Why Physical AI Is the Next Platform Shift

Starting a company is a rollercoaster. So there's like low lows and they're high highs and My view is that you should just ride the rollercoaster like people try to mitigate it, but just ride it and try to have fun Please welcome Eric Lando with the co-founder of Anchored who is building the data layer for physical AI, right? Yep. Welcome Eric. Thank you. Thanks for having me So as I was going like preparing the the chat and looking at your background I was actually pretty surprised I kind of like just very surprising that you did big data particle physics and Then you became a quant for 10 years Before creating Anchored. Okay, like lead us with that. Like why like is there a common thread here? Yeah, the common thread is almost the kind of trajectory of AI development in the time that I've had my career So it's it's quite interesting because do you know the the bitter lesson? I mean the bitter lesson is this idea that AI You know, it used to be that to build an AI system. You would think very carefully about it you do all the feature engineering you try to induce as much domain expertise into the the system before kind of wrapping it up in machine learning and they found that Actually, it's better to not do any of that and just to scale the system Just add more data more compute into it. And when I started in physics, I was doing particle physics the Kind of systems and the strategy was based off of the physics itself. So you Filter the data based off like the way the particles, you know You know, we're moving the momentum and and then in high-frequency trading it was thinking about the markets what factors actually move the market one way or the other and then Really crafting that and then putting it together into machine learning model afterwards And now it's just throwing data into a big pile of machine learning and then seeing it hoping for the best. Yeah Before speaking about anchored I realize we were chatting quickly yesterday about that, right? I mean, it was not the best at your bed to become wealthy I mean, you've been 10 years as a quant I guess you were making way more money as a quant that you did as a founder. Is that right? yes, and in fact when I quit I quit during kovat and Everything just I quit to go start a company in that month The market went absolutely crazy. It was the month where oil was trading negative extreme high volatility People were making a lot of money on Wall Street bets My desk made more money in that month than I think they made in the previous three years combined. So I just quit And I was sitting on my couch fighting Python dependencies in a machine learning library and I was like, what have I done? But obviously I didn't do it for the money. You left a lot of money of the table Well, you didn't know you left as much as that And so what did you do it like what led you to actually quit a very well-paying job to become a founder? so when my co-founder I really believed in AI and to us it felt like This was the Technological paradigm shift of our generation. This was the early days of the Internet. This was the early days of computing and Whenever I was a quant for quite some time and for the last several years When I would go to sleep at night, I would like hit my head on the pillow and I would think What am I doing with my life, you know? and When I started a company even when all the craziness happened in the markets I would go to sleep at night and just think about the problems of the company. I wouldn't have the Existential fear anymore. It was all like more of a very concrete concrete one, but we are doing actually both providing value to someone Well, did it at first? No, it doesn't at first but you believe you're part of something which is bigger than you and it's very rewarding It's extremely rewarding to be in a company and seeing it grow and seeing the impact that can it can have And so we're in 2021 is that that that's when you created on Cartwright How long did it take for you to could be convinced that There was something here like the early years were not that easy, right? No, we we were in the desert for a couple years So why like what made you continue and not give up a pivot? We know we still believed in AI and you know If you look in the long long horizon the power of the technology the generalizability of it But it was hard to convince people That it was as important as it was and then chat GPT happened and the kind of tenor of the conversation changed which Definitely helped us quite a bit the chat GPT release Actually unlocked the market like and is that because people needed more data or because people starting to realize That AI was important. Yeah, it was more they were open to the conversation So it's funny because product market fits for some folks it's so product market is both product and market and these things have to come together and so Some people they get it They achieve product market fit by setting a market constant and then compounding their product to hit that market Sometimes it's the market comes to the product you know the world changes and suddenly the thing that people didn't care about now care about but most cases it's them kind of Both moving at the same direction and for us Our product was just continuously getting better and better and better and then after chat GPT I wouldn't say there was like a big binary threshold change, but the market started coming to us Faster than it was before and so a lot of people a lot of companies you talk to them They they did a product launch. They changed their product and then they suddenly felt immense product market fit That never happened to us. It just felt like it was a little bit better every day It just kind of was slowly slowly compounding and looking back on it. My co-founder he thinks like time when we have product market fit was When we saw, you know a sale closed in the gong Channel and we didn't know what the company did like what is this company? and that was very Satisfying because we were doing all the sales ourselves We were like intensely involved in every single part of every company that we were trying to sell to and when you see something They're giving you money. You don't know what they do. That's like, oh, we made it You are not involved in the tail. That's what you mean. Okay, when was that? Two months ago. No, it was it was some some years ago, but it's only looking back There was never a single moment. We're like, oh we have got product market fit. Yeah, I mean It's kind of like actually one of the topics here about I mean you were crowned the physis And you had to be the sales dog at some point. How were you able to do that? How did you hire the first test people and when? Yeah, it was I wish we had a more elegant Answer to this, but it was trial and error. So how many do you have to fire? Yeah, so we had a we brought in a sales team. It didn't work. We bought another sales team It didn't work by the third sales team. It started to work and it's funny because You know the the advice that they give you in YC It's one of the things that we found is like it's almost always right. So in the advice of voices always yeah No, it's almost always right, but you still make the mistake and I remember very distinctly There is a particular hire that we were about to meet or actually we made and we talked to you This was maybe a couple years after we had finished with it YC and you were you said? Oh for this role You should hire you should look out for this and this and this and don't do this and this and this and we're like oh, no, because we had just hired exactly what you said not to hire and We realized oh we made a big mistake But we still brought them on and we had to learn that it wasn't a fit We should just cut it there but That one was I think probably in nine months and then oh nine months, okay, not nine days or nine We've gotten a lot more efficient at learnings here. But yeah, this was something that we learned the hard way, right? So most of you are not there yet Well, there is no world in which you keep someone nine months that you know is bad Yeah, and it's never the thing is it's never sometimes it's like bad Man, this is bad. But usually it's one of those things where you kind of know in your gut and or you know, you have the feeling but it's just difficult and there's never a strong for there's never a strong reason you just know It's probably not a fit but there's nothing that pushes you to it and what we've gotten a lot better at over the years is just following your gut and Making the right decision faster. Yeah, I cannot emphasize that more Let's go back to the product now Because you didn't start with physical AI. Physical AI is kind of like the it felt it feels like you've built it a few times Still always around data for AI, but the target applications have changed over time, right? What did you start with? Was that healthcare? Yeah, we started in vision and our thesis originally started with vision was it was the hardest modality to start with and so we Always wanted to be multimodal and because our thesis was that AI systems would be multimodal naturally because humans are multimodal The best way that you make the best decisions is to take as many sensory inputs as you can you can accept So vision was a very dense data type we developed a lot of expertise into it and then we moved into multimodality and what happened with physical AI is that That's one of those cases of you know, your product goes up and the market comes down The market started really coming in our direction and the biggest application of multimodal AI now is in physical AI So now we work in a lot of robotics autonomous vehicles logistics manufacturing But we started in Other use cases which were less, you know less popular now it feels like In some sense physical AI is still nascent as a market I don't know like when you say that I think robots and I don't see robots every day yet But what you're saying is that the market is already there. It's already here for you and so there are a lot of robots that are working in the real world, but The opportunity in physical AI is enormous 80% of economic activity is Manipulating or moving things in the real world. So we live in the real world and If you talk to a lot of the robotics folks, they think in five to ten years There will be more robots than people so in this room All this other seats will be robots that are listening to the talk and then transmitting it back into some centralized server I may not need the robot for that. Yeah Yeah, you all this will be a room or a world packed with Autonomy, let's say all right Your company is doing great and you are betting the future of the company ahead of the curve on a new space. Is that fair? How did you come up with that decision? Did you see new customers coming from physical AI? How was that the result of a deep conversation with your co-founder or something? It's yeah so with AI you always have to you can never be comfortable you always have to go try to think about the next thing and The best way to do that is just to talk to people like talk to as many people you can talk to smart people And have conversations with your co-founders. So we're constantly constantly Even talking to prospects to customers people that are ahead of us in the space these at these YC events Talking to other co-founders in AI and really just trying to absorb what's going on What are they what problems are they thinking about what are they doing and then synthesizing those things into the the next bet that we Can we can make so it's not us sitting in a room quietly trying to imagine the future It's really being out there and getting as much data as we can Okay, so you shift the narrative you shift how you speak about the company itself. What about existing customers? Well, because we did you still have a lot of healthcare models? Yeah, we still have a bunch of healthcare companies We have companies in basically every application and the most fun one People always ask what the most interesting one. It's facial recognition for cows So it turns out that Humans are are really bad at recognizing cow faces, but algorithms could do it quite. Well Probably better too. Yeah, so We've seen every we've seen everything in AI and that's the one of the who is annotating the the data Yeah humans that are experts well you can there's other ways of tracking cows to the barn and but you want to know their health and things and yeah, humans are Are only a part of the annotation process as well Like yeah, once you know who is the cow like you tag it then, you know, you can recognize its face many times or something Yeah That's cool. That's cool And what would say are the biggest use cases today then Physically, I like you said autonomous driving as one. Yeah autonomous driving. That's one. That's Further along in maturity curve, so it's closer to actual production It's actually in production in some places. There is. Yeah, I'm close to kind of mass-scale production And one of the bigger the biggest growth for us now is in robotics. So humanoids consumer robotics Manufacturing there's so many different applications for robotics that we're seeing And how do you collect data? How do we collect data? so we do have a Facility in the Bay Area that we have robot systems and operators that are collecting data for various tasks A lot of the data is collected in production as well. So our system Ingests production data from the customer and then uses that is that video? Dream is that it's a video stream sensor data audio Language, it's it's multi-modal data. So how does that work that facility in the area? Like people come with their robots and you have like some setup. That's right. Yeah. Yeah We have it's almost like film sets Where you kind of build a thing you have a robot do a thing in a specific environment collect data in that that environment but Data collection is usually for the companies that are earlier in the stage Collecting enough data for pre-training a lot of companies already have production loops So what's the rest of the product the rest of the product is doing the management of that data at the curation? So selecting the right data, it's we say it's like Finding a million needles in a billion haystacks there's a lot of data that you need to use but even more data that you have and Then the annotation and enrichment of the data and finally the eval You have a can share with us like me some sense of the scale of the operation Whatever well, we have hundreds of customers we deal with more data than was used to train Well from what we know GPT for we don't know the latest amounts of data for GPT-5 Yeah, we're dealing with multiple petabytes of all different types of multi-modal data So you just mentioned that you built a facility in the Bay Area, but you're based in London You're open and you still Expand you to beat that facilities there. Why? Yeah, so my co-founder is in San Francisco. I'm I'm still based in London. We started the company in London and Our general philosophy is you should just be where the important people for your company arm So for us, it's customers and then talent and then investors in that order But different people have different weightings of those categories So if you're building something in deep tech, you don't care about customers you might over index on just the talent bit and then you can just be close to where the good talent is for us most Of our customers are in the u.s. And of those the vast majority are in the Bay Area So my co-founder moved there. We built a facility there. We built a lot of operations there just to be close to have those Huh? Talent is is mixed. But I think London has some great talent or for AI and engineering great universities people from from other companies So there's a comparative advantage for for talent right now in in London Europe Although there's now this London maxing term, which is kind of arbing out the the talent Yeah Slightly less intense in Europe than it is in NSF. Yeah You are competing with that kind of crazy good companies too, I mean when we speak about the data collection I'm thinking about scale And in Merkur and many other all our u.s. Companies. How do you stand out? How do you? Win against them. Yeah, and so one when we started the company The the main Pushback that we get was like this space is way too competitive There's too many companies and then over the years a lot of our competitors actually They pivoted or they consolidated or they ran out of money and for a long time There actually wasn't much competition in in the field now It's getting competitive again and our view is actually we want it to be competitive we prefer to have really good competitors because they they push you they make you better as a company and Standing out we all have different kind of flavors within but the ones that you mentioned our focus is really on physical AI and scalability of the data like being able to Operate on the multiple petabyte level of data, which is hard for people to do and which we spent a long time You know building the foundations for if anyone here is building in physical AI Should they try like and they're small and that's just starting. What should they do? Use such a product like yours, I guess you have go to market over time and your product may be not affordable for them Yeah, so usually where people come to us is when they've built internal tools They tried POC and they're at the inflection point where they need scale So they're there their models are going from POC to production or they're just getting a big influx of data And I usually I don't want to like oversell our product when it's too early It's like when you're ready, you should use a product like ours But you can start with you know, open source tools like what like any any product anything to recommend? There's a there's a whole host of them. I don't want to undercut my own My own book here there. That's okay before including here any As well on the topic of advice any advice you give to founder in that space physically I the advice that I think is important for basically all founders and One of this came from one of the YCG retreats that the CEO of reddit was talking about which is Starting companies a rollercoaster. So there's like low lows and they're high highs and My view is that you should just ride the rollercoaster like people try to mitigate it But just ride it and try to have fun because whenever you're at the low, you know that a high is coming whenever you're high, you know Low is coming and just accept that the that you're on a rollercoaster Speak about like hacking your brain If you can hack your own brain to enjoy the fires And then when there is a new fire a new challenge you just run to the fire run at the fire like because you it's exciting Then it becomes so much more enjoyable as a journey because yeah, we're going to go through crazy highs and crazy lows Yeah, roller coasters are fun Did you think that when you started? No When did you start thinking that? Only relatively recently actually when you go through enough of these these sinusoidal fluctuations you feel like okay I can just make this fun. That's interesting. It's rare that people think That it becomes easier Do you think it become became easier? You you become more you realize that you have more control than you think over your own mental state and then you can kind of accept what the reality is and how you how you react to it and That's something that has helped a lot in Seeing the fluctuations in increase over time. Awesome. Thank you. Eric. It was great to have you with us today. Thanks so much