Uncapped with Jack Altman
· Chetan Puttagunta
[00:00:00] S1: You just have to roll back to the end of 2022 when we happened to get involved with these spectacular entrepreneurs. The thing that became clear to us sitting around a table was that AI was the thing. Even if it wasn’t the thing, it didn’t matter. That was where we were drawn to. It was the thing that had this gravity pull for us. All we wanted to do was spend all of our time talking about it, thinking about it, meeting all the people working on it. Then you have to overlay your value system on the thing that you’re excited by. The thing that we laid on top of that was: what kind of relationships do we want to get into with companies, in that moment in time? We decided we wanted to be in business with companies where that ethic of being the primary partner—board partner, lead investor, first investor, principal investor, principal believer in the mission—is how we wanted to practice investing in AI. That meant we were looking for spectacular entrepreneurs with unique approaches to the market. If you want to do that in a place where you want to be the first investor and back teams with two, three, four, five people, you’re often meeting people that are probably a little bit early on whatever the next curve is. Remember what was happening in 2022 and early 2023. Everybody wanted to start a foundational lab and everybody wanted to aggregate GPUs. There was a big drive to aggregate capital to buy GPUs that would be utilized for training runs. It wasn’t that we had some hypothesis or some macro view of why we don’t want to do that or do want to do that. We were looking for companies and entrepreneurs that resonated with us and wanted to partner with us. In that moment, we met a lot of people that were working on really aggressive ideas. We thought they were spectacular people with spectacular approaches. As a result, you saw this list of companies that we compiled at that time. That was when we did Sierra, when we did Fireworks. We did LangChain, we did Mercor, we did Levelpath, we did Legora, Manus, et cetera. All of that came together. When Eric did Cerebras, the Series A, it’s all of that same stuff: partnering super early with founders working on something that they’re deeply passionate about. Frankly, it’s cliché to say, but all of those investments at the time were a little bit non-consensus. I think you have to be.
· Ev Randle, Eric Vishria
[00:00:00] S4: It’s interesting because from the outside in, before I joined Benchmark, if you look at the investments that were made in that ‘22, ‘23 time period… You had an inference cloud with Fireworks, you had a data infrastructure platform with Mercor, you had a horizontal AI play with Sierra, you had a vertical AI play with Legora. Obviously it’s only four of the ten great investments that were done in that era. It was easy to ascribe a kind of thematic nature. [00:00:00] S3: Then you got here and you were like, “Oh my God!” [00:00:00] S4: “These guys had no idea what they were doing!” [00:00:00] S3: That is correct.
· Eric Vishria, Ev Randle
[00:00:00] S3: That is correct. [00:00:00] S4: But you come in and you’re like, “Oh wow, they had vertical, they had horizontal, they had data. They must have done a market map.” Then you come in and start asking about each of the investments and the story behind each of them. 90% of the story on every single one of the investments is the person, the founder and the entrepreneur, and the relationship that they built and why the entrepreneur was so special. That was so revealing to me coming into this organization and this partnership. The founder centricity just bled off the page in terms of the stories of all those investments. The way we’re approaching it today, obviously we love to talk about all of the newest and greatest things going on in AI every single week. But in terms of the actual investments, it’s always founder-centric. It lends itself especially to this era, because the sands are shifting beneath the founder’s feet so quickly in AI and things are changing so rapidly. Founder centricity as an investment strategy matters more now than anytime in the last decade.
· Eric Vishria, Ev Randle, Jack Altman
[00:00:00] S3: I agree. The underlying technology substrate is changing very quickly with AI. Any software that you could have built in 2022, you could have built in 2010, plus or minus. Once we had the cloud, we got little APIs here and there, but for the most part, for twelve years it was pretty stable. If you compare that to today, you’re getting more change every quarter than we did in a decade in terms of the substrate. A founder’s ability to navigate that and actually understand where their edge is and where their edge is going to come from, and how quickly the moats are deteriorating—because they’re deteriorating really quickly—and how do you build the next one… It’s critical. So I think it becomes even more important. We see all the same things. It sure feels like the infrastructure cycle is going to continue. It sure feels like five years from now, or maybe even sooner, we’re going to have really interesting things in robotics. [00:00:00] S4: Agents are getting really good. [00:00:00] S3: Agents are going to get really good and the applications are going to get better. All these things seem really clear. But that isn’t enough to make an investment. [00:00:00] S0: It’s impressive to me. It would’ve been so tempting to answer my question with some high-minded thesis, and you all just said “founders.”
· Peter Fenton
[00:00:00] S2: Also if you step back and think of these as forces, not specifically AI or social or mobile, there are windows where the disruption is so high that entrepreneurs come in and they see something with such clarity that they can’t not do it. Then there’s this lag effect where years later all the other people come in afterwards. In a sense, that happened already with AI. It happened with the first generation. It happened with the people saying, “Okay, now we’ve got to do our model company.” But when your brother and the team of people were at OpenAI, before it was obvious, lightning struck. My belief is that it is likely to happen again at least one or two times in the AI cycle. There’s something so disruptive that no one can fully understand it. Then there’s a completely new kind of entrepreneur that emerges. What we’ve been watching for the last three months, post-Opus 4.5, is that the force of that disruption has awakened a whole group of entrepreneurs that were otherwise not seeing things— they’re seeing them now—three months ago. I’m a big believer in what you’re doing with Sunday and robotics, because that’s a world where we all know that in a decade we’re going to have these things in our house. But the path from here to there, is that an incumbents’ game where the big companies are going to push down complex products? Or will there be entrepreneurs? So much of this is, where is there disruption? But then the lag effect of our industry is that 90% of the capital flows in afterwards, after the entrepreneurs have figured things out. By the time it’s figured out and there’s the next thing, it’s not for us. That’s our faith: Silicon Valley is an adaptive landscape that will continually have these disruptions. As soon as there’s a winner, we’re uninterested. We move on.
· Jack Altman
[00:00:00] S0: It’s a great place to end. Thank you, guys. This was really fun.