing to say, but when you're investing, it's something to think about, which is there's a lot of things that could be potentially wrong with Prime Intellect-

Mm-hmm

... for sure. But the thing that I really liked that drew me to them is what, if they were right about a couple of things, what could go fantastically?

Which is distributed-

Distributed training is one of them. Access to talent, I think is one of the things that I underwrote for them. The ability to hire fairly great people f- away from people, like, other labs is, is, is really hard, and so they, I think they can do that.

And the third thing I think is there's a broader vision to Prime Intellect that is not yet realized yet, where the first step of that was a distributed compute and, uh, and we'll see if they realize that. Um.

Yeah. Well, uh, you know, Will Braun's been on the podcast multiple times, uh, and he's, they've launched kind of like a verifiers SaaS platform or something, or a marketplace. I'm not really sure, uh, what exactly. I should probably try it out, but, uh, it's very interesting.

And the other thing I'll just say out there is, like, like, everything in AI changes, like, every three, four weeks, so, l- I, I'd be a fool to say, like, I could tell, like, what this company's gonna do.

Yeah. Well, well, you know, all I am tr- trying to do is I try to capture for people who are, like, not in the loop on, like- You know, that this, these are the companies that people are talking about, right?

Okay, so, so le- let's, let's at least hit on OpenRouter and maybe one more, uh, of your choice that maybe is, like, less known but you want people to, to know more about it. OpenRouter we have to cover.

Uh, big deal obviously. I, I, like, I, I do think, like, this one I was, like, relatively early on in terms of, like, I, I, I saw the, I saw the products, I saw what he, what he was trying to do, and I mean, it clearly has, has done really well.

I did not know he was taking investment or I would have invested.

He wasn't.

Okay, say more. Say more.

OpenRouter was sort of my, like, uh, i- i- you know, like, maybe I don't want to make this about me. It's really about them. But to, in my mind it was my, my, my darling deal-

You can be proud of it. It is about you

... because I'm just like, man, I b- entered venture and I'm like, that is the company I would have built.

Yeah.

Um-

And can, and for- ... I, I think we're skipping a bit. Let's explain who Alex is-

Right

... what he did before, like-

Right. So I'll give you, let me give you the background on OpenRouter. Alex is a phenomenal, phenomenal founder. He started a company called OpenSea before, which was the NFT company. Obviously that at its peak was, I think, a $14 billion, what, more than $10 billion company.

It did not, uh, meet that valuation's expectations, but look, there are many things out of, out of control and, uh, in your life. Then Alex started this company called OpenRouter, and w- what gravitated me towards it initially was w- two things.

One, it was very clear from my time at Glean that this is a perfect problem where engineers all think it's easy until it becomes so annoying to keep maintaining this. That's the sweet spot, 'cause no other person, no other company will gravitate towards it, yet it is so...

It is kind of thorny to be able to maintain a portal that accesses a bunch of models. The nuances are quite tricky and annoying and boring. So that's one thing I liked. Second thing I liked is I was pretty convinced that if there was a market for anything like this, it would have to be a PLG motion.

I think go so far as to say for in any SaaS market, if there can be a PLG motion, the PLG motion will win. What I mean by that for, like, if you're not f- people are not familiar with venture words like PLG, is-

Yeah

... all users have to be able to access and self-serve the product and try it in order for that to be successful.

Without talking to anyone.

Without talking to somebody, and like the classic, like, get on the phone, uh, on a SaaS website. So those two things really drew me to the business. And then of course, third one is just quality. Like, there's these small details at OpenRouter, just, like, beautiful website, beautiful landing page.

It's not some, like, SaaS trash of, like, "Here's what we do," and he- product, solutions, about us. Like, I am so sick of that. You land on the page, it's a developer page. It's like, "Here's how many people are using what models."

Love it. I'm like, this guy knows what his users really want. And all of those were compelling. I went out to New York to talk to Alex. He ignored me a bunch of times forever. I'd write him what I call love letters.

I'm like, "Hey, man. Love it, dude. Like, it's so cool. I don't even want to invest." "Just talk to me." Like, "I don't really care. I just want to meet you." "I have so many ideas and interesting things."

And, uh, it was one of those companies where I genuinely felt that way. So when I did meet him, um, we, you know, started jamming on things, and I don't know the VC motions of how to sell, so I wasn't really even trying to do that.

But when I told him, like, "Look, if you are ever going to raise, I will make it happen. I just love everything about this." So that's how we ended up doing the round. I think the company is interesting and from a business model perspective, I get this question a lot, how does this business model scale?

And I think right now the business is doing fairly well.

Volume, right?

Well, there's, there's that-

He takes like 5% of, uh, everything.

There's that business model, but then there is a, a reasonable threat factor where, you know, what if the spend on the net goes down over time as tokens go up. So you do take, you do carry some risk of the prices of LLM falling to a point where the business does stops working.

And I know many other companies take that risk as well. Um, so that's one risk of the business on just pure consumer spend. Second risk would be, you know, keeping enter- people on a, like, h- a lot of hobbyists use OpenRouter, and they tend to churn, and then, and a lot of enterprises will use OpenRouter to evaluate and then go pick a model that they want to settle with later.

So that's a problem to fix. And so those are two of the risks, but overall I think they've just, like, been executing phenomenally.

Yeah.

How do you think about the Vercel AI Gateway, for example? I think that's been... It, I mean, I'm a fan of OpenRouter as well.

Cloudflare will also do it, Vercel. Yeah.

Yeah, I'm interested where you already have, like, I use Next.js, right? And it's like, well, I just use AI SDK. AI SDKs comes with AI Gateway. It's like-

It's free

... kind of makes sense to do it. How do you think about this market and, like, how tied you need to be to, like, the actual application development versus if you're just kind of like this, what's our lane?

Hey, we don't have... You know, OpenRouter doesn't have a developer framework, for example. You know, if we're in a partners meeting, that's maybe what I'd, what I would ask.

Like, my simple answer is I don't think the AI gateways of other products are ever gonna be their first priority. And, uh, the other simple answer is I think OpenRouter has this, uh, mindshare and momentum that just doesn't go away overnight.

So it would be similar to asking like, "Hey, I'm OpenAI in 2020. What if somebody else does this?"

Mm-hmm.

Like, yeah, I mean, they could, or 2022. Like, well, they could-

Yeah

... but, like, we are so far ahead in some ways already. I think the last thing is, uh, I think that they have built a lot of smaller things that are non-obviously useful that other people probably won't sweat the details to go out and build.

And so when I say that, I, I'm like, it's everything from, like, here's something that nobody ever, like, even cares about, about OpenRouter, but they have a feature flag where you can only want to go to certain LLMs that do not retain your data.

They go to that level of granularity of thinking about what is, what do the users actually want, and that's o- one example. Another example is their detail on the provider level. Uh, almost nobody has provider insights. There was a very interesting side study of how Kimi K2 did this whole study of different, um-

The verifiers?

The verifiers.

Okay.

But I think that's interesting. Like, the fact that- People don't really acknowledge this, but the same open source model or the same closed source model can be served by different providers and have different context windows, different quality, different latency, different throughput.

Where would you go to see all that information? Well, you see it on OpenRouter. And, uh, and there's some, like, elements of scale where there's enough people using the different providers, so you get that data. So all of those things I think are somewhat defensible on OpenRouter, and, uh, hopefully more over time, so.

Yeah, and I think their leaderboard charts are, like, one of the best growth hacks because-

Very good graph- graphics.

Yeah.

Yeah.

Especially people that are into open source AI are always posting these things saying-

I, I, I'm trying

... "Hey, open source is up. Uh, we're, we're back."

And one, one thing-

Scream for OpenRouter

... I, I, I used to joke about is, uh, OpenRouter is the only non-Elon company that Elon has tweeted the most about. For obvious reasons.

Grok code fast one.

Oh.

Number one right now.

Yeah, yeah, yeah.

I'm sure that's M code free plan, but-

There was like a good week where I was like, every day it's like, "OpenRouter, OpenRouter, Open..." I'm like, yeah.

Yeah, yeah. Yeah. And, and so, so for those who don't know, because that's because Grok Code Fast is, like, a top model.

Yeah, 'cause it's free also, but-

That... Yeah, yeah, 'cause it's free. Yeah, yeah. Yeah, there, there's a lot of gaming, right, of this, of this stuff where it's like, "Oh, we'll, we'll give it to you for free, but then we'll, we'll say we're very popular."

I'm like, yeah, you're free because you're popular.

Right.

Yeah.

Yeah. You're popular because you're free.

The other way around. The other way around. Okay, very cool. Um, and okay, so there, there's, there's a bunch of others. We're not gonna go, go through all 40. What comes to mind? What, what, what do you wanna talk about?

What do you think maybe is a, a very interesting company in your portfolio that, like, more people should know about?

I'll talk about, um, Whisper and Inception are the two I wanna talk about. Um-

Inception. Inception's not even here.

That's why I was, huh.

Yeah. Oh.

So, so, so, so-

Okay. You, you, you can, we can say, we can talk about the company without saying the name.

Yeah. Okay. Let's, let's try that. Let me, let me try that, and then if we don't get there-

But I mean also, like, Inception, it's like, if I Google Inception, it's not like I'm finding it anyway.

Let's talk about these two things. So Whisper I can talk about first. That's a clear one. So Whisper is a company that does, you know, a very, in many people's eyes, something very commodity, which is voice dictation on your, uh, l- phone and laptop.

The things that I really liked and that stood out to us about Whisper was, um, in that, quote-unquote, commodity market, they are, in my mind, like the, the fastest and best and most, uh, delightful product that kind of in many ways set the frontier of the nuances of how to make this easy.

S- like press your function key on your Mac, talk to it. It's always on. It has fantastic accuracy. As you're dictating, if you ever stutter and go like, "Oh, no, I didn't mean that. I actually meant this," it knows what you meant, and it goes and corrects it.

I find that they have this metric they use called zero edit rate inside, uh, which is, uh, you know-

Amount of times you don't need to edit.

Correct. And, uh, their zero edit rate I think is north of 80%, which is insane for a voice dictation product. So I, you know, many other risks of that business too, but one thing I, I think I love is users love it.

Users stay on. The retention is great. And, uh, it might make voice suddenly work because if you think about computing, people type slower than they talk. And so it could, like it is unlocking this new faster way that people feel comfortable talking to their computers that really didn't happen in voice dictation before.

And it's not just a Whisper model, which is a common question I get. So

Yeah. For, for people who don't know, it's W-I-S-P-R.

Yes.

Uh, which, you know, you gotta spell it somehow. The, I mean, the question here is always like it's the same thing, right? Like voice is very commodity. Uh, I actually happen to use Super Whisper. Um-

Yeah, same.

Right. Yeah, right. Um, mostly I- influenced by Jeremy actually. And then Granola is d- very popular. Notion has like this Notion v- speech thing. Like how, what, what, what's the, what's the plan?

This is every AI investor meeting.

Yeah, every, yeah.

This is, this is why I'm not an investor. How do you survive basically? I don't know.

Trying to reason about why you should be the winner is, is so hard.

Even ChatGPT Desktop has like the, you know, has some shortcuts for stuff. I don't know if it like does exactly the same thing, but like, you know, it's not that far away. Anyway, you're excited about it. I do see a lot of tweets about Whisper, and I, it's one of those things where like, yeah, the PLG's getting me, man.

Like I, I, I like I, I'm like, should I switch? I don't know. Like my thing's fine, but like- ... what if it feels better on the other side? I don't know.

Well, we'll see. We'll see how that plans out. There's some interesting plans to get, uh, it to be a, a cooler product, but we'll see. The other company and I, again, one-

Okay. We'll call this Stealth Co.

Stealth Co. One thing I find very interesting about Stealth Co. is comes in the purview of research. We talk about different architectures all the time. One of the most compelling alternate architectures for AI is diffusion models. So o- one thing that I think is really interesting about it is like you do talk a lot, Sean, about like the p- the Pareto frontier of, of-

Yeah

... latency, cost, quality.

Yeah.

Diffusion models today are, I would say 80 to 90% of the quality at one-tenth the cost and latency, so has huge implications on obviously the stock market, which is kind of Nvidia. Um- ... and, and, and many other things, but also like there is clear examples that you can show of use cases where that might be very valuable because there are many applications that work in volume that do not require high quality, but d- definitely require better latency, and everyone could use some cheaper models.

So you know, there are, I think there's an interesting, uh, area of research there. Maybe it gets to frontier, maybe it doesn't. The one thing I want to draw attention to with diffusion that I think is particularly interesting is left to right reasoning for code doesn't actually really make sense.

Mm.

Because in code we don't, like we might sometimes write code left to right, but after you write code, you go up and down and figure out, hey, is this variable set? Did I do this? There are many bi-directional dependencies, uh, in code.

So it, there's a natural tendency to lend itself to diffusion models where you can imagine like as you are denoising, you fix partial issues in different parts of the code at once versus this reasoning paradigm where you kind of have to figure everything out and then go give your final answer

Yeah. Yeah, I like that a lot, uh, especially for like, uh, syntax structures like C-like languages where you need to open and close a bracket and, and all that and hold that state. I think, like, it's, it, I...

The question is always the qu- sort of, quote-unquote, "the hardware lottery" of transformers. Like, transformers is all you need, and like, uh, diffusion is kind of like a different branch off of that tree of research. Uh, they are related, but, uh, we might be too far gone down the transformers tech tree to come back and then go down diffusion.

Like, it, it being the point where, like, they might never be frontier because we've just had like four more years extra of like transformers LLM research.

Yeah. It's true. I, I think about this all, all the time. Like, like thinking about in the course of history, what are the significant moments where if only something forked off a different way, that maybe there would be a completely different paradigm of outcome?

Yeah, and usually the worse tech, worse tech wins, like Blu-ray/DVD, uh, HD/DVD or something like that. I think there's like a, a lot of variations of this. Even like, I think there was a discussion about AC versus DC currents like back in-

Hmm

... like Edison's days. Like what, there was this like whole big fight, uh, between, between, uh, Tesla and, and, uh, Edison. Um, I don't know if you-

I mean, I'm, I'm aware- ... of the very, very basic details. But, but like it's so interesting, right? Because like just you take something like this and then the question becomes like, okay, do we bet on it or is the timing just off because something took off and n- and we can't pull this like rocket ship back to earth and so we've lost that fight?

I don't know. I'm not a purist scientist anymore where I believe like the best ideas and things win. I think in markets it's very obvious that that's not true. Um, I think a lot of things go into winning and-

Yeah

... uh, sometimes it's out of your control.

Yeah, yeah, yeah. It, it, it's very true. Like, uh, and, and you know, speaking of Anthropic and like things that happened this year, MCP happened this year. And I was... When MCP came out, I was, I was sleeping, and then when, when w- they came and then did the, the workshop with me and, and I think as you see a lot more noise and I was like, okay, there's something to this.

And, and like now it's like basically kind of de facto one as the interop layer for all the labs and all, all the, all the models. And there's no reason why this could have won versus ev- anything else apart from like it was well specced out.

It was backed by Anthropic. It's, it's kind of a similar thing. Like I don't know if it's like the best, but like it was good enough.

Yeah. It's happened, it happens so often. It kind of makes it tricky to, and not in just investing, but in general to think about ideas. We see this with startups as well. It's, it's very heartbreaking every, every once in a while you'll, you'll meet a founder where I'm like, "Your idea is fantastic.

Your execution is great. I just don't see-

Hmm

... it work because the market dynamics are not in your favor." And maybe I'm wrong about some of them, but you know-

When you say market dynamics, is it, uh, TAM or something else?

No, it's, it's-

Competition

... sometimes it's like I don't see the... Like you are a small group of people trying to wedge something into a market. We know how long that takes, and we know the other forces at play, and if I don't like, I just don't see...

Imagine a single person running in a tunnel with a light at the end, but the tunnel's closing in on you. You could be the fastest runner in the world, and you might not make it out of the tunnel.

That's kind of the analogy.

I see.

And, uh, and so you might be doing everything right. It's just that that window is not there, or at least I might not think that window is there. Um, I do think a lot of companies fall into this bucket of ideas and so...

To me, in a way, I almost think of companies like most HTML in a way, which is like, hey, we got this amazing team. We can help you fine-tune models and yeah, but nobody, you know. The market dynamic just there's really nobody fine-tuning models.

And part of it is like the open models are not that good, and part of it is like people don't really have good data. They don't have the expertise. And again, if you go back now, now there's like, you know, RL environments and like RFT is like the next wave of that, and it's like maybe they'll be able to get in the window.

But it's just interesting how, you know, now I'm saying it's-

And yet then the other flip side of that is, and yet they get acquired for this amazing price.

Right. The mar- but yeah, it, because the market is just so big. I mean, even if you think about something like, yeah, diffusion models for text, right? It's like, you know, it's a be- it's like if you sell it for a billion dollars, right?

It's like 0.01% of like NVIDIA's market cap. And so it's like, okay, well, the amount of money being spent in this space is large enough to justify betting-

Yeah

... like the same way Instagram was like 1% of Facebook market cap. It's like this is similar-

Yeah

... where it's like, man, if-

Databricks is rich enough thing.

Exactly. It's like, you know, like-

They really want you to know that they're an AI company.

Exactly. And now they're worth 100 billion. I, I mean, you know, like without MosaicML-

It worked. It worked

... exactly. It's like without MosaicML, maybe they're not on the same trajectory. It's like, I don't know. Maybe they are because, you know, Ali is great and all that but-

I, I don't know if you guys have ever talked about like the roll-up companies, which is my favorite like little, uh-

The PE roll-ups?

Yeah. Well-

I didn't know that was a topic of yours.

It's not really a topic of mine. I just find it quite interesting to see how t- speaking of AI companies and markups, it's there are companies, obviously I'm not gonna name them, but there are s- companies who go like, "Hey, here's like a small company that does a million of ARR completely with humans.

I'll buy it for two million, and then I'll do some of it with AI, but now I'm an AI company and a million of ARR in AI company world is $100 million valuation." And so, you know, it's, it's, it's pure like multiple arbitrage on-

Yeah

... the category that you're in.

Yeah.

Um, so-

But like, y- yes, that's the like cynically haha.

Yeah.

But then like what if it actually works? Because like the hard part is getting the customers. The hard part is like getting the domain expertise. You b- drop a bunch of software engineers in there and like, you know, automate it and make it scalable, make it cheaper and like, yeah, maybe it works.

No, you're right. You're absolutely right. I'm-

I think you're just pricing-

It, it, it-

You're just pricing the asset

... he, he funded the company that bought a ta-

Yeah

... a tax firm, so.

Yeah.

Yeah, a law firm.

No, look, I, I-

Accounting firm or tax firm?

A law firm.

Law firm.

Yeah, yeah.

If it works, it works. I just think what was interesting to me is like you can 50x the value of the company before you actually landed anything with AI yet.

Yes. But then, then you use that funding and the equity to, like, hire the people and... It's, it's weird.

Yeah.

So, uh, there's this concept I always talk about, uh, which that I'm surprised n- people don't really understand, is reflexivity. The belief that something can be true can make it true even though it's not true at the time that you believed it.

Yeah, that's venture capital.

Yeah.

Just give money and everybody's like, "Oh, they raised 300 million. It's a great company."

Yeah.

"I love that company." It's like, "Yeah, I'm an investor in it, so I love it, too." And it's like all the employees are like, "I love this company. My stock are, is worth a lot of money" and-

There's also that effect that's very clearly in venture capital where not just what you said, which I agree also happens, but imagine there's, there's times where people funnel so much c- money into a company before it's really, like, prime time, which dissuades anybody else from entering that market and then they become-

Yeah

... the de facto owner of the market-

Yeah

... 'cause they cancel the competition with funding.

Yeah.

Mm.

Um, and you can think, and I'm not gonna name the categories, but you can think of numerable categories-

Mm

... in this market, in, in, in this paradigm-

HR payroll software

... where that's, that's already happened.

Yeah. And I feel like even in AI it's, like, maybe two and a half years ago when ChatGPT came out, it's like this is cool but, like, you know, a lot of enterprises were, like, maybe skeptical of, like, is this trend gonna continue?

But then once you start seeing tens of billions of dollars being put in OpenAI and Anthropic and it's like, it's gotta work. But-

Especially when you deploy new hardware.

Yeah.

Which, uh, you know, I think, like, you're- at that point you're building infrastructure, and infrastructure very capital intensive and, like, you, you actually can do the math. It's not, it's not humans anymore. It's, like, machines and land-

Yeah. Yeah, exactly

... and power.

Like, like Amazon is building all these, like, training chips and, like, all this infrastructure for Anthropic. It's like, do you really think they're dumb? Like, you know what I mean? I think at some point it's like, same with Stargate.

It's like, do you think all these people are dumb?

Yeah, yeah. So the-

And, like, you're saying the models are not that good. It's like, you know?

Uh, the, the podcast we released today with Kyle, like, he was still kinda skeptical that they had 500 billion for Stargate. And I'm like, not only do they have the 500 billion, they have the next, like, trillion, like, lined up.

Mostly 'cause, like, the, the projections... A- a- and I think, like, I, I, I've been talking about this a lot and I'm very out of my depth 'cause I'm not Dylan Patel. But, like, I think it's the most big- it's probably the biggest story of the year, like, beyond the models, like, the just the infra build.

The infra build, yeah.

Of, of like, um, you know, like, a- and- and I think, like, people don't understand, like, the, the, the, the roadmap is very, very strong for them to, like, the rest of this decade at least for OpenAI to go from, like, two gigawatts of, of, of compute this year to 30 with everything they've already announced, and then there's a plan to afford the next 125.

Like, the United States uses 300. It's, like, crazy ambitious.

Do you think, like, I, I guess as a question for you guys also 'cause I, I don't have a good answer yet, is the, the belief is always obviously bitter lesson pilled, right? Like, you buy more compute, therefore you-

Right

... get the most models.

And is it, by the way, it's an Anthropic relevant thing.

Mm-hmm.

Right?

And so but, like, is... I guess is that necessarily true? Like, i- there could also be a world where that's just not, not true. So-

Yeah

... you know, you are kind of betting-

It's like this is what makes it bitter. It's like what if it doesn't apply to me this time?

Right. Right. And I think, you know, being in Sam Altman's place that's absolutely the right chess move to play but, you know, I do wonder what happens if, like, all this investment in compute doesn't actually lead to economic gain/better models/everything else.

But I feel like we have reached the point where, like, the models are good enough that even if the next generation is not 10x better we'll be able to use the compute. I, I mean, and again, a data center is like, you know-

Yeah, that, that's the cope.

They're writing it down for, like, 30 years.

Mm.

So it's like, you know, can you run GPT-5 Pro over the next 10, 15 years?

But do you think, like, given the amount they're spending on compute... And this is a general question, I'm not criticizing OpenAI at all.

Oh, no.

Is even if everyone ha- was using Clau- like, whatever, Codex, Claude Code, whatever, all the time, like, inference demand is not that big globally, right?

Not yet.

So you, you would have to believe-

Not yet.

So what would you have to believe for that to be true? 'Cause there are 800 million weekly active users.

This is what Greg Brockman tell, says, like, a GPU for every human on Earth.

I, I, I'm, I'm somewhat shitposting.

Yeah.

I'm somewhat shitposting but they actually say this on their official comms, so I'm just repeating him.

I, I, I, I don't, I don't necessarily disagree. I'm just trying to work backwards to, like, what do we need to believe to get there? 'Cause ChatGPT compute is not that much.

Correct.

Right? So, like, they're not doing it, like-

Yeah

... agentic stuff. Maybe they will be in the future. Most people are doing basic Q&A type queries.

Uh, by the way, I put it up on chat so, so, uh, if, uh, people are watching on YouTube they can see this, which is, uh, this year OpenAI spent $7 billion on compute. Only two of that was for all of, uh, their inference.

Right.

The remaining five was R&D. So all of ChatGPT, all 800 million users, all of Sora, all of, like, all, all the, all the sort of, like, uh, API volume, two billion and they have two and a half times that for R&D.

Right. And so my, my, my, uh, my point being, like, yeah, if inference is one thing, I don't know how that will scale to that volume, but then you'd have to believe that the rest of it goes into R&D and therefore produces models that are so much better that therefore b- have more demand, et cetera.

But if in any case that, like, I don't know, the incremental marginal is not that big, then, you know, that's the risk of, of, of the-

Yeah

... of the bet.

Yeah. So all your... Like, to disrupt OpenAI you need to have more efficient research because right now it's pretty inefficient. You know, spend five to get two. So, you know, so, like, what OpenAI did to Google is what the next OpenAI has to do to OpenAI.

Yeah.

Right? You know what I mean? Like, Google was spending a lot of money. Facebook was spending a lot of money, and, like, they didn't come up with anything. OpenAI did, and it was, like, a small, tiny, little, uh, you know, startup.

And, you know, they had, they had, you know, GPTs and now, like, Ra- Redford, but, like, someone else will, like, may or may not come up with that. Um-

It's like that classic quote, "Your margin is my opportunity." Like, Google was milking those margins, and they didn't wanna spend-

Yeah

... the, the compute for every search query and so-

Yeah

Now OpenAI will, is willing to.

So we've covered a lot of topics. Uh, I, I think this- Thanks for indulging. Like, I think this is like, for me, it's like a survey episode of, like, here's everything. We're also catching up with a former guest and it's always nice.

Maybe we can end it on this, like, coding interview thing-

Mm-hmm

... uh, which w- which literally you tweeted about today. What is the situation that, you know, I guess e- engineers should be aware of? And I think this, like, maybe ties into LLM psychosis a little bit.

You know, like, like, so I tweeted... I'll just cover the tweet first. So I tweeted about this, uh, guy who wrote a blog post about he was in an interview from a ps- I, I didn't think it was a legit account.

He thought it was a legit LinkedIn message where he was interviewing for the company. They sent him a coding interview. They said, "Clone this repo, run this code, make this edit." Kind of not untraditional. So it's pretty, pretty run-of-the-mill type interview.

It happens. And in that interview, he s- claims that he went to Cursor and asked whether the code had anything, any vulnerabilities or anything he should be aware of, and it revealed that it had some link. It had a byte array that compiled into a link that would go and take a bunch of private information from you.

So that was the TLDR and, uh, and I tweeted about that saying, you know, like, the, the world... Interestingly enough, it was solved by vibe coding, but it could very easily, the world of vibe coders who don't really look at code, I imagine are more susceptible to being in attacks like this and in the future.

Mm.

And, uh, and it got me thinking about a lot of things, like what is- what do attack vectors even look like if people aren't looking at code? There's so much that can go wrong. And what are the implications on model safety and how models behave in those environments?

So that's one, but I think the broader thing, and, and I'm curious what you guys think about this, is what I've been noticing more and more is I was having this conversation yesterday with some of my close friends where, you know, some of the joy of coding used to really be you're stuck on this annoyingly hard problem and you just bang your head against a wall and you want to kill yourself, and then eventually you're like, "I've figured it out," and then you solve it, and that's, that's the muscle that, that you build and when you improve and get better.

And now I find myself even doing this. It's so hard to do if you just have a constant slot machine that might give you the right answer, and who knows if it will, who knows if it doesn't, but you just pull it all day long.

"Please fix, please fix, please fix." And, uh, and what does that mean for the craft of engineering or software engineering in the future? I, I don't know. Like, this vibe coding stuff, I mean, great for the rest of the world that was not an engineer, but I'm, I'm now seeing how it's affecting the, the trained software engineers and it's kind of like a drug for them.

Mm-hmm.

And it stops them from, like, living their own life which is-

Yeah

... doing the engineering because-

It turns your brain off

... it just turns your brain off.

Yeah. I, I think, you know, self-driving cars, people thought about this first. This is why when you drive your Tesla you have to, like, keep your eyes on the road, uh, because they don't want you to turn your brain off and we don't have that equivalent in, in, uh, developer environments yet.

Maybe we should-

Hmm

... like, watch your, watch your eyes.

Yeah.

We remove one- ... one word in the code. Which one was it? Write it back.

So my ans- I mean, I happen to have shipped a model today, uh, two models, um, and, uh, part of that is actually what I've been calling the semi-async value of death, um, and a lot of it, I think, is, is my reflection on coding agents in terms of, like, uh, we started with Copilot, which was tab autocomplete, and then when we went all the way to Claude Code, which is, like, very async, very, you know, like, uh, just it could take 30 minutes, it could take 30 hours.

I, I don't know. It just, just, it just, it just runs. And I think, like, something that Cognition's very interested about is fast agents or something I've been writing about more is fast agents, is where, like, under a certain level you actually want to just be in a mind meld with the human and AI, uh, to have, like, fast responses so that you can get helpful assistance if it helps, or you can get out of the way if it doesn't help, and, um, it like, that is actually where you do your hardest problems, and then the async agent is where you do the commoditized, dumb, boring labor stuff that you know how to do, you just don't need to do it.

But when you are actually very deep work and focused and you're working on a hard problem, you are, you are, like, you should be applying your human intelligence augmented by AI in an unintru- unintrusive fashion-

Mm

... which I think is the way that, obviously, I think it's like, it's a pro-human message, but it's also, like, a really interesting area of research for us.

But that's almost like, to play devil's advocate there, that's like telling somebody, "Well, I'm gonna put the cigarettes right here. I know you love smoking, but so please don't do it."

It's not, it's not a cigarette.

"It's right here." It kind of is. It's in... There, there is an analogy, right, to be made here. It's, it, it's a cigarette for your brain because you do not think anymore when you pull that button.

Mm-hmm.

And, and over time I feel like, you know, the brain will get weaker if you don't use it for that task. And, and, and I like your, your, your message. I mean, I would ideally, like, i- if I was, had a team of engineers, I would also tell them the same thing.

But I mean, I, I, I worry about the reality which is that's not what they do-

Yeah

... in many cases.

But, but I mean, you gotta ship the thing, right? Like, I, I agree, but at some point you gotta close the ticket and merge a PR.

Mm-hmm.

So how are you gonna get the code done? Right? It's like th- they are doing it or they're gonna get fired if they're just, like, generating-

They, they, they are doing it one way or the other

... they're writing the enterprise integrations.

One way or the other.

They're writing the B2B SaaS. Yeah, it's interesting. Uh, okay, so, um, maybe I'll put it this way and I'll see, I wanna see what, how you respond. Okay, so, um, we have the formu- the, the fundamental formula for coding agent performance, okay?

It, it basically is find the right files and then write the right files. That's it. Or, like, so read and write. Like, read the right files and write the right files. That's it, right? So actually what fast agents can do or, like, what, um, you know, what, what I just, uh, did today was basically the equivalent of a heads-up display.

Like, give you more info, but you take, you still take all the actions. So we help you read, uh, read faster, read more efficiently, read, uh, with more focus, uh, but you still write.

Mm-hmm.

And so I think, like, that's still, that's not a cigarette so much as, like, we try to be helpful-

Yeah

... and, like, we're, we're evaluated on the helpfulness of the, the reading and the comprehension so that you can hold everything in your head.

Hmm.

That would be the, the pitch.

It's, it's true. I, I, I think there, if... I, I don't know how the product looks. I would love to eventually, uh, play, play with it, with the sweet grep and all of that stuff.

Yeah.

And but, um, there's a world where I think the, the product decision also get, goes a long way into how people use it. So if it is like that, then, then maybe, and I, and I think when people use f- even for example, if someone uses a cursor, a lot of people like the fact that they can see the code and then they kind of have to hit the final accept.

Yeah.

Um, so-

Human in the loop

... human in the loop. But, you know, I still, I worry. I still worry.

Yeah, yeah. It's, it's valid.

But, um, and I worry the most about, like, the younger kids, right? Like, the, you, you think about the people growing up in college. How would you ever get yourself to think if you just had this, like-

You're cooked

... clearly more intelligent thing than you?

Yeah.

At least for, like, I, and I don't wanna, like, rate myself too highly, but if I'm working in a domain that I understand, I can at least tell AI, "Yeah, model, you're, you're doing the wrong stuff. Like, you don't, definitely don't do that.

That's, don't write that at all. That's a terrible file. Why are you hitting four files for this?" But if you think about what it looks like to a 18-year-old CS major freshman, they're just probably like, "I guess that's how you do things."

Um, and, like, they can't hold it at bat.

Yeah.

So when they, like, their, their training is just a little bit different.

Cool.

Yeah.

All right, Deedy, thanks for indulging and welcome back and thanks for coming back.

Thank you, guys. Always fun chatting with you guys.