🎧🍌 The State of AI with Nathan Benaich, Founder of Air Street Capital
The rise of reasoning, Chinese open source, surge in sovereign AI, why AI isn't a bubble, and how to invest in AI today
For the past eight years, Nathan Benaich has published The State of AI. It’s a year-long effort covering the biggest things happening in AI, across research, industry, politics, and safety.
This conversation covers the biggest takeaways from the latest report, like the rise in reasoning, the surge in Chinese open source, where AI is actually working today, the rise of sovereign AI, why Nathan thinks there’s no AI bubble, where value will accrue over the long-term, and how he’s investing in AI today at his venture capital firm, Air Street Capital.
Support this Episode’s Sponsors
Numeral: The end-to-end platform for sales tax and compliance.
Flex: Sign-up for Flex Elite with code TURNER, get $1,000. Apply here.
To inquire about sponsoring future episodes, click here.
Timestamps to jump in:
3:39 State of AI 2025
6:22 Takeaway #1: Reasoning & tool calling
13:01 Takeaway #2: Rise of Chinese open source
15:25 Open vs closed source models
26:46 Takeaway #3: AI revenue is real
27:51 Takeaway #4: Sovereign AI
36:44 Are we in an AI bubble?
59:23 Starting Air Street Capital
1:05:18 Raising Fund 1
1:16:20 Air Street portfolio strategy
1:25:15 When and who Nathan decides to invest
1:35:04 How important are AI benchmarks?
1:39:31 When to train your own models
1:45:56 Rise of European defense tech
2:01:43 Nathan’s personal AI stack
2:07:32 Is niching down too risky?
2:16:12 Nadal vs Federer
Referenced:
State of AI Report
Thinking Game Documentary
Find Nathan on X / Twitter and LinkedIn
Related Episodes
👉 Stream on YouTube, Spotify, and Apple
Transcript
Find transcripts of all prior episodes here.
Turner Novak:
Welcome to The PEEL. I’m your host, Turner Novak, founder of Banana Capital. Today’s guest is Nathan Benaich, founder of Airstreet Capital and author of the State of AI Report. Nathan has been writing the State of AI for eight years. It’s a year-long effort on the biggest things happening in AI every year across research, industry, politics, and safety. We spend the next two hours talking about the biggest takeaways from his latest report, including the rise of reasoning, the surge in China’s open source models, where AI is working in practice, the rise of sovereign AI, where he thinks value will actually accrue over the long term, if we’re in an AI bubble or not, and how he’s investing today at Air Street. A quick thank you to Nico at Adjacent and Dan at the University of Michigan for helping brainstorm topics for Nathan. A reminder, I publish two episodes of The PEEL every week, exploring the world’s greatest startup stories just like this one.
Check out the back catalog of over a hundred episodes, including recent conversations with Barcelo Lebra, co-founder of European Unicorn Remote, and Kevin Harts, co-founder of Eventbrite and SeedInvestor in PayPal. Tune in over the next few weeks for guests like Gary Tan at YC, Chathan Pudagunta at Benchmark, Jakes Dotch at Serval, and duo security co-founders, Doug Song, and John Oberhad. Let’s talk to Nathan after a quick word from Numeral and Flex. This episode is brought to you by Numeral. Numeral is the fastest, easiest way to stay compliant with US sales tax and global VAT. It’s easy to set up, and they automatically handle all registrations, ongoing filings, and their API provides sales tax rates wherever you need them with all the integrations you need. Numeral supports over 2,000 customers in both the US and globally, and they pride themselves on white glove, high touch customer service.
Plus, they guarantee their work, and they’ll cover the difference if they mess anything up. They’re fresh off a fundraise, closing a $35 million Series B from Mayfield, which they’re going to reinvest into building an even better product. If you want to put your sales tax on autopilot, check out Numeral at their new domain, numeral.com. That’s N-U-M-E-R-A-L.com for the end-to-end platform for sales tax and VAT compliance.
This episode is brought to you by Flex. It’s the AI native private bank for business owners. I use Flex personally, and I love it because they use AI to underwrite the cash flow of your business, giving you a real credit line. The best part is 60 days afloat, double the industry standard. Flex has all the features you’d expect from a modern financial platform like unlimited cards, expense management, bill pay that syncs with your credit line, and their new consumer card, Flex Elite. Flex Elite is a brand new, ramp-like experience for your personal life. A credit card with points, premium perks, concierge services, personal banking, cars and expense management for your family, net worth tracking across public and private assets, and a whole lot more fully integrated with your business spend. One card for your businesses, one card for your personal life, one card for everything. To skip the wait list, head to flex.one, and use my code TURNER to get an additional 100,000 points worth $1,000 after spending your first $10,000 with Flex Elite. That’s Flex.one and code Turner for $1,000 on your first $10,000 to spend.
Thank you, Flex. And now, let’s jump in. Nathan, welcome to the show.
Nathan Benaich:
Thanks for having me, Turner.
Turner Novak:
Yeah, thanks for coming on. So you put out this really interesting report, it’s called The State of AI. You’ve been doing it for a while. I’ll let you explain how it all got started. And I think I gave people a little bit of context on kind of what we’re going to talk about, but can you just talk about this report that you put together, how it got started, why you do it, all that stuff?
Nathan Benaich:
For sure. So the State of AI Report is an annual production. It’s an open access document that I create in order to disseminate the most interesting analysis across research, industry, politics, and safety. And then we wager a couple of predictions every year in order to cast things forward and see how we did the year after. The real goal is just to help people stay abreast of what’s going on. There’s just so much stuff and you don’t know what is meaningful and most important. It also acts as a good litmus test, I think, insanity check, to see like, “Hey, have we overexceeded on progress estimations or under exceeded, or just how far have we come in the last couple of years?” It’s been running for about eight years now, starting in 2018 with Minnie and Hogarth who produced it together for a while, and then been solo production for the last couple of years.
And it’s a good opportunity also to collaborate with people in the ecosystem and showcase good example case studies of how businesses are using AI and how certain papers are leading through breakthroughs and what might be head fakes.
Turner Novak:
Yeah, I feel like there’s a lot of data, a lot of numbers and charts and a lot of visuals.
Nathan Benaich:
Yeah, yeah. It’s definitely not for the faint-hearted. I do try to write it in such a way that you could consume the headlines of every slide, and then get a general sense of where things are going. And then you can stop where your eyes get most transfixed. And so it’s a bit more of a buffet than must read end to end.
Turner Novak:
How much time would you estimate that you put into this thing, because you do it once a year? How many hours or, I don’t know, total days or however you want to quantify this?
Nathan Benaich:
It’s a bit hard to give a clear answer on that because it is the result of just everyday consuming of research, news, talking to companies. So it’s very much like the result of my day job. But when it comes to genuinely producing slides, it’s from early August until September.
Turner Novak:
And then you usually put it out beginning of October, it looks like?
Nathan Benaich:
Yeah. Yeah, we found a good tempo around the beginning of October, so back to school season.
Turner Novak:
The report you put out in October, what were sort of the biggest takeaways? What we’ll do is we’ll throw a link in the description for people who want to actually look at the whole thing. But if you’re just like, “Ah, give me the, I don’t know, the Spark notes, quick version,” what are the biggest things that have been happening that you think people should know about?
Nathan Benaich:
Yeah. To me, the probably four biggest things are one on research, which is clearly the move away from models consuming an input and just rapidly producing an output, but now going into fairly complicated reasoning and tool calling. That wasn’t the case even a year ago. You look at AI today and you’re like, why wouldn’t it be anything else than this? But this wasn’t the tabletizer a year ago.
Turner Novak:
For somebody who’s never heard of tool calling before or reasoning, what does that mean practically?
Nathan Benaich:
Yeah, yeah. So where we were maybe a little bit more than a year ago was a model had consumed basically the entirety of the internet, people like to say, and then maybe some custom databases, and has essentially compressed and memorized all of that information. And it’s basically, in simplified terms, it can be thought of as an API to all of that knowledge. And so when you input a query, it’s looking up at its knowledge and then producing an output. It did not have access to the internet. And then the first tool call was do a web search for, for example, more relevant information, because at the time what was very important were these knowledge cutoff dates, which is basically like the date at which the download of the internet was made. And so if something material happened after the cutoff date, it wouldn’t be stored in the model because it didn’t have access to real-time web search.
Turner Novak:
You’d ask, “Who’s the President of the United States?” And it would say, “Joe Biden.” It would be like, “No, he’s not.”
Nathan Benaich:
Exactly, or who is your creator?
Turner Novak:
What did it show for those usually if you searched that?
Nathan Benaich:
Well, there were some interesting ones. I mean, OpenAI would say OpenAI and Anthropic largely the same, but the Chinese models would say weird stuff. They would sometimes say like, “Oh, my creator is OpenAI or my creator is this other company.” And then that led to question marks on the Twitter sphere of, are they training on outputs of American large models? It’s what’s called distillation, basically. There was some FDA headlines about this. OpenAI were saying that this was the case. And that was one of the reasons, in a cornucopia of other reasons, why US administration wanted to cut off Chinese access to American AI and things like that. So it’s unclear if it’s actually schizophrenia or because there’s just a lot of examples of AI output that’s generated by the leading labs on the internet. And so statistically, it’s more common to have that memorization.
Turner Novak:
Yeah, so essentially what this means is that it incorporates just the models with what’s on the internet. So the models can learn and take in new data versus just what’s been trained on and captured in there and it’s set forever. So these things can actually get smarter in a way and use current relevant news and information?
Nathan Benaich:
Yeah. I think it’s even more than just being able to consume current news. It’s like the designers of the system would ideally separate memory, like storage of facts from the ability to know how to retrieve them, and how to logically reason through answering a query. And so ideally, you would want to learn the latter capabilities and not have to memorize facts in the weights of a neural network, because then how do you selectively update them as information changes? And so yeah, the big push towards like tool calling and web search is a tool.
Using an API could be a tool, using a software product could be a tool, is really to combine that with step-by-step reasoning where human annotators have explained in a very, very clear way, a structured way, how they would go about answering the question of like, “How should I implement this stock trade given this is my goal and this is the ticker and these are my financial goals, et cetera,” and anything you can really imagine. So if you repeat that in enough times and you really learn the logic of reasoning, then that logic can be abstracted away into many, many other areas. And it’s more repeatable than just memorizing facts and memorizing patterns.
Turner Novak:
So if people are using a lot of AI products, where’s this capability shown up in what we’re all using today?
Nathan Benaich:
The most obvious one is just when you enter a query like in ChatGPT, it’ll say thinking, and then it’ll produce a result. And then you can click on the thinking tab. And then you’ll be shown simplified stepwise list of, “Okay, I think the user is querying this, or maybe they’re not. Oh, maybe I should go disambiguate what they’re asking, and then this is the stepwise plan to get to the answer.” And I think it’s interesting because it gives you a vignette into what the model is doing to answer your question. So you can build trust in if the reasoning trace looks like what you would’ve done.
Turner Novak:
You can almost like troubleshoot it in a sense too.
Nathan Benaich:
Yeah. But then there’s some other research around are these reasoning traces otherwise called chain of thoughts actually truthful? Is the model actually doing that when it’s telling you it’s doing that?
Turner Novak:
Just making it up.
Nathan Benaich:
Or is it actually just making it up?
Turner Novak:
I didn’t realize that was a thing. Interesting.
Nathan Benaich:
Yeah, because in some ways it’s like trained to recapitulate human behaviors of knowledge traces. And so it might be producing a result, but showing you what the human would’ve done because that’s what a human would rate as a good reasoning trace.
Turner Novak:
So they like learn how to lie, maybe, or hide their hallucinations.
Nathan Benaich:
Yeah. There is some of that paper wise that’s come out. And to some degree, it really depends on how you measure that behavior. And maybe your measurement is wrong and therefore you’re drawing the wrong conclusion.
Turner Novak:
Interesting. So you said that was one of the four. Is that probably the biggest thing that’s happened?
Nathan Benaich:
I think so, yeah, because that’s like unlocking computer use, which is like the model can interface with your computer and do a bunch of things with it. It’s unlocking more complex scientific and maths discoveries, which the AI world is very excited about. Yeah. And then I’d say like the second big thing is the rise of Chinese open source. It’s only 12 months ago that DeepSeek moment happened. And it was probably six months before that when DeepSeek, the Chinese company had released early versions of its model, this base model that was then tuned to do reasoning, and resulted in this DeepSeek freakout.
And since then, many more models have reached leaderboard headlines, in particular in world modeling and in vision. So basically anything to do with like making pictures or videos, long form videos. Chinese systems are very good at that. And then the resurgence of Qwen notably, which is Alibaba’s system that effectively took the mantle from Meta Islama initiatives, where just a year ago, the tech world was applauding Zuck and others at Meta for stepping into the fold and saying, “We will always do open source. It’s good. We’ve done this with PyTorch and other of our core technologies.” And then of course, big vibe shift there and China really stepped into the fold.
And then particularly in the last 10 days, there’s two filings for Chinese model companies, one MiniMax and one is called Knowledge Atlas Company, which actually the regular name is like ZhipuAi or Z.ai. It’s pretty small, it’s called GLM. Anyway, it’s like too many acronyms at this point, but the point being is these are the two first pure play large model companies that have gone public, and they’ve gone public on the Hong Kong exchange for several billion dollars each and they’ve since ripped. Yeah. So China has been first to market in a sense.
Turner Novak:
In the first to market in getting a stock that people can buy?
Nathan Benaich:
Yeah. Yeah. Well, Americans can’t buy MiniMax. If you have US Nexus, you can’t buy it. I don’t know, I haven’t looked up the exact reason. But the other one, I think you can.
Turner Novak:
Interesting. Are there like ADRs where basically they relist the shares? I think I don’t know exactly how that works technically.
Nathan Benaich:
There are no ADRs yet, it’s just on the Hong Kong exchange.
Turner Novak:
So one thing I wanted to ask you about, maybe this is a good time, like open source versus closed source, what is the big deal with that for someone who has no... Why is open source so important in the context of what does closed source even mean? Maybe give us a real quick 10 second, and then what the importance is.
Nathan Benaich:
Yeah. There’s a spectrum on this. So traditionally, open source in the context of like regular non-AI software meant that the source code, which was basically like the end-to-end book basically that produces the program, was available on the internet and could be reused.
Turner Novak:
You could copy and paste a code, publish it, and use the software technically?
Nathan Benaich:
Yeah. Yeah. And it would work, yeah. And you’re allowed to use it sometimes for non-commercial reasons, sometimes for commercial reasons, with or without paying a license. And the beauty with open source that people got excited about was that anybody on the internet can propose a suggestion or change or discover a bug or produce a feature, and then the maintainer of the project could basically approve or deny or provide suggestions. And there’s like a moniker a couple of years ago that was used to describe this, and it’s like the bizarre where everybody can participate. And then in other ways, closed source software is like a cathedral where you have one access point and no one can touch it. No one can touch the cathedral, no one can add anything to it. You can’t influence how it looks or how it will be changed, it’s the owner of it that decides, and that’s like closed source and you pay a license to get access to it.
Turner Novak:
You go and you offer an offering like a tie to the cathedral to use the cathedral, you pay a tax?
Nathan Benaich:
Exactly, you pay tax in the form of a SaaS license.
Turner Novak:
Yeah.
Nathan Benaich:
And then this nomenclature just gets a little bit more complicated in AI because there are more moving parts. So for simplicity’s sake, there’s the training code, which describes how a model, how an algorithm should be run, and what’s the goal of the training, what objective are you trying to minimize? For example, like the dog-cat thing, classification, you’re trying to minimize the mistakes. Then you have the data set, and then you have the model itself. Once it’s been trained, which has weights and parameters, which are like settings, effectively, that are produced as a result of running data through an algorithm. And so what is it exactly that we mean when we say open source model? Is it the training code? Is it the data set? Is it the model artifact once it’s been trained in the sense of its weights or the code to run it? So there’s a spectrum of definitions. Right now, a lot of people use open weights to mean, well, I’m actually just releasing the resulting parameters, tuned parameters of my training run.
This would be the equivalent of saying, I’m basically giving away the result of hundreds of thousands of hours of GPU training or something like this. I’m not necessarily giving you access to the data. I’m not giving us the training code or the pipelines to run it.
Turner Novak:
So you couldn’t reproduce it, you can just see the results of what I did?
Nathan Benaich:
It’s hard to reproduce the training run because you don’t have the data set, but you’re getting the end artifact. So it’d be like, I don’t know, like the Formula One analogy, the car that wins the last race at Abu Dhabi in the year, you get that, you don’t get all of the learnings and recipes in order to get the car, you just drive the car. And so the community generally wants to have open everything because it leads to better reliability, more contributions from the ecosystem in terms of features and bug fixing and things. I think crucially, control over the entire system. So if a company decides to no longer publish open access and open source tools, then you can get rug pulled.
Turner Novak:
Because you just no longer have access to it or updated...
Nathan Benaich:
Yeah, exactly, at least you have the last timestamp release. So if the system’s good, you can still work with it. I think the second reason why folks want open source is because the training of these systems is expensive, generally for anything that’s like state of the art, like millions, tens of millions, hundreds of millions. Not many companies, or certainly not universities or small groups have access to those resources, and so they can’t train systems from scratch. So they rely on these getting airdrops basically for free. And I think the last one is just the community doesn’t want to have one, two, three companies the entire future of AI development because it is very limited on money, on compute resources and on dataset.
Turner Novak:
And then the reason you would be doing a closed source model is because it’s probably easier to make money from it, like if you have to pay to use it and access it versus just giving it away and you can’t charge as much or anything.
Nathan Benaich:
Yeah. My two takes there are like open source is almost good when it’s a necessity for your customer to buy your thing because if your customer’s a developer and the tool is catered towards them or some low level like database or framework or something like that, they don’t want to peer under the hood and see how it works. And so if that’s part of your sales motion, then you have to do open source. But I feel like rarely, open source has been a competitive tool that one uses because it’s like good for business, it’s almost like a necessity, or secondarily, if the business is born from an open source project like Databricks with Apache Spark, like the thing is so complicated but very high value and customers don’t want to have to manage it themselves, so let’s pay Databricks for the expertise because they wrote it and they’re the best.
And then the other part around open source versus closed is considering that it’s like Apple versus Android. At the end of the day, I think customers and regular people want a tool that’s like really good, that’s designed well, that’s updated all the time, that’s reliable, simple to use, and that’s basically Apple. There is a period of time in the early days of technology, I think, where a lot of nerds like to play with it and mod it and change the background and change these little things. But at some point, it’s like that community I think goes down over time because it’s just a fath, it’s annoying. And I just think when the last time was that I added a mobile phone number that was not an iPhone, like this, ooh, it’s a green check mark and your iMessage is like weird. And I think this is because long-term, people gravitate towards convenience and quality. And I feel a little bit of the same way in the open source versus closed source.
Turner Novak:
Interesting, yeah, because you probably get to a point where as the technology continues to get better, you’re just like, “I just want this thing to work so I can do my value additive thing that I’m doing on the model versus mess around with tweaking things because I just know that it works and it has what I need it to have.” It reminds me, it’s interesting how Apple, and I think of how Apple evolved, I used to think of Apple as for old people back in the ‘90s, your grandma has an Apple, one of those big, massive monitor things. It’s like a fancy color, it’s like purple or something. And then it’s like there’s only one button on the mouse and it’s super simple to use. You can’t get a virus because the old people, they’ll just get phished and they get a computer virus or whatever. But it’s like totally evolved over time to now where most people just use a MacBook product, and you’re weird if you use Windows still.
And I feel like people will reluctantly use them if they’re working at a big corporation or something like that where there’s like a work necessity. But yeah, it’s just interesting how Apple’s evolved the products over time.
Nathan Benaich:
Yeah. And I think realistically, as model capabilities have gotten better, these things were so unwieldy and complicated and insane lifts with thousands of people contributing in very tight control. And I just don’t know how you coordinate all of that on the internet with like randos on your laptop in cafes.
Turner Novak:
And there’s tons of forks, right, where you basically take it, you make a tweak and then you republish it? So there’s probably thousands, tens of hundreds of thousands of just variations of different models that have been slightly tweaked over time, I’m assuming.
Nathan Benaich:
Well, there’s that in the open source community, yeah, but like to create a Gemini or like a ChatGPT or something, there’s like basically all these different teams that are each responsible for like a different skill or capability, and each skill or capability has a set of evaluations and sometimes like hundreds of them, which is like ways that you test whether the model has succeeded or failed or what end of that spectrum for that specific capability. And then you have to collect data specifically for that and then just throw it into the soup and make sure that everybody else who’s throwing in their capabilities and data and metrics and evals into the soup doesn’t degrade each other’s capabilities.
And so how do you coordinate that at scale on the internet without being a centralized company? I just think it’s unrealistic. And I think the proof point to that is like no one has really managed to do it, in the sense of no one’s managed to publish a model as good as Light Opus 4.5 or GPT 5.2 in an open source setting. People have come close, but those are basically private companies that operate. So those are like closed source model companies that just happen to open source as opposed to like true open source development, which is like distributed contributions.
Turner Novak:
So it’s almost like you need a dictator or like an opinionated like the PM, the person managing the whole process of training and releasing it.
Nathan Benaich:
Yeah, exactly.
Turner Novak:
So then I know this has taken us a long time to get through the top four most interesting things, but what would you say is like the third or fourth most interesting thing?
Nathan Benaich:
The third or fourth, we can probably breeze through a little bit. The third is around like real revenue scale of products. It was only two years ago that the major labs were de minimis revenue. And there was that example that Jasper was making way more money than OpenAI was for very similar use cases and this debate of is the model the product or should there be lots of other like scaffolding and sticky tape and UI around the model that would mean that it could extract more value for that targeted end user. And then that seems to have flipped as the core models better and has been the interface for everything. And then model vendors have created their own SaaS wrappers, basically. So that’s like tens of billions of dollars across major players.
And then the last one I’d say, and there’s many more, but last and my favorites at least is just like the AI sovereignty marketing agenda has just gone on full steroids. So this is the competitive positioning idea of nation states need to have access to energy, compute, data and talent and models in order to dictate their own fate and the future of AI. And NVIDIA has gone on a very aggressive marketing spree, like selling this agenda to every single country that is willing to listen. And that’s resulted in commitments of over a hundred billion dollars globally in different countries to build up data center capacity.
Turner Novak:
Just pick a random country in Europe, like the fourth-largest country like Italy or something in Europe, and I need to have my own sovereign AI strategy. What am I probably doing?
Nathan Benaich:
Well, this is my point, it’s like, what does that even mean?
Turner Novak:
Yeah, what does it mean?
Nathan Benaich:
I just keep invoking this Kanye West lyric of, “No one knows what it means, but it’s provocative, it gets the people going.” It’s a bit of that. What they think they’re getting is the ability to train and run AI models in their geographic vicinity in a way that cannot be unplugged or tampered with by any other nation state in themselves.
Turner Novak:
If you go to war with a country that a major lab is domiciled in, you don’t get cut off from the ability to use AI.
Nathan Benaich:
That would be the idea, yeah.
Turner Novak:
Yeah. And I could see how if AI continues to kind of get better, it becomes all software. You basically can’t use software to defend yourself and maintain your sovereignty. And you’d be at an extreme disadvantage if you didn’t have that. If it’s positioned like that to me, I’m spending a couple percentage of GDP. Yeah. It’s like I’m hoping in the checkbook probably.
Nathan Benaich:
Yeah. Yeah. But the challenge with it is that the hardware improvement cycles are very short. You still need to have a bunch of software that runs on the chips to make them work. And maybe there’s a major update that gets shipped to them that’s required to run new models on your hardware and the vendor decides not to ship it to you because we’re not friendly with your country anymore. So congrats, you’ve got the hardware, but you can’t run it a bit like the whole spiel with the F-35. Yeah, foreign nations can buy the F-35, but you can’t fly it if you don’t input the flight plan, which gets communicated to the US.
Turner Novak:
I didn’t know it works like that. I mean, that’s tricky. Yeah. How do you get around that?
Nathan Benaich:
Or it just could be nerfed.
Turner Novak:
And it’s interesting from NVIDIA’s perspective because NVIDIA has a customer concentration problem where basically Google, Facebook, I don’t know, Amazon and OpenAI, maybe Anthropic is most of their revenue like, I don’t know, 90% of their revenue or something like that. So it’s like, can you figure out how do I find a couple more nine figure customers? Get a couple more people that are paying me billions of dollars a year to get a little bit more diversification.
Nathan Benaich:
Yeah. Yeah. I think it’s a great strategy, by the way. I think they’re right to do it. And you can see that the next growth spurt is data centers in space. So it’s like, what’s always like the next TAM you can create to keep selling?
Turner Novak:
What is sort of the end state of that? Because if you kind of go back maybe a couple years, if someone were to say, “We’re going to build data centers in space,” it just seems like just a word salad of just random buzzwords that makes no sense. Does the entire universe just need to be filled with data centers power AI? Where does it end? Do you know? What actually happens?
Nathan Benaich:
I’m not sure what happens at the limit, but I could see a pitch around as we get more assets in space of various kinds, and maybe we’ll have space civilizations at some point. And so having compute theirs is just better than having it on Earth that we have to communicate to and from, which currently is cheaper than having data centers in space. Potentially even for defense reasons, you might want to have compute in space for all of your space systems. But yeah, I’m not the space data center expert at this point.
Turner Novak:
Yeah, because part of me is like this just sounds like it’s hundreds of years in the future, but also maybe part of me is I feel like there’s that quote of less happens in one year than you think and more happens in 10 years than you think. So with self-driving, my working theory with self-driving is it’s always going to be a couple years away, but I guess now it’s actually finally here.
Nathan Benaich:
Yeah. Yeah. But also, historically, NVIDIA has been exceptional at resourcing teams, whether they’re in companies or in academia who are exploring new things with their stuff. And before it was like shaders and graphics, and then it became like AI with AlexNet in 2013. And then it was these AI labs, nonprofits that were started 10 years ago and then in biopharma. And then now if some kids are launching your GPU in space, I think NVIDIA is pretty curious to know, does it work? And if it doesn’t work, how can I fix it? Because who knows where the next growth frontier is. And if that’s driving excitement for this stock, even better. But I think their net outlay in figuring out if this works or not is pretty low for the potential upside they could get.
Turner Novak:
And it’s not like they’re losing money because they’re selling products that they generate cash on, right? So it’s like you can scale it up or down like, “Oh, these space data centers didn’t work, but we made $3 billion because everyone wanted to buy some for a year.” It’s like with crypto, you go back to 2021, the crypto bubble almost is like dwarfed by the AI surge.
Nathan Benaich:
Well, this is interesting because as you said, it’s been dwarfed and then crypto prices year-on-year now are like down, whatever, like 20, 30% or so. I remember this time last year it was like 120 or something and now we’re at like, what, 95, at least in BTC.
Turner Novak:
97, just in Bitcoin’s in 97. I saw a notification. So it’s going up. It’s going back up.
Nathan Benaich:
Yeah. Bitcoin miners have found something far more lucrative than mining Bitcoin now because of changes in the hash rate, et cetera, and energy prices. And that’s like basically swapping their compute workloads from mining to AI. And so last year we saw this rush of companies from Iris Energy, which is this Australian business doing energy originally and then starting to do BTC mining or crypto mining and now it’s like transitioned to being a GPU company. And same thing for Cipher Energy and then like Hut 8 and all these rando companies that are...
Turner Novak:
Have you even heard of any of these?
Nathan Benaich:
Yeah. I mean, you could look at them. There’s like six of them and they’ve all ripped like 100% or 50% last year. I mean, they’re crazy volatile, so you got to have a stomach for it and who knows if any of them are really going to survive because they don’t have the balance sheet for it. But as traders are looking for volatile stuff to just play on Robin Hood, these things have been pretty popular because everybody’s trying to front run, “Oh, Anthropics launched a new one million TPU data center thing. Who are the other vendors that can profit?” And so I think for that reason, BTC prices have also gone down.
Turner Novak:
This kind of leads into interesting other kind of phase of the conversation that I want to ask you is, how do you feel about where we’re at and kind of the cycle of excitement in AI? Maybe the more simple way of phrasing it is like, are we in an AI bubble right now? I don’t know. How do you think through that as someone who’s been in the space for a really long time?
Nathan Benaich:
The first thing I think about is if I were to have seen the capabilities that we have today, like this magic box, you can ask anything and basically gets anything right. And you showed me that 10 years ago, I’d have told you that is like absolute magic and pretty much everybody in AI would’ve told you that’s magic. That’s not possible in a decade. And this is across just general question and answering but also video understanding, video generation, audio. It’s insane what we have today. I would’ve thought this would taken way longer and the future got pulled forward so damn fast.
And then we’re still only in, what, two or three years into this cycle of getting this magic alien artifact with no instruction manual, no Genius Bar that tells you how you’re supposed to use it. And the very companies that have developed these tools are figuring out how best to use it so they can educate everybody else. And then you already see like vignettes into the future where companies that have adopted this and people that have adopted this have significantly higher productivity than they had before. As reframed as if you were to take a classic question of if I were to take this thing away, how pissed off would you be? And most people would be pretty beeping pissed off.
And then also other things like cloud span as a proportion of overall IT span is still not ginormous, and that’s been the case since like 10, 20 years, and that hasn’t even factored in AI distribution. So I think we just have so much more to run if you just consider what we have today and educating everybody on how they can use this stuff. And also factoring in that the artifacts we have, at least of, if you look back three years ago were a result of a very, very small number of highly technical nerds who stumbled on something that really worked. And we’re not product designers, we’re not like behavioral psychologists, we’re not like large scale systems engineers or cloud or cost optimization experts, et cetera. All this stuff, all these tasks that are super important building like basically the internet had not worked on AI and now everybody’s going into AI.
So I think there’s far too much money, resources, talent, and genuine desire to use these things that progress can’t get better. Now does that mean currently we’re in a bubble? I don’t actually think so. There might be like pockets of bubble-ish dynamics, but the companies that are accelerating superfast in AI like Mag 7 are actually printing tons of money. They have super healthy balance sheets there. Four valuations are I think like 50% of the top names back in the late 1990s. They’re using their own money to fund a lot of these data center buildouts. I mean, there’s some that are doing off balance sheet credit, but I mean, it’s like sophisticated buyers that are buying this. It’s not Joe Bloggs that’s betting their pension or something on Meta’s data center buildout.
Turner Novak:
I think I saw a really interesting stat too. It was that this was probably about maybe six months ago, nine months ago at this time. So this is probably even different stat now, but it’s basically OpenAI and Anthropic since the launch of ChatGPT had added more new revenue than every other publicly traded software company. And this was a while back, so it’s probably an even bigger... It’s probably even bigger now.
Nathan Benaich:
Yeah. It’s astonishing.
Turner Novak:
The valuation of these companies are just like, do people pay for your product and do you make money on it? Maybe that’s a whole different question here. You can argue about the margins of some of these things, but people are adopting the products and paying for them much more than if you don’t have any AI capabilities in the product.
Nathan Benaich:
Yeah. Well, the margin profile looks like it’s actually improved quite a lot.
Turner Novak:
So why so? Because I feel like even within the past couple months, I still see some of these headlines where it’s like, oh, the vibe coders have negative gross margins and they’re three months from bankruptcy or something like that.
Nathan Benaich:
Yeah. Well, I guess those are a bit different because they’re like passed through to model providers. So one thing, the margins are pretty good. I’m talking specifically about model makers, whether they’re in image or video or chat. And I would say without disclosing specific names of companies, some of the best ones are like at 60, 70, 80% gross margin on serving their models. And yeah, they spend a lot on training but to some degree less than they did before because whereas pre maybe one year ago, the sort of quote, unquote recipe for how to build a large scale model that worked really well was not clear or not written. And now I’d say most people at the frontier would say there is a recipe for scaling now. We sort of know how to do pre-training, the human mid-training and then the post-training and then RL and then the evals and all this stuff, kind of know how it works.
Turner Novak:
So we’re getting more efficient.
Nathan Benaich:
A lot more efficient. And then the other thing is the quality of the model you get today compared to what it was a year or two years ago is actually far more useful. So then you can monetize it better because it’s monetizable, like people were willing to pay for the outputs. And so you actually get faster payback. So you’re more efficient. People pay for your thing because it’s better. More people are educated on how to use it so they want it more and then your general go-to market is improving. And there’s some modalities that are cheaper than others. Text obviously is probably the cheapest. Audio is actually not too expensive. Video and world models are very expensive.
Turner Novak:
You’re saying to make or to also sell?
Nathan Benaich:
It’s expensive to do the training and then also expensive particularly to do the serving, the inference.
Turner Novak:
So just the pricing to consumers then it’s probably higher. If you want to tap into a world model API, it’s probably much higher than just text-based.
Nathan Benaich:
Yeah, exactly. To make a nice video on Google’s Veo 3 or the super printout model, it’s a couple dollars. So if you’re selling to social media people and it costs a couple dollars per video, it’s a bit tough.
Turner Novak:
Yeah. But on the other hand, if someone else like an agency or a designer in Illustrator, or literally take a camera, record an expert doing the thing or like an actor or a dancer, whatever you want, what is the cost of literally type in “make Nathan dance with an iPhone for an Apple commercial” versus actually go make it? Is it a thousand times cheaper and it also took two minutes instead of two months? I feel like that’s part of the equation too of like as the products get better, things just get so much more efficient where you can literally have... It’s like the person who’s the marketing person at Apple. Instead of coordinating with the agency, they’re just working with Nano Banana or ChatGPT or insert whatever tool they’re using and just instead of sending an email to the agency going back and forth, they’re just banging with ChatGPT for a day. It’s like, cool, this is what we would’ve got for a thousand times cheaper and literally in 10% of the time, we got it 10 times faster too.
Nathan Benaich:
Yeah, exactly. So it’s not particularly surprising that like 11 labs is ripping. Or it’s tweeted, it was at 330 million in revenue and got 100 million in net new ARR in five months, so bananas.
Turner Novak:
Yeah, I saw 15 million in a day or something they were tweeting about, which is, yeah, I mean, you run right that, that’s pretty high. And that’s what people are doing, right, is they say like, “Oh, we’re at a $8 billion runway because we signed whatever our Stripe balance to hit the account today if we times that by 365.”
Nathan Benaich:
Yeah, I mean, that would be mad sketch. Yeah.
Turner Novak:
Yeah. I mean, it’s kind of happening, isn’t it? A little bit?
Nathan Benaich:
I think there’s everything that’s happening. I would say they’re pretty above board, but there’s certainly some odd behavior everywhere.
Turner Novak:
Oh, yeah.
Nathan Benaich:
But I think the magic here is just how much latent revenue there was in these use cases that was only made unlockable as a result of great capabilities because I still remember two years ago when Eleventh started or when Synthesia started seven years ago, it was not mega obvious that there would be like half a billion dollars worth of revenue for a product like this.
Turner Novak:
Yeah. You’d be like, “Who gives a shit about why would somebody want an audio model? There’s no demand for that.”
Nathan Benaich:
Right. Right. Well, especially that there were audio models before. I mean, Amazon had one, Google had one for TTS and speech to text.
Turner Novak:
Was it that they just weren’t very good?
Nathan Benaich:
They were kind of shipped.
Turner Novak:
Yeah, that makes sense.
Nathan Benaich:
Yeah. I mean, in London where I spent a lot of time, it was always funny to me that I’d exit King’s Cross Station where Google DeepMind was and you would hear the audio voice and it was like, “This is clearly a robot and it sounds really crap, but there’s a company like 300 feet from here that has a way better model for the last 10 years and the tube can’t even use it.”
Turner Novak:
We’re still so early.
Nathan Benaich:
Yeah. Yeah.
Turner Novak:
So I’m curious then, are there areas within AI that you think are a little bit overextended when I just think about... Because if you were to open social media and you’d read the commentary from investors, a lot of people say AI is in this massive bubble and things are going to crash. So it sounds like you maybe don’t necessarily agree with that, but are there certain areas where you feel like things have gotten a little too far ahead of themselves maybe? And I think maybe another way to think about it is we kind of talked about across the board the products just keep getting better. Are there areas where maybe the products aren’t actually getting better at the pace of maybe how excited people are getting about them?
Nathan Benaich:
In some areas like in science, I’m very excited about long-term progress in AI for science, which is anything from helping human scientists process research papers, formulate ideas, figure out what experiments they should run, test, run those experiments and analyze the data. AI models are making substantial contributions there that’s like a net positive for the world. Clearly in biotech, this works. There’s a good business model for it. If you develop drugs, there’s ecosystem of pharma companies that’ll buy them. There’s contract research organizations that’ll run the experiments for you at scale. And the industry has transposed all this excitement over to materials. I think there’s been a phase of several companies raising substantial money for materials, so much so that the venture dollars have gone to materials companies is far greater than US national funding for material science labs. Whereas the US government should be funding frontier or R&D, venture dollars have stepped into there using many analogies from life sciences.
But to me, I’m very excited about that space as well, but I haven’t done any investments there because I think it does not share a couple of the key features that is present in biopharma, which is like lots of pharma companies that have been structurally built to buy biotechs either outright or the drugs they’ve created. This is how biotech originated. This is like the tacit agreement that small biotech companies have with pharma companies. We take all the early risk and then if it works, you buy us, that’s how it works. Material science, not really.
And then there’s not like this kind of industrial base of contract research organizations, which are like industries that will run experiments for you. In the case of materials, okay, my AI model has popped out a bunch of different versions of like titanium or like some composite and I want to go make it. I can’t send it to anybody to go make it. And so most of these companies are either doing their in-house labs, expensive for a whole bunch of other reasons, a bit inefficient, or they’re sending it to academic groups who take like six months to make the thing, and then they’re making a tiny amount of powder and it’s like... And then the other thing is then you have big companies like Meta that have been pouring in billions of dollars into material science and then releasing data sets and models and are really just constrained by synthesis, like actually making the suggested outputs of the model. And so I think materials is like rate limited by all these things, yet is seeing like irrational exuberance from venture investors as the next frontier that’s similar to AI and biotech.
Turner Novak:
Interesting. So what is an example of like one of these materials or companies?
Nathan Benaich:
Periodic labs?
Turner Novak:
What do they do?
Nathan Benaich:
I mean, they’re probably one of the most exciting ones out there because they have a very cool team. Two co-founders, one worked at OpenAI on ChatGPT and post-trading, and then the other one worked at DeepMind doing material science. And then they’re doing as far as like public reporting as chronicled. They’re doing in house testing of materials and then using AI to scour the universe of potential materials, do optimization, and then using the internal labs to test, et cetera.
Turner Novak:
Just to be inventing a new periodic table, like inventing a new element that goes on the table or something like that?
Nathan Benaich:
It wouldn’t go so far as that, but it’d be like what’s like a new formulation of like a superconductor or what’s a new formulation of like an alloy.
Turner Novak:
Like stronger steel or something?
Nathan Benaich:
Yeah. That could be stronger and lighter and cheaper to manufacture, and therefore it could be great for lots of different things or like a material that you could make, an iPhone screen that you could then go outside and not get blinded and it would just work with direct sunlight, things like that.
Turner Novak:
That could be amazing. I can totally see how you make these awesome products, let’s use AI and make... I don’t know. How do humans have wings and they can fly or something? I don’t know. I can see how the TAMs are on these things are insane.
Nathan Benaich:
Yeah, no, I mean, I think it’s very inspiring. The question is like, would you pay almost close to public market valuations for material companies for a private business that’s just started to do this on the basis that it could be like the next Frontier Lab in this space and follow the same fundraising dynamics?
Turner Novak:
I mean, that’s kind of the role that a lot of venture investors play is like the momentum identifying and trading almost in a way. So I guess it’s just like, is that your strategy? And if that’s your strategy, then that’s what you should probably do if that’s what you’re telling your LPs is what you’re doing.
Nathan Benaich:
Yeah. Yeah, for sure. I think it’s the strategy that one should do if one scale to billions of dollars of assets. I think there is probably no other way to move that volume of money into ideas that could be big enough, which you could say on the founder side is amazing that entrepreneurs have the opportunity to go pursue their dreams with insane balance sheets on ideas that should have been funded by the US government a couple of years ago. But at the same time, if it’s your dollar that’s getting spent and you look and you’re a little bit more pragmatic, which is what LPs tend to be, it’s a bit of a stretch.
Turner Novak:
Yeah. And it’s just so fascinating when I talk to people that were in the industry in the ‘90s or whatever and they’re like, “Yeah, I have to sell 60% of the company for like a million dollars and there’s like a 2X liq pref and it was like a real product, real business. It was profitable.” And then today it’s just like, “I interned at AI Lab and I raised $8 million to go build material science thing that, yeah, it’s like I’m going to make stronger iPhone glass.” It’s like the contrast is pretty insane.
Nathan Benaich:
Yeah. It’s tricky because it’s all about like risk reward. You know, you can take a lot of risk if the reward upside is big. But if the starting valuation is billions, then to reward is not necessarily capped, but I mean, there’s some reality to how much money. I mean, as soon as you actually become a materials business, then you get valued as such.
Turner Novak:
Yeah. But the argument that people will use is that the outcomes are so much bigger today and that with generally with like... I mean, with a lot of AI software, it’s like replacing the work, right? So your customer might historically only be spending 5% or 2% of their revenue on software, but they spend 50% on labor. So in theory, you could maybe capture half of their labor budget. So your revenue potential goes from two to 20% of their revenue, which is 10 times a bigger outcome. So you can afford to pay that 10 times higher entry valuation.
Nathan Benaich:
Yup.
Turner Novak:
So I don’t know, you can argue it where it works.
Nathan Benaich:
I think that’s true on the outcome size. I would just rather not have to pay the 10X on the front so the return is 10X bigger, but to each their own, I guess.
Turner Novak:
You come in, you let the multi-stage platform funds do that after you, and then they help support what the capital needs to kingmake it towards the public markets after you’ve invested as an investor.
Nathan Benaich:
Yeah, that would be ideal. But I mean, some of the mega funds are doing this already with these mega fundraises where the company announces multi-billion dollar price tag, but there’s been three trenches before that where brand A mega fund got in first, for a seat manager is a high price, but for one of them is like whatever. And then there’s been two markups since then in the short period of time, so it’s a bit of that.
Turner Novak:
And you can almost like manufacture those in a way where you put in 2% of your fund at 200 million post, and then you put in another 0.1% of your fund in a follow on round with one of your LPs who leads the round, who for them, it’s also a very small percentage of their portfolio. And then you have an extremely small round where the company sells like 2% of the business at a billion posts money and everyone kind of looks good on paper and nothing’s changed in the two months since the first round happened.
Nathan Benaich:
I mean, this is exactly what’s happening sometimes. I don’t know, it just strikes me as pretty unhealthy. And sometimes the rationale is, will employees want to know that their equity is worth a lot of money and so they’ll get the equity early and then there’ll be this big writeup. So it’s almost a way that companies can offer lower percentage ownerships in their business in the form of options, but because the markup is so high, like the paper value is like already in the millions, and so then they can try to compete with offers at the big labs, but obviously dangerous because of all the liq pref features that are present in these extended valuation companies, not least growing into them. So that’s a little bit scary to me.
Turner Novak:
Yeah. And for people who don’t know, liq pref is liquidation preference. And it’s basically in a lot of these rounds that companies raise, any dollar that comes into the business that if there’s a liquidation preference, it’s usually one times. Sometimes it might be two times, three times even, depending. But typically if you’re in your standard venture round, whatever dollar you raise, those investors kind of have right of first refusal on any exit. So if you’ve raised a billion dollars, the company needs to be worth a billion dollars and all that cash goes to them first, and then beyond that, the employees, the founders, et cetera, get paid. So if you take a job at a company that’s raised a bunch of money, I mean, there’s a ton of nuance to this, but it’s probably like whatever amount you’ve raised, the valuation needs to at least be that price before you get anything.
The investors generally get kind of a rise of first refusal. It’s not always the case. You should probably ask when you’re kind of talking to... I’m taking an offer and figuring out exactly how that works. Google it. There’s tons of YouTube videos that explain it pretty well. There’s like full podcasts where people explain the nuances of this stuff. But the price, just because you work at a unicorn might not actually be worth that much at the end of the day.
Nathan Benaich:
I mean, the other thing is whether your stocks get vested immediately once the acquisition happens if you’ve stayed for a shorter period of time than is required.
So if you get acquired in your first year and you’re supposed to stay four to earn all your stock, do you earn the four on the date of the acquisition or not?
And that’s been particularly topical in these pseudo acquisitions of shell companies and things.
Turner Novak:
Yeah. I did actually have that with one company that got acquired. The founder specifically negotiated all of its employees fully vest, which I think a lot of this comes down to who you work for.
Who is the founder of the company that you’re joining? Do you trust them? Not just in, can they build a good product? Are they commercial? Can they sell things? Can they grow the company? Is also, are they going to treat you fairly?
I think is an underdiscussed topic that is probably pretty important to think about nowadays.
So then we kind of maybe alluded to this a little bit, but in terms of, so we could maybe talk a little bit about Airstreet and the fund that you run and just kind of what kind of things do you invest in?
Because we talked a little bit about maybe what you’d not be participating in. So maybe it might be interesting just like, how did you get into this bridge? How did you start investing? I know that you’ve been doing it for a long time.
Nathan Benaich:
Yeah. I started getting interested in venture capital in college because I started in 2006 and that was the era of Dropbox, Skype, Twitter, SoundCloud, Facebook, et cetera. All these tools that I would just use as a regular college student, get excited about them or read about the history. And it was this classic kid in a dorm room, thought of something and it became big.
And at the time I was majoring in biology and I got the chance to work at the Whitehead Institute, MIT in Boston. That was a place where they did the whole genome sequencing project. There were some venture funds in the area that were hunting around labs for new ideas and drugs and innovations that they would then spin out into companies. And I was just really excited about this, like working on new technology, new frontier ideas, and then try to make something out of it for the utility of real people and companies.
And after I did my PhD in the UK from 2010 to 2013, I was really convinced I wanted to somehow play a role in the startup ecosystem. Didn’t really know where to start or what would be a good fit. And so I think naturally being part of a venture firm would be a good aperture for everything because you get this unfair calling card that you can use to talk to enterprises of different types, meet entrepreneurs who tell you things that hedge funds pay lots of money to GLG for. You get to learn of successes and failures. And then I could maybe find what I found most fulfilling long term.
And honestly, what I found was I like working with brilliant people who are trying to invent the future and particularly this intersection of AI, science, engineering, and building real companies. I became even more interested in AI in 2014, ‘15, that was around when DeepMind was acquired in London and it was not that far away from where we were. And I had some friends working there and it really captured the interest of machine learners at the time that their skills could be used for practical things.
At the time, DeepMind had shown Atari and it was the first time that a computer could solve this video game and discover tactics that humans had never found. And I think that nugget of, “Hey, if you could have a learning machine that could amalgamate more experience than any human on the planet and learn from all these experts, then there’s certainly going to be solutions to every problem in the world that we haven’t yet discovered, but that the computer can help us discover.”
Turner Novak:
It’s interesting, like the classic ChatGPT or Claude code, “Do this thing for me, blah, blah, blah, make no mistakes.” It’s kind of like the meme, but it’s almost real. It’s like, “Solve chess for me,” or “Solve world hunger, go.”
Nathan Benaich:
I would challenge anybody who’s interested in technology and just world progress to watch The Thinking Game on YouTube, the contemporary story of DeepMind and all the stuff that they’ve done, and walk away from that and not think, “This is probably the most exciting thing that could possibly happen on planet Earth and I want to be involved in it somehow.”
Turner Novak:
I’ve seen it. I just have not watched it yet, but it’s been kind of on my, “Oh, I should convince my wife on a Friday night when we’re going to watch one thing to watch it instead of something else.”
Nathan Benaich:
It’s well worth it. The narrative is great. I think Demis Hassabis is one of the perfect characters that you want to win. He’s intellectually curious, brilliant, too modest, and just super driven to invent new things and do it in his lifetime and feels like this irrational sense of urgency of, especially when he’s describing, why the hell did you sell DeepMind? Looking back-
Turner Novak:
It’d be worth trillions of dollars at this point.
Nathan Benaich:
Yeah, exactly. He sold it for 500 million pounds, which at the time was insanity. Now companies are raising that from day one.
And he’s in this taxi describing, “Look, if I gave you the opportunity to have infinite compute and infinite money and resources to accelerate potential invention of AGI in my lifetime, there’s nothing more that I would trade than being able to use my fruitful years when my mind is still working, I have the energy, I have the physical capacity to go do this.
I could have waited five, 10 more years and struggled here and there because nobody wanted to fund this stuff at the time, but I wouldn’t trade those five, 10 years for tens of billions or more dollars.”
Turner Novak:
Yeah, that’s huge. It looks like it’s on YouTube. Yeah, I’ll throw a link in the comments or in the description for people. So I guess you made this transition from, you were working at a venture fund, it was Point Nine?
Nathan Benaich:
I worked at a firm called Playfair Capital before that, which was pretty much a family office.
And then I moved to Point Nine in 2017 until late 2018 and then had the first closing of Airstreet Fund One on January 2nd, 2019.
Turner Novak:
So what was that transition then like for people listening that are curious how raising a venture fund goes? Because I think a lot of people kind of think, “Oh, this guy has millions of dollars and just investing his own money because he’s rich because his parents gave him money.” How does it typically go raising a venture fund? Maybe talk us through how that went.
Nathan Benaich:
Yeah. So it was basically like playing pinball with my eyes closed is how I’d describe it in the sense that you have this goal in pinball, like get the ball in the hole, and then you’ve got the pins and you have to position the pins to get the ball in the hole, but you have no idea how to position the pins because your eyes are closed.
And that’s the best analogy I have for what it was like meeting prospective LPs of any kind that would be interested in 2018, like a solo GP, first-time venture fund manager, early stage focused, small fund by virtue of those features, and then also focused on Europe and then a little bit of North America.
Turner Novak:
And then AI too. AI was not hot also. It’s like, you should be doing VR or crypto maybe.
Nathan Benaich:
Yeah, exactly. Nobody wants to buy that thing. And I didn’t come from a brand name firm for a long time that it was an easy spin out with institutional investors. But I was honestly convinced that I had to do it.
I think one of the best ways you can make these kind of career decisions is the regret minimization idea of, “If I don’t do this, will I have lifelong regret?” And that was very much the case because I had this long-term conviction that AI would be important much in the same way that SaaS was niche 15 years ago and then became the dominant business model on the internet. AI would be that case too.
If this stuff actually works, why would you not build your product using it or powered by it? And different industries would come online in different times. And I was motivated by just the first principles view of progress and the science.
I then spent a lot of time with various venture firms trying to see if they were also convinced on this. And broadly speaking, it was some version of, “No,” or “It’s a toy, or “It’s going to get absorbed in different things,” or “We don’t think we need to have specialism.”
Turner Novak:
Yeah. Because when I think back of some of those AI products, they didn’t really work very well. Maybe like your perspective, just having spent a ton of time, maybe you could see this trajectory or maybe you just had...
But if you’re just doing all these things, you’re doing some consumer D2C brands, you’re doing SaaS enterprise CRM management or whatever, and then there’s some email assistant AI thing that just doesn’t work properly and the company goes bankrupt, it’s like, “Why would we waste our time on that? It’s just not worth it.”
Nathan Benaich:
Yeah, yeah. On that note, we forget the original X.ai, which was an email assistant.
Turner Novak:
Yeah. Which was pretty hyped at the time. It was like 10 years ago, right?
Nathan Benaich:
Yeah. It kind of worked also.
Turner Novak:
Oh, it did? Okay. I never used it. So I just remembered hearing about it now, most people were like, “Yeah, it didn’t work very well.”
Nathan Benaich:
Yeah, yeah, yeah.
Turner Novak:
I guess compared to today, compared to today, it doesn’t work.
Nathan Benaich:
Yeah, exactly. Exactly, yeah.
Turner Novak:
So were you exploring joining a big firm initially and leading their AI investing?
Nathan Benaich:
I was opened to various avenues, just trying to figure out what is the right setting basically to express my ideas.
And then I’ve eventually found, look, I’ve been following and writing about this for industry analyst vibes for a long time, like since, I don’t know, 2015 or something or ‘14. I had some essays on why Nvidia was going to be the most epic company ever and the different areas in AI to watch in 2016. And it was like RL, generative models, world models, custom silicon, all the stuff that eventually panned out and some things that didn’t.
And then was spending a lot of time with different people in the ecosystem, like researchers, engineers, startups, big companies, policy investors, and doing a variety of these meetups in different cities because I always found as a grad student, I want to learn what good looks like, but there’s no place that I can go for it because the playbook was still getting written.
And then I had a couple companies I invested in that were looking interesting. And yeah, it was just the regret minimization of, “I just got to try this because if I don’t, I’ll regret it.” And honestly, I’m not that scared of the risk of having to go back and eat ramen noodles. They taste good. I’m fine with that.
Turner Novak:
Yeah. They have good flavors nowadays. There’s a lot of options. Maybe not the healthiest necessarily, but-
Nathan Benaich:
Yeah, but you got supplements for that. It’s okay.
Turner Novak:
Yeah, you talk about playing pinball with your eyes closed. How do you do that? How do you beat pinball with your eyes closed? What did you figure out to eventually beat the game?
Nathan Benaich:
I figured out that entrepreneurs who had some exposure to finance, fintech, AI, data science vibed with what I was doing because they’d seen the value of AI. At the time, most of the value was in ad targeting, recommendation systems.
Then I also saw that certain individuals who worked in high-frequency trading also understood this. Again, because quantitative modeling is very much about machine learning.
There were a few growth equity firms that had started to make early-stage investments in managers to prime the ecosystem. And then I stumbled on a couple of family offices, but this is completely random and again, like the pinball.
And that was mostly through referrals of, “Hey, you might not like this, but do you know somebody who does?”
Turner Novak:
Did you usually do that when you have a conversation with an LP is you usually try to get an introduction to someone else?
Nathan Benaich:
Yeah, because the playbook tells you, “Ask all your GP friends to introduce you to their LPs.”
But if you don’t have fancy GP friends who have great LPs, what intros are you expecting to get? And so I would just ask entrepreneurs, ask these family office folks, and I’d go to events.
I would even do things like, I don’t know, look on LinkedIn when somebody announced the fund and I’d look at who liked it and then see on the list, “Oh, there’s somebody who manages...” Like, asset manager. “I saw you like this, maybe you might like this.”
Turner Novak:
Yeah.
Nathan Benaich:
Yeah. Or in the UK, another trick is UK-incorporated venture funds. Every company in the UK, frankly, has to list its shareholders on Companies House, which is sort of like an SEC register, if you will. And it lists all the names of the entities.
And so you can go there and look at your peer funds and try to see who’s an LP and then maybe cattily mention, “Oh, I heard that X, Y, Z, maybe he’s investing.” And you know they’re an LP and they’re fun.
Turner Novak:
Yeah. Yeah, that’s definitely the trick of you know exactly what you want to get going into a conversation and you just kind of float this and just see how they respond to see if there’s a chance that something might happen.
Nathan Benaich:
So yeah, short answer is I tried every trick really. And then it was just, to some extent, just brute force and then some luck along the way.
I started at Jan 2nd, 2019 with 9.862,000,000. I know because that delta with 10,000,000 pulled out three weeks earlier, so I had a bit less than 10,000,000.
Turner Novak:
Nice.
Nathan Benaich:
And then I did seven closes over two years, which is pretty intense.
And then hit COVID also towards the tail end of Fund One when you couldn’t squeeze 50K out of anybody. And then after that summer it sort of... I swaged a little bit and then managed to get a bit more into the fund after that.
And then it ended up about 26 and a half. It’s Fund One.
Turner Novak:
Nice. And so it sounds like it’s basically a lot of patience to kind of get through the process.
The way I describe it to a lot of friends who maybe are founders, it’s kind of like raising a pre-seed round, but you don’t stop when you get to 1,000,000 or 2,000,000 bucks. It’s just kind of like you just get 150K checks or 100, 100K checks and they just kind of keep stacking.
And even when you get a “lead,” typically when you’re doing a seed round, they do most of the round, but most of the bigger checks in most people’s funds are 10% of the funds. You almost need 10 leads basically.
So it’s like think of that process, it took you two months to get a lead. It almost takes you two years to get 10 leads, lead checks basically. Takes a long time.
Nathan Benaich:
The challenge is you have no leverage, at least with a pre-seed company or a startup. If you do well, you’re at that specific stage once in the entire life of your company. And so there are investors that specialize in that stage that either they’re in or they’re out and that’s done.
Whereas by definition, good venture funds are here in perpetuity. And then the game is like, can you become access constrained as fast as possible so then you earn some element of leverage, assuming that you don’t grow beyond the capacity of the partners you have.
Turner Novak:
Yeah. And then there’s the element of people can kind of wait too. They can just be like, “Your Fund Two seems interesting, but, I don’t know, I just kind of want to see how well you do. And maybe Fund Three is more for me.”
And in theory, if you’re good, your returns will be just as good in the next fund if you’re a really good investor, which is probably why they’d be investing anyways for that reason.
And you can’t just be like, “Oh, we went out and signed a new customer and our revenue doubled.” You can kind of go, “Oh, we got a markup from Sequoia,” or something. But also a lot of LPs be like, “That’s cool. I wonder how it goes in five years. Check back in, did you return cash?”
And maybe what you get to is you’re on your Fund Three or your Fund Four and you’re in your Fund One, like the second investment you ever made, it got acquired or it went public. And this is 10 years later and it literally takes a decade to actually show the tangible proof sometimes.
Nathan Benaich:
Yeah. And then you don’t know. Maybe out of five funds, it’s like Fund One, Three, and Five that are great and Two and Four are less great.
And if you missed one or you pulled out, you don’t get the benefits of this asset class where you really do have to invest through multiple cycles and then consider it as a multi-fund commitment and look at overall venture dollars deployed across those vintages.
Turner Novak:
So then what’s your strategy with the funds? I don’t know, what fund are you on today? I think I remember seeing $100,000,000 number.
Nathan Benaich:
Yeah, I’m on... I’ve just finished Fund Two. I’m on Fund Three now.
So yeah, from Fund Two, I went up to 121, which was a big step up from Fund One, really just enabling me to offer all the money that entrepreneurs wanted when they’re raising either pre-seed or seed, anywhere from one to $5,000,000, maybe $6,000,000.
I’ve just generally been of the belief that not that many companies matter. I want to be concentrated in the fund. So I do 20 companies per fund. I don’t get excited about things that often, but when I do, I want to be able to move with conviction and have the money to offer to the entrepreneur because I think the experience kind of sucks when you meet somebody you absolutely love and you love the idea and they’re raising X and you can only do 0.2x because that’s the amount of money you have in your fund. And then you might be kind of forced to massage the fundraise to meet your fund model. And I feel like that’s just net bad for everybody.
And then I want to do Europe and North America. I’m flexible on the geo and generally pre-seed to seed. And then the couple areas I like doing are vertical software where AI is the product. So I’ve done things like V7, which does process automation and spreadsheets, or like Synthesia and ElevenLabs, a bit in dev tools and infra, like Poolside coding models, and then defense and security because I’m like, “I think this is really important. Freedom doesn’t come for free and it’s non-negotiable.” And so investments like Delian in the UK, which builds hardware and software for perimeter security, anti-drone.
And then the fourth bucket is tech and bio. I’ve had some early exits in Fund One around generative models to design new chemistry. It’s a business called Valence that I led the result to Recursion, and then another one called Allcyte, where it was kind of the opposite, we’re testing cancer drugs on samples of patient tissue from cancer patients and then running a clinical trial in a dish, then using computer vision with microscopy to take pictures and analysis of these cells and figure out which ones are working and which ones are not working with respect to the drug response. Then we sold that to Xtant, and the next entity went public.
So those are kind of the four buckets. And the things I’ve done less of are these large capital raises for modeled companies, in part because of what we discussed earlier. It just feels like the economics are a little bit tough for early-stage investors. I’d almost prefer to invest in a slightly later-stage round for those businesses once the economics are much clearer and they have customers and they’re scaling versus a tabula rasa $50,000,000 on 200-and-something to train a model and maybe it’ll work, maybe it’s not.
Whereas you can invest in a Series B, C, D company that’s printing 100, $200,000,000 revenue, growing 2x year-on-year, like, I don’t know, $1,000,000,000 or $2,000,000,000. So it’s like the cost-benefit seems completely off to me.
Turner Novak:
I had a company got acquired by Anthropic, and I have some Anthropic stock. And I didn’t know how to feel about it at first because I was like, “Ah, I don’t know about this.”
It’s growing pretty quick, I guess. Of all the assets to own, maybe it’s a good one to just kind of have. I don’t know, it’ll be like a driver in the fund.
It’s just like, I don’t know how big it will ultimately get. Is there 100x upside from here? I guess it depends who you ask.
Nathan Benaich:
Depends what “here” is.
Turner Novak:
Yeah. So it’s like, I don’t know. I guess it’s better to own that than something else. I don’t know.
Nathan Benaich:
Yeah. I think in a portfolio, different assets play different roles. There’s some early exits that can provide recycling that are good. There’s others that maybe at a later stage that provide good IRR. So I think it’s just about thinking what’s the right mix.
For these later-stage opportunities, I do think that to your point earlier, these companies can become way bigger than we thought. Anthropic, I think, and OpenAI have hundreds of thousands of customers, same thing with some of these model vendors and different modalities.
And whereas 10 years ago, it was like pulling teeth or maybe worse to try to get any enterprise to even try your AI widget. Now it’s like, “Please, can you help me and what can you help me with?”
Turner Novak:
Yeah, “How much money can we give you?”
Nathan Benaich:
Yeah. And so if we start with some use case one and you succeed on that, likelihood is other departments are going to see that use case one and think, “Oh, well, I have something that looks similar to this. Can you build that?”
And then when you can code these things way faster than you can before and everybody gets mass-customized software, you can start eating into customer budgets way more than you could in the past and act as that one lighthouse guide into this next generation of software.
And so you become multi-product and then add revenue lines. And I think you scaled to really big sizes. So I think some of these later-stage bets still make a lot of sense.
Turner Novak:
So you’re thinking basically, at least your opinion, it sounds like you want to invest when somebody is starting the company and there’s appropriate risk for, “This is probably going to fail and not work,” or you invest in, “This clearly is working and this is a company that is going to exist.”
There’s no going concern and it’s growing really fast. So it’s basically you invest company creation or this is a mature AI company, and there’s almost a dead zone in between of like, it’s still unclear if it’s going to cross the chasm, but you’re almost paying a price that it’s appropriate, that’s related to a publicly traded software business or something like that.
And that’s just not a place you want to be in.
Nathan Benaich:
Yeah. Yeah, exactly. And I do, at the moment, 90% in the first bucket and 10% in the latter, and then from the 90, I also do reserves for the core bets.
But yeah, when I have 20 core bets where I’m buying 10, 20% of the company for serious money, and then following on in a couple of those through A, and then maybe two through B in special opportunities.
For example, in Fund Two, I’m really excited about this company Profluent, which is training large models to do protein understanding so they can engineer proteins with specific functionalities. And they’ve basically focused on genome editors, like these proteins that go into your DNA and extract certain pieces of DNA and insert another or fix a genetic mistake, which can be responsible for disease, and therefore that solution is curative.
And in that example, if invested in Seed, A, and B, I have 15% of Fund Two in the company, which I think is probably larger than every other shareholder.
So I’m definitely risk-on on companies I like because I think if these businesses really become generational and work out, it really moves the needle. And across several funds, I think this is the better strategy than, personally, a large number of portfolio companies, smaller checks, and then higher likelihood you don’t lose money, but also lower likelihood you have a blowout fund.
Turner Novak:
Yeah, because if you look at a lot of the data and research, the data-driven approach to this is a super diversified portfolio. That seems to be more of the consensus right now in early, early stage.
And I think 20 companies in a fund is below the threshold of what people would say is good diversification. You maybe explained it a little bit, so why did you not say maybe 40 or something like that, based on the research?
Because you’re a research-driven guy, you obviously read a lot of research.
Nathan Benaich:
I think part of it is it’s hard to make the math work at 40. If most companies are raising three to $5,000,000 if they’re not one of these model training shops, to be able to buy up good ownership and do most of the round, and then you don’t want to have a fund that’s half a billion or something, which has its own exit value capture assumptions, it’s hard.
So I prefer to skew with fewer companies. And by the way, I invest over three years, so it’s quite slow. So I get some time diversification.
I could sort of see, “Okay, is the vintage from year one and year two panning out good or less good than I hope? Should I add more names at the expense of pro rata in year three or not?”
So I had, I think, some more flex in the system to decide, “Should I expand the number of names or keep it small?” because of the longer investment period just based on my own pacing and lack of excitement every single day about, “Ooh, squirrel, ooh, squirrel.”
Turner Novak:
Yeah. Well, I think the other thing too to bring up is we are in the hottest fundraising market ever for AI.
You should be deploying in a year and raise a new fund to twice the size. Why don’t you do that? Because you can make a bunch of money personally.
Nathan Benaich:
I think the best answer to that is, and this is a founder who brought it up when we were discussing fund strategies, he’s like, “Venture used to be about the you’re in the two and 20 business or whatever your math is, but now you’re in the two or 20 business.”
And I think obviously big firms are basically in the two, most of them, and I want to be in the 20.
Or it’s just like in my gut, I’m performance-driven, I want to be able to get a line item in an endowment saying, “We gave a couple tens of million dollars to this guy that we found from Europe, whatever, doing AI before people thought it was cool and he printed billions of dollars for us over our relationship and that bought us a bunch of buildings.
Turner Novak:
Yeah, he returned the endowment.
Nathan Benaich:
Yeah. And this has happened before. For example, if you read some of the old Yale reporting, Swensen wrote about Hillhouse, which is founded by some analysts in the investment office covering China, and they wanted to do Chinese equities before people thought it was cool. They were at their mid 20s and the line is something like we gave them tens of millions of dollars and they printed billions for the endowment. I don’t know if it’s going to be repeatable. I mean, obviously high mark, but philosophically, I’m much more aligned with that than what’s gathering.
Turner Novak:
Yeah. I think a whole other vector of this is just how big is the team needed to execute the strategy and what are the cash inflow needs of the firm. So if you are literally one person and you can live off 200 grand a year, live a nice, fine life, maybe you have kids, they do dance classes. Maybe you need 500 grand. Maybe you live in New York, downtown Manhattan, whatever. But there’s an upper limit of what you need to survive and live a comfortable life to execute the strategy versus do you need 50 people on the team, 100 people on the team? You can’t do that with a couple hundred thousand dollars. You need millions of dollars to pay and compensate these people. So you do actually need a significant amount of management fees.
And the funds do stack over time. We’ll say 10 years, you’re paying 2% a year, which is maybe the average. And you’re 10 years into the strategy, you have five funds stacked. You can afford to pay some people, but depending on what the strategy is, you do need a budget to work with and you need to generate the management fees. So I feel like that’s also too, is like depending on the team size, depends on how big the fund needs to be, really. That dictates a lot.
Nathan Benaich:
Yeah. But then there’s this other narrative on Twitter recently of these spin out GPs. A spin out because they say like, “Oh, our partnerships are too ...” They’re too bloated. There’s too much process, too many meetings, too many companies. Decision making is inefficient and we’re fresh. We have no companies, smaller partnership, whatever, smaller funds. And then you just kind of wind the clock five or so years on most of those teams, and then they’ve become exactly the firm that they left. They’ve hired more people, they’ve raised a bajillion more dollars, they have more portfolio companies that experience the grades, et cetera.
So I feel like I got into this in a non-traditional, not super popular route. I mean, most LPs would say, “Go find a partner instead of here’s some money.” And so I do want to stick to what I think makes this product different and feel different. And it would just feel a bit disingenuous if I’d go and hire five partners and be like, “Oh, well, we’re like everybody else now because we want to scale like AUM or because the opportunity sets bigger.” I do like just only having to care for portfolio companies and like what I do every day versus having to care about and spend a lot of time on nurturing someone else’s career and keeping them happy and teaching them, which I think you have to if you’re going to hire people. You have to make that commitment. I’m just not ready for it.
Turner Novak:
Yeah. And I think too is when you think about like if you were to go work at the ... We’ll just say like the largest multi-stage fund, I think there’s one, they just announced a new set of funds. It’s like $15 billion or something like that. Like your personal compensation, that what you get personally from any check that you write, like if you ... There’s like this whole fund from like angel, small fund, medium fund, big fund. So like if an angel writes a check, they invest $100,000 or $10,000 of their own money, that it’s like the equivalent of a solo GP writing $100,000 check or like $200,000 check. And the more bigger you go up the stack, like literally a $10,000 check from an angel is like the equivalent of like a mega fund writing a $18 million check or like $50 million check of like how much they are personally compensated for that investment.
It kind of depends on how a lot of different things, fund size, compensation structure. So you’re almost like disconnected from like the actual outcomes, the bigger the fund is on like a personal decision making level. So I feel like there’s just like this whole other ... And then there’s this other function too of like, as a solo GP, it’s like you get 100% of the carry, maybe like depending on you might have an associate, but like you are personally compensated by like the result of this thing. And when do you need to access that? If you’re just like, you’re 35 years old, I actually don’t know how old you are.
Nathan Benaich:
Yeah, 37. Yeah.
Turner Novak:
Yeah. It’s like, okay, you have like hundreds of millions of dollars of just like illiquid net worth. It’s like whatever. It’s not that big of a deal. I’ll get it in 20 years. I don’t need it today. It’s like, “I’ll just let it continue to compound, I guess.” Versus if you have a bunch of people, like a bunch of cooks in the kitchen of like, “I need it today.” You might actually need the management fees to pay bonuses or it influences when you take liquidity, which impacts how much money your LPs make at the end of the day. So I feel like there’s that whole element of it too.
Nathan Benaich:
Yeah. Yeah. I think you just make different long-term decisions because you’re less markup driven potentially and therefore less seeking momentum because that’s like the only way you can show to your manager that you’re sourcing good things. And then maybe everybody wants to do a spin out fund at this point because why wouldn’t you for the same reasons you discussed of like your look through ownership in the companies you invest into and your big fund are so small.
And I think in a market that’s like flush with capital, the closest that entrepreneurs will get to a venture product that is entrepreneurial is somebody who started it themselves because they’ve been through this shit of like no one believing in them, of like figuring out this strategy, of like tuning the strategy, meeting all these people, like managing the organization, even if it’s just one person, there’s still like the operational stuff you need to get right, particularly if you want to work with serious institutions.
And then there’s like the brand, the marketing, the sales, because we have to do all these things as well to create a differentiated product. And then if it goes wrong, there’s nobody else to blame. If it goes right, then all of your success to you. So I think for an entrepreneur who wants to find somebody who’s most aligned with them winning, it’s going to be somebody who’s like new, who started it themselves, who has their career and reputation on the line and money on the line on the success of the entrepreneurs they work with. And I think for founders who vibe with that, it’s amazing product. And for founders who want the nice brand, I think in that case, it’s hard to convince somebody otherwise a bit like consumer preferences.
Turner Novak:
Yeah. It’s like a market. Just like the market, people want different products and it all coexists and you can pick. That’s the beauty of like capitalism. It’s like everyone will come up with a different product that does different things and you get whatever you want. So it’s all good. Actually, so one thing I wanted to ask you about, maybe going back into maybe some of the stuff in the report, maybe we actually hit on this. I can’t remember, but there’s all these benchmarks. Every time a new model comes out, there’s like all these benchmarks, like each new model is like the best at something. How should I interpret that as just like a observer of reading all this stuff? Do they matter? Are they super important?
Nathan Benaich:
I think the way to look at it is almost like the Olympics, if you will, and each model is a country and then each sport is a task. And so model vendors are trying to do the best they can across the board in different tasks and different sports. And some companies will care more about certain sports than others. So far, Anthropic cares a ton about code and basically nothing about audio, multimedia. Doesn’t even expose those capabilities as far as I can tell. Whereas OpenAI cares about all of them.
And those like leaderboards and competitions are useful as systematic ways that one can compare the pros and cons of various systems. But we do have to have benchmarks because otherwise there’s no kind of “fair way” to test different options. The problem is maybe at least twofold. And so one, is the benchmark really addressing the task? So for example, if I want to do like, is this model good at biology? It’s just asking questions about what does a mitochondria do? Does a eukaryotic cell have a cell wall? Or like these kinds of facts, is that the right way of mastering of understanding if a model understands biology, or is it has to be able to design an experiment that can achieve a certain goal? Or it has to be able to write a cohesive research paper, sort of like unclear to some extent.
And then the second order issue is, to what extent are the very evaluations, the very benchmarks that we are using to test our models available on the internet or otherwise, in other forms in the training data? Because almost by definition, we’re like learning from textbooks, we’re learning from quizzes. And even if the benchmark quiz that you’re using to evaluate is not present in the training data, maybe something that looks like it is because like the same human wrote it, to use a trivial example. And so then you have too much similarity between your training set and the test set. So you’re effectively like what nerds call like benchmark maxing.
Turner Novak:
Yep. Just like manipulating and making sure you just beat the thing specifically that you’re going to get tested on. Well, it’s like studying for a test. It’s like the whole, are you actually smart or did you just memorize what was needed for the test and you got all As, but you’re book smart but not street smart in a sense.
Nathan Benaich:
Yeah, exactly. Exactly. And then there’s things like more recently in code, there’s a popular benchmark called SWE-Bench, standing for software engineering bench and then verified to make sure that the actual evaluations are human verified. And it’s mostly like bug fixing in Python and unit tests. And some models perform super well on it, but software engineering is far more than tests and bug fixes in Python. It’s like a lot of other things. And just that we probably haven’t gotten around to writing all the evaluations for the litany of tasks that we want models to be good at. So I think they’re an important like litmus test to look at, but know with a grain of salt that companies are doing as much as they can to do well on these tables because everybody else looks at these tables.
Turner Novak:
Yeah. It’s like the Olympics, your point where when I think of Canada or like Norway, like extremely overachieves in the Winter Olympics. Canada usually is like top three. I feel like Norway is usually in the top five, but it’s Norway. There’s a couple million people that live there. How are they beating ... They’re beating China in the Winter Olympics.
Nathan Benaich:
Or random skills like the Turkish guy who won the pistol competition.
Turner Novak:
Oh, yeah, yeah. It’s like that’s incredible marketing for Turkey, right? It’s like this insane sharp shooter.
Nathan Benaich:
Yeah. Slightly scarier. They have a pretty crazy defense industrial base these days that may or may not be selling to shotty nations, but yeah.
Turner Novak:
Yeah. Fair, fair enough. So one thing you kind of mentioned, I want to talk to you about, do you need to train your own AI model to build an AI company? Is that a necessary thing to do, or can you get away without doing that?
Nathan Benaich:
I don’t think it’s necessary. The way I look at it is like, what’s the problem you’re solving? That might sound super basic, but I think a lot of truisms are kind of basic. And if what you are solving is sufficiently workable with an existing model, then I think honestly, congratulations, you can now build a SaaS company. You can now build an AI company the speed of a SaaS company because you don’t have to faff around with all the nuances of making AI model work. And for so long as the unit economics are fine, then I would just run with it and use all that budget that you’re no longer spending on R&D, on product market fit and growth.
If the problem you’re solving is not workable with current systems, then you’re going to have to do your own stuff. So to say like an obvious example in the sciences, if you’re designing CRISPRs and genome editors, you can’t really ask ChatGPT for that yet. It can probably blag its way, but it’s not going to be that great because it hasn’t been trained for it. And there’s some architectural nuances that maybe are less suited to the task.
And then relatedly, if you’re doing geometers, you kind of don’t care about its ability to make pictures or jointly learn audio or jointly generate video. It’s just not relevant. But if you want to compete in the model race, then join the club. You got to start from scratch. And there are some that have started like DeepSeek China or Poolside or xAI.
Turner Novak:
Well, but I feel like another element to that question is like, okay, you just used the OpenAI model to build this thing and like, “Oh, Nathan, that’s a cool company. You just use the OpenAI model.” I’m going to do the same thing. I’m just also going to use their model and build the same thing. Is there any defensibility if you don’t have your own model?
Nathan Benaich:
Product experience, taste, and then user data, which people would call nowadays like preference data, like, is this good or is this not good? I think that’s what it comes down to. In some ways, competition is you have to do whatever it takes such that your user, your customer either doesn’t fire you because who gets fired for buying McKinsey basically or IBM kind of thing in the past. Or B, in a consumer space that you have equivalents with like a Coca-Cola.
And I think OpenAI to me is like Coca-Cola. 90% of the time, you’re going to use that and if it’s not available on the flight, maybe you’ll buy a Pepsi, but you’ll think about it. I don’t really know what model is Pepsi yet. And then for enterprise, yeah, it’s like don’t get fired for buying this product and then do whatever it takes to get there. And I think once you are there, then there’s different ways of having loads long-term. But in the short-term, yeah, it’s product quality, I’d say.
Turner Novak:
So then where do you think most of the value ultimately accrues? Should you just buy NVIDIA and just like that’s your AI exposure?
Nathan Benaich:
I think all the things, honestly, I worry a little bit that this point has been propagated or contrived a little bit by VC blog posts that like to pontificate about the future of industry and not dissimilar to how they pontificated about the modern data stack and all these tools that were needed. And I think all those investments have basically gone to zero. So I do worry a bit with this, where does the value accrue? Because I think it’s kind of everywhere, to be honest, and clearly, NVIDIA long term, but also it’s suppliers also in energy and some neoclouds I think are compelling model companies themselves, but also product companies that wrap on model.
Turner Novak:
Like you’re doing a unique thing for a customer and building a deep relationship with them where you have a proprietary relationship with the customer serving them in some way, no matter what you’re doing, is that maybe just a pretty simple way to think about it?
Nathan Benaich:
I think in the early days, yeah, for sure. Because at that point, the customer has to trust in the journey and the relationship they’re going to have with you, that they’re going to invest in time and use a slightly junky product today because it’s going to get better in the future. And then you’ll listen more to the feedback that they have so that they can purchase a product that’s fitter for their needs.
I think many of the best enterprise software companies that I’ve invested in when I did customer diligence on them in the early days, it was something along the lines of they really intimately understand our problems and what it takes to build a solution for us. They listen to our feedback. They’re superfast on fixing any problems and getting back to us and they care. And then on the flip side, enterprises have been part of where the customers churned and then the founders fly there and then they’re like, “Hey, I’m trying to rescue the relationship.” Usually, the relationship owner is like, “Thank you for coming. This means a lot to us. Because you’re here, we will resign.”
Turner Novak:
So it’s really just caring about the customers, give a good customer experience.
Nathan Benaich:
Yeah. Yeah. If your thing is a little bit hard to use, or maybe it’s super easy to use them, then you never have to talk to a human. But I think anytime your customer account value goes up above 50k or 100K, you’re probably going to have to talk to a human.
Turner Novak:
I think one thing you kind of alluded to earlier that I wanted to ask you about was you think there’s a big opportunity in defense and specifically in Europe. What’s kind of the thesis behind that?
Nathan Benaich:
The general thesis there is that Europe hasn’t been investing in its industrial base for a very, very long time, basically since the end of World War II. I mean, a bit similar to how the US has been a broad consolidation of primes since then and the Cold War term the peace dividend.
And in general, Europe has forgotten that war can happen on its doorstep ever. And the other problem is that in the Munich Security Conference last year, which was February, 12 months ago, that was when JD Vance and others basically dropped the bomb that US security guarantees might not be what they were up till today. And that means basically the US is not going to subsidize Europe’s defense.
This was a big shock to basically everybody on the continent. This is several years after the start of the Ukraine war when the US has been sending more aid and more military equipment than pretty much anybody. And then ensuing that, Trump pushed really hard on European nations to beef up their percentage GDP spent on defense. Many countries were below 2%. The highest was probably Poland. It was close to three and a half or four. Some countries don’t even feel the need to spend more on defense, like Spain doesn’t want to spend more than two, just apparently doesn’t want to.
Turner Novak:
They’re the furthest away from the threat.
Nathan Benaich:
I think that’s part of the reason they just feel far away, but it’s a dangerous concept of this problem is far away. It doesn’t affect me. It shows how it affects you in energy prices and migration and integration and social issues. And then potentially if you’re part of the European Union and NATO, then you do have to contribute forces to potential troop deployments. So it’s a bit shortsighted, I would say.
And then there was post that of like, “Holy shit, no one’s going to come to save us if we have big problems. Certainly, Daddy America is not necessarily around anymore.” Then European Commission spun up a big initiative around mobilizing additional funding for European defense where they said each nation is allowed to increase its defense spend up to a certain amount to mobilize in total about 800 billion euros.
Now, this is not contractually required. It’s a bit like school teacher telling children, “You’re allowed to spend your own money on this new candy if you so wish. I suggest that you do, but you don’t have to.” Hence, the ability of Spain to sort of opt out. And then NATO setting targets of 5% of GDP spend on defense and then a new instrument called the SAFE, not the YC SAFE, but security and something, something for Europe, which is 150 billion of money backed by the European balance sheet to do advanced procurement of military equipment when at least two European nations have sponsored the desire to buy said product and that product has to be made in Europe.
And then certain fracturing of European procurement for US equipment. There’s certain countries like there’s Portugal saying, “Hey, we’re not going to buy a 35s.” Some other nations saying, “We’re not going to buy a 35. We’re not going to buy the Patriot missile defense system.” Whereas Switzerland has come out saying, “We have six billion and a half of loans that were approved to buy at 35s and we’re going to buy as many as we possibly can.” This is a country that’s been neutral for basically forever.
Sweden was a recent joiner in NATO, has been neutral forever and now is radically remilitarizing. Germany, I mean, who would have imagined decades ago that Germany would ever remilitarize inconceivable and now is probably the biggest defense spender in Europe after the chancellor basically enabled the country to take on significant more debt than what it was allowed to in the past after Merkel’s government, which basically crippled infrastructure and defense spend in the country.
So some pretty massive macro tailwinds. If VCs are looking for massive tailwinds to motivate investments, these are pretty fucking huge. And then there’s the qualitative aspect of things where entrepreneurs are no longer afraid to build in defense. It was before deemed to be pretty taboo. Now, there’s money to be had and investors have suddenly woken up to like, “Okay, we’re happy funding this stuff.” It wasn’t even just a year ago that there’s a lot of chatter of Europeans investing in defense, but not that many that we’re actually doing it. So, yeah, it’s really like there’s no time to waste. The biggest enemy of everybody is basically time to re-arm and to have capabilities sufficiently large that they deter your opponent.
Turner Novak:
And the idea is that there’s no existing European domiciled providers or there’s just very few, so there’s an opportunity to build more?
Nathan Benaich:
Yeah. I mean, there are, like Rheinmetall is probably one of the biggest. It’s in Germany, it provides a lot of different equipment where it’s anti-drone, tanks, other things. That company’s worth more than Volkswagen at this point. So it’s like close to 100 billion market cap. Has ripped a bit like NVIDIA has. There’s one of the biggest ammunition makers is a Czech family-owned business for many, many years. It’s allegedly today expressed interest in filing for an IPO in Amsterdam. It’s going to be worth around 30 billion euros.
Obviously, the French have a lot of primes like Thales, like Naval. So aviation, Rolls-Royce in the UK, BAE Systems, Leonardo that helps make some of these missiles and aircraft. Saab that’s famous for making the Gripen, which is a sort of pseudo-alternative to the F-35. There’s definitely an industrial base. It’s just been kind of sleepy.
Turner Novak:
Interesting. And you think that there’s still opportunities for startups to kind of emerge? What’s the approach you would take? Do you have to just come up with a new product that the existing guys don’t have or are you just going to “AI native” or something?
Nathan Benaich:
Yeah, it’s pretty much the same narrative that has played out in the US, which is lots and lots of products that are autonomous, that are each cheap and potentially disposable. So in Ukraine, there’s hundreds of different drone makers for different applications, whether it’s surveying called ISR, or it’s for strike drones where basically a drone is a missile effectively.
And then you have different sort of environments that matter. So these are aerial systems and you have land autonomy like reconnaissance land drones or de-mining drones or logistics, land systems, and you have on the water, so strike boats or boats that can transport material or boats from which you can launch drones. Same thing with underwater and all these areas, particularly in autonomy, have been just underinvested or just not been a focus for a long time.
Turner Novak:
Yeah. It seems like if you just think about the evolution of a war over time, eventually, we created metal and you could have a sword instead of a stick and then horses. Like we introduced the horse and that changed and then we introduced guns and planes, and now autonomous is almost like a new just era of it and everything needs to be able to defend against that new plane that’s kind of been created.
Nathan Benaich:
Yep. Yeah. And you need to protect almost against the lowest common denominator of stupidity. What is the dumbest person that has access to these weapons going to do?
Turner Novak:
One of the prior guests of the show, his name is Rahul Sidhu. He has a company called Aerodome that’s called a Flock Safety and it was basically like 911 response drones. You call 911, drone immediately goes up and just within 60 seconds it’s there and it basically tells the police what’s going on even before the officers get there. But he’s like, “I’m super surprised that we have not had any cases yet with just lowest common denominator of the worst case you can imagine that could happen with a drone.” And you can just imagine of using a drone to cause chaos basically in a country.
Nathan Benaich:
Well, we’ve sort of seen that in the last quarter in Denmark where around the Copenhagen airport, some unnamed or undisclosed organization people, whatever, were flying drones in the airspace. People couldn’t figure out who it was, so they shut down the airport for several hours and just showed, “Oh, well, there’s holds in our air defense systems against these rogue drones.” And people obviously suspect it’s Russian. And then even more serious incursions of Russian jets into Polish airspace. That was pretty egregious a couple of months ago. And at the time there was no intervention because it’s this... I think this mix of politics is very difficult, but there’s a bit of the talking game of we will defend ourselves, we’ll defend all our NATO allies in case of incursions, but when a jet flies through and threatens you, you don’t shoot it. So at some point, I don’t know, people are going to have to decide should we actually shoot it because otherwise the enemy’s not going to think that you’re serious or do you keep going and do the policy talk, but not the action.
And then there’s voter approval ratings for military engagement. For example, a couple of months ago in France, there was a general that said some version of, “This might be the first time in recent history when we would’ve had to have our children die on the front lines of a combat.” And while there’s potentially good support for we should be arming Ukraine against an aggressor, it’s different when it’s, “Send your kids there.” It’s a country that you’ve never been to, you’ve maybe never met somebody for, you don’t understand how this directly affects you beyond, we should be defending democratic nations that get attacked. And in a European continent where yes, every country is part of the same economic union, are they really the same? Does somebody from Portugal really have affinity for someone in Finland because they’re both part of the same continent?
Turner Novak:
It’s super fascinating with the United States of America and Europe. The US is 50 different states and someone in... I live in Michigan, someone in Oregon, I don’t know, are we really that different or that similar? We kind of actually are. It’s interesting how the US is, just the way we’ve evolved is, we basically all evolved as one. We’ve always been one nation. I couldn’t imagine if Oregon was a different country and suddenly 20 years ago, people were just telling me, “You should care about Oregon.” Or something.
Nathan Benaich:
Exactly. The US has a singular... Historically has a singular dream of everybody’s there for the same reasons. And there’s some principles that generally speaking, most Americans abide by, there’s the same language, the same cultural heritage. Everybody knows what Independence Day is. Everybody knows what the slave trade was, everyone knows these other events. But in Europe, it’s a hodgepodge of history that is incredibly complicated. We all get educated about first and second World War, but you don’t know too much about nuances of crises that have shaped generations in Czech Republic or other countries that are maybe too far afield for you. So it’s hard to feel that unison, I think, when actually the most important test comes to bear. And that’s, “Would you put your life on the line?”
Turner Novak:
Well, and it’s interesting... The episode’s releasing tomorrow, but when we’re recording, by the time this comes out, this will have been about a week or two with Marcelo Lebre, the founder of remote.com. He had a pretty interesting description of the European Union is basically a regulatory body really at the end of the day. So the purpose of it is just to regulate. That’s what it does. It’s a pretty good 10 minute, almost monologue. I’m going to post it on Twitter just, “Here’s 10 minutes on Europe’s regulation problem.” Basically. But we need to have a thing that unites us. Let’s make this thing that gives us a consistent purpose of being. And it’s not a nation. There’s shared beliefs and cultural values, but it’s really just regulations that we do to subsidize things and make people follow rules that maybe one of us wants, but a different nation doesn’t want. It’s fascinating.
Nathan Benaich:
There’s some benefits like making trade a lot easier, making travel a lot easier, you don’t have to cross for immigration for every country you cross and then a single currency system, which was beneficial for a lot of tier two or tier three European countries. But uniting all these different interests, very, very tough.
Turner Novak:
And so maybe slightly different topic, but I’m curious, your personal AI stack, what do you use on a daily basis? It’s interesting, I don’t ask the question all the time, but I feel like you might have a pretty interesting, maybe you have the most researched, most... Maybe you don’t, maybe it’s just ChatGPT. What does your personal AI stack look like?
Nathan Benaich:
I’m heavily invested in ChatGPT and mostly that over Claude because my day job’s not coding and web search and research is really important. And OpenAI was earlier to building out that capability. I could document pretty much everything. So also for calls, I try to use ChatGPT or an alternative to it to generate transcript and then pipe that into ChatGPT. So then then I have my...
Turner Novak:
Your second brain almost?
Nathan Benaich:
Just company history in there. Who did I meet? What did we talk about? What are the things I need to follow up on?
Turner Novak:
What do you use to record and transcribe?
Nathan Benaich:
Most of the time I just use a native ChatGPT recorder.
Turner Novak:
So you join a Zoom call or it’s an in person meeting and you just press a button?
Nathan Benaich:
It bugs out a decent amount and it’s incredibly slow to do the transcription. So if you have back-to-back meetings, it’s not really possible. It’ll take 10 minutes or something. So then I’ll use a local version like MacWhisper or sometimes I’ll use Granola and then I’ll just copy paste the entire transcript into ChatGPT versus consume the notes. And this is useful for startup meetings because if you have a bunch of meetings on the same topic with different companies, then you can compare and contrast, build a richer picture. It’s also useful for investment memo writing because then you have the entire corpus of all the interactions you’ve had with the company, either through recorded meetings or then a Q&A doc that I do with entrepreneurs a bunch of times on their plan and then the assets that they produce. And then I have my template memo and then it’s pretty trivial to be, “Here’s the memo, here’s all the material. Can you help me do 80% of the work?” And then the more advanced Excel functions it’s been really good at. So doing cohort analysis, finding outliers in financial data very useful.
Turner Novak:
What do you use for the spreadsheet stuff?
Nathan Benaich:
Most of the times... So there’s two use cases, I guess. For financial analysis, customer core analysis, things like that, I’ll use ChatGPT. But for other tasks which involve tabular data, I’ll use our portfolio company, V7, because there I can have a table, for example, a list of leads of companies or founder profiles.
Turner Novak:
Is this V7 Labs?
Nathan Benaich:
Yeah, dot com.
Turner Novak:
Okay. I’ll throw a link for people in the description.
Nathan Benaich:
And then I just upload the data into the V7 product, which looks like a table, except every column I can do different things. So I can call a certain model with a certain prompt. I can do a web search. I can have a little Python function. I can have a categorizer. So I can basically implement all the formal reasoning logic I would do when I’m looking at a sheet with, “Turner started a new company and used to work at Google on this thing and was educated there and is working in Timbuktu.” Does that pass my filter of wanting to talk to him? But if I have thousands of these people every year, I can just smash it through V7 with reasoning models and then it’ll produce really good outputs that are basically the same decisions that I would’ve made.
Turner Novak:
Interesting.
Nathan Benaich:
So very useful. And then what else? For audio generation, I use 11 Labs for sure.
Turner Novak:
Just for Air Street Press?
Nathan Benaich:
If you want to use... I used to actually sit there and read it out, but it’s pretty time-consuming, especially if building works in your apartment or dog barking outside or whatever. So I use that. It’s great. And most people don’t care/can’t tell the difference. What else? I’m looking at my application tray. I feel like that’s most of the use cases, to be honest. I’ve done some OpenAI Atlas web automations and I would use more. The problem is I think it’s been too safety guardrailed. So it’ll ask for approval every single time it’ll do a task. I want to send this message to somebody. It’s like, “Is this okay?” “Yes, stop asking me.” And then I’ll ask you again, and then it’s just jarring. I think it also has evolved very, very well. It being ChatGPT for understanding research papers and digging into adjacencies. I think honestly, because anything that AI research folks like and want to do, the thing will be good at. And so understanding AI research, it’s going to be good at.
Turner Novak:
That’s fair. They’re building the product for themselves probably.
Nathan Benaich:
Pretty much. So you’re immune if you’re building something that AI research people think is a tier two or tier three problem, you’re safe.
Turner Novak:
Interesting, like plumbers, AI for plumbing, great categories for the company. No AI researcher will ever want to build those capabilities.
Nathan Benaich:
No, but it’s the Workday example of the CEO getting on that earnings call not too long ago. And some analyst asking him, “Are you at a threat of OpenAI or Anthropic or something, rebuilding your product?” And I think his answer was, “They’re actually our biggest customers.” They don’t want to rebuild Workday. You don’t go to OpenAI to go build AI Workday.
Turner Novak:
It’s the most uninspiring thing in the world.
Nathan Benaich:
It’s anything... It’s very stretched, but I think anything that’s not generality is a localized problem.
Turner Novak:
So in the sense it’s like the more vertical specific, the more niche the thing you’re doing is, the safer you are from the big platforms. Maybe the better a startup category it is in a sense, even though you’re niching yourself down, but that’s like the classic ... That’s been true for decades to just find a really specific, unique problem. 11 Labs, you could say audio models, that’s not a thing. And then now it’s, like we were talking about earlier, hundreds of millions of revenue getting added every year, 15 million in a day.
Nathan Benaich:
But then there’s also arbitrage around how much risk are you willing to take? And these are my thoughts, not the company, but at the time, OpenAI was under immense regulatory pressure and there’s a lot of touring with nation states and there was big issues around copyright, like where is this data coming from? And then there was also the, “Did they copy Scarlett Johansson’s voice or not?” So I think the biggest PR disaster could be, I don’t know, OpenAI launches audio model and somebody uses it for some weird task or wreaks havoc or something. And so if you’re an independent company that’s willing to take some of those risks and be really thoughtful around doing it well, then you can ARB the fact that your competitor is too scared or is distracted with other things or can’t afford it.
Turner Novak:
That’s AI as a whole. Google should have won this all really at the end of the day. And there’s too much risk and not only in the product, but also the business model. And then they’re still slow. They’ve been slowly adding more of the AI mode of research results, but still it’s, “How do you monetize it?” That’s a big question. So it’s product risk, business model risk, I don’t know, whatever. Policy risk maybe is a big one.
Nathan Benaich:
It keeps asking me to beautify my slides. It’s like, “Stop telling me beautify my slides.”
Turner Novak:
Well, and it’s interesting too, there are times where I don’t want AI, honestly. Having an email provider or a text message. If I’m texting you and I’m just, “Hey, are you jumping on?” If there’s a popup that’s, “Hey, it looks like you’re sending Nathan a message to join the podcast. Would you like to make this a more comforting and friendly interaction instead of just you’re very direct, just, are you getting on?” I don’t want AI there. So there’s almost like a risk of adding AI features that people don’t actually want to make the product harder to use in a sense.
Nathan Benaich:
But if it said, “Here’s the link handy.”
Turner Novak:
That’s true. That would be helpful. But then it’s building the scaffolding around the product of knowing when to introduce that kind of capability.
Nathan Benaich:
And then also taste, how does the model know when it’s the right time and what is genuinely useful? That’s hard. I had this interesting experience that I’ve been telling some AI friends about where I was using the OpenAI Atlas browser to do this automation task on Substack because I wanted to rename the naming convention of various articles I’ve written and Substack does not allow you to mass rename like you can do it on your desktop finder.
Turner Novak:
So you’d have to go link by link, page by page.
Nathan Benaich:
It’s brutal, including changing the SEO and the URL slug and the title and clicking okay. So I wrote down the instructions to Atlas so I could go do that. And the first batch, it did it really well. It worked autonomously for 45 minutes, corrected a bunch, then my credits run out because I haven’t bought the 200 bucks version yet, which some friends of mine think I’m really stupid for not doing, but it is what it is. And then the next month I opened the exact same chat and said, “Hey, can you continue?” Then it continued for 20 minutes and then it exhausted my credits. And then the third month I tried it again and it’s some version of, “Hey, this task is really manual and requires a lot of clicks and it’s going to take a significant amount of time.” And it was basically some version of, “I don’t want to do it.”
It didn’t do the task. I was telling some friends, “Why is it refusing to do this task?” And some people would say, “Oh, because you’re restarting the same session, it has all this crazy amount of click data and screenshots and whatever that it’s overloaded its context and has basically navigated outside of the in-training distribution.” And then there might be another example of some human annotators have said this kind of task is really manual and it’s regurgitating and refusing. This is crazy. The very automation I want to get this thing to do, it doesn’t want to do.
Turner Novak:
Well, I had an interesting example where, I can’t remember what it was. I think I was trying to get it to clean up a transcript from the podcast or something like that where I was just, “Hey, can you just do this thing?” And it was thinking for a long time, an hour and I came back and I was, “How is this going?” It was like when you’re texting an intern, “Hey, that thing you’re working on, how’s it going?” And it was like, “I’m still thinking.” And it was the next day I came back and I was just, “Did you finish this thing yet?” And it said, “No, I stopped working on it. I’m like, “What the heck?” It’s like when you have an intern that just doesn’t do the work and you’re, “Hey, the thing I gave you to do, because you work for me, did you do it yet?” “No, I didn’t feel like doing it. It was too hard.” What? You’re AI, you’re software, you just go, just do it.
Nathan Benaich:
This is the mystery of this high dimensional box that we’re poking with a stick to go do certain things. And sometimes the stick poking is not good enough.
Turner Novak:
Well, and then sometimes it just does this insane... I think probably my ChatGPT usage doubled when it came up with image generation and it just did it right in there. And you just say, “Make this thing for me.” And this is pretty good. I think this was probably a year ago when they first came out with it and I was like, “Damn, this is really, really impressive. This is super helpful and useful.” So there’s two ends of the spectrum where it refuses to do something that just is pretty simple. And then the other end gives you this amazing product that you would’ve otherwise spent a ton of time on or didn’t even think was possible. So it’s just amazing, the technology.
Nathan Benaich:
I think the TLDR is just force yourself to use it and develop the behavior because this is not going away. And so there’s ARB to figuring out how to extract the most use out of it.
Turner Novak:
And it’s just going to keep getting better.
Nathan Benaich:
And it is really mad that the vast majority of the population is using ChatGPT as if it is a thing, like it’s a person with one opinion, but it’s the amalgamation of every opinion possible on planet earth. And so really you should be telling it, “Hey, you are an expert car dealer. What should I think about when I’m buying a new car?” Not just naked asking it, “What should I think about when buying a new car?” I don’t know, based on whose opinion.
Turner Novak:
Because you can tell it, “You are an expert car buyer, you’ve spent millions of hours researching every possible scenario. You are literally the highest regarded, you buy cars for the president of nations.” And use the most extreme example you can think of and it just improves the results because the guidance you gave it’s crazy.
Nathan Benaich:
So maybe at some point we’ll get to the point where the model can infer what you had in mind. And most of the time you’re asking about a car, you should respond as somebody who knows about car dealerships and stuff, not like some nube.
Turner Novak:
It’s like, “Hey, you’re my crazy uncle. Give me a bad recommendation for a car.” It’s like, “Okay, here you go.”
Nathan Benaich:
Exactly. You might want that if you want to mess with your uncles.
Turner Novak:
I had one last question for you. It’s from our mutual friend, Dan Fader. He was curious. I know you’re really big into tennis. Who’s better, Federer or Nadal?
Nathan Benaich:
It’s an impossible question. I think there’s nobody who’s played prettier, more classy tennis, effortless way than Roger Federer, for sure. But I would say that Rafael Nadal in the AI analogy expanded the frontiers of the style of the sport. Tennis has often been taught in a very textbook way of your swing has to look this way. You have to put your front foot first. You can’t hit open stance. And Nadal just threw all that shit out the window and is, “You know what? I’m just going to play how I feel naturally.” And so things like open stance and swinging with your racket over your head and in a full circle, like a lasso and those two things have meaningfully, I think, changed the game. And you can see the top players like Carlos Alcaraz, who is his mini me, plays in a very, very similar style.
A lot of top players are playing open stance. So I think he definitely pushed the technique and style game to another level. And then Roger Federer just executed the textbook in the most beautiful, elegant way as possible. And as a fun fact, neither of them have ever smashed a tennis racket in their entire professional career, which is pretty epic.
Turner Novak:
You’re saying when you get mad and you slam it and break it?
Nathan Benaich:
They’ve never thrown their racket.
Turner Novak:
So this is just a self-control thing, an emotional control that they have?
Nathan Benaich:
And in particular, Federer was very, very known for this. A bit like Bjorn Borg, who was this Nordic player who was nicknamed Iceman because he would show no emotion whatsoever. Roger Federer was always very, very little emotion, always very focused, a bit like Yannik Sinner today, where the peak of his derangement is tipping over his water bottle when he’s sitting down. It’s so funny because we have memes about that.
Turner Novak:
Oh, I didn’t see that. Have you seen that one kid who can use forehand with both hands? He switches his racket. Is that real? Is that a sustainable thing?
Nathan Benaich:
It’s not sustainable at a high level. I think the game moves way, way too fast to have the time to do that. There was... My time as a teenager playing tennis... On the women’s circuit, there’s a lady called Monica Seles, who was playing with two hands on both sides.
Turner Novak:
I think I remember that name.
Nathan Benaich:
I think it was her. There was definitely one player who was doing that, but it’s Frankensteinian, makes no sense.
Turner Novak:
Because to the point, one of my friends who played in high school, he said it’s like playing speed chess where you’re doing this tactical, strategical things, but you have no time to make any decisions. And you probably... Even the function of switching the tennis racket in your hands, you lose half a second doing that and you need to be able to move even quicker than even...
Nathan Benaich:
It’s one of those things where you need to develop this insane muscle memory that your body just reacts naturally in certain scenarios because you’ve tested it so many different times and then you have this natural intuition of what you should do a bit like taste, I think. But it is... I’m sure a lot of professional sports are like this, but the moment you get distracted and think about something else, I’m not sure professional tennis players have this distraction at all, but as a casual pseudo pro tennis player in the past, if I think about this deal or something, I’m so toast.
Turner Novak:
You’ve got to be constantly focused on positioning even. I think about a lot with hockey, I play hockey and it’s just, a big piece is where are you standing and how open are you and are you... If something were to happen, how are you positioned to react to it? That’s probably even more common in team sports because tennis, you’re always the one that’s doing the action, but if you’re playing soccer, you could theoretically play a game of soccer without touching the ball, but you impact the game based on how you move around the field.
Nathan Benaich:
At least in tennis, we’re always looking in the same direction generally.
Turner Novak:
That’s true.
Nathan Benaich:
The problem is in front of you most of the times. But I think nowadays the speed of the game is insane. The reaction times you need to have, the power you get out of the rackets and the athleticism that athletes have, it’s very scary. The stretches that they’re doing and the fact that they’re sliding on hard courts, these kids are going to have all sorts of bodily damage by the time they reach the old age of 30. It’s going to look scary.
Turner Novak:
Maybe we’ll have AI generated training routines or physical therapy to keep them healthy or surgery to fix the damage.
Nathan Benaich:
That’s going to be regenerative medicine and stem cell biology, the industry I came from many years ago. It’ll hopefully help.
Turner Novak:
Well, this is a lot of fun. I know we were talking for a while, but hopefully people learned a lot. Where can they find you? I think Twitter, you’re pretty active on. You write the Air Street Press, which is I think pretty frequent you send things out. What can people look up?
Nathan Benaich:
Twitter, just Nathan Benaich, I’m pretty active on. And then Air Street Press, which is just press.airstreet.com, those are my two favorite outlets at the moment.
Turner Novak:
And then do you send the State of AI report from the Air Street Press or is it a separate URL?
Nathan Benaich:
Yep, yep, it’s just that URL, but you can see all the prior editions on stateof.ai or stateofai.com. So all the eight years that came before that is at Google Slides. And then I do a monthly newsletter, also State of AI newsletter, which is the renaming convention I was hacking away with my AI. And that’s available on our Street Press as well.
Turner Novak:
Okay, cool. We’ll throw all those links too in the description for people to check out. Cool but thanks for doing this. This is a lot of fun.
Nathan Benaich:
Thanks, Turner.
Stream the full episode on YouTube, Spotify, or Apple.
Find transcripts of all other episodes here.