Making Humans Irreplaceable: James Lo & Anish Acharya on building & scaling Ethos

0:04Well, a very warm welcome to this very special live episode of 40 Minute Mentor. I'm thrilled to have Anish and James here. They are incredible entrepreneurs and investors and they have had stellar careers. I can't wait to dig in. This is a record for the podcast because James has been on. This is the third time James has been on. So I am incredibly honoured and grateful. I've said everything I wanted to say. Yeah, yeah. Okay. So I'm sure there'll be some spicy takes. And Anish, this is your debut. So very warm welcome. Thank you for having me. I guess we should probably, for those who haven't met you before, we should get in some intros.

0:40So I would love to hear, maybe we'll start with you Anish, given it's your debut. Do you mind sharing a bit more about your background and current role and yeah, how you've got to where you are today?

0:50Anish Acharya:I grew up in a small town in Canada, in Ontario. I started my first company after spending four years at Amazon in 2008. I sold it to Google. I started another company. I sold it to Credit Karma. And then I found my way to Andreessen. And yeah, it's been great. It's a great opportunity to be at the edge of technology and work with the best people in the world. I love that. James, are you team? Yeah. Well, I grew up in Hong Kong until I was 16. Rejected all my university offers, went on the streets, became a pro-democracy student leader in Hong Kong at the time. There were massive protests where like a 13-year-old kid called Joshua Wong was leading the protests on the streets.

1:22And I was trying to join as much as we could. Went from there to finally come over, studied politics, economics, thinking I was going to go back to Hong Kong and be a politician. That turned out to not happen at all, and I sold out completely and became a McKinsey consultant instead. Went offered to McKinsey, and then went from there to SoftBank, where I worked very closely in the turnarounds team. So we were a tiny squat team at the time, there were four of us. And then from there, I started my first company around six years ago. I was one of the early companies building on GPT-3 in late 2020.

1:47We tried to build an AI education platform on top of that. That was the first time I came on, I think, way back. And then eventually in 2024, I started Ethos with my co-founder, Daniel, who was at Dmine for six and a half years before. We decided to create this company together. So yeah, that's me. Thank you very much. So Anish, you've seen both sides of the table, founder, investor. Can you share with our audience why you decided to join A16Z when you did and how your experience as a founder is helping you as an investor?

2:15Anish Acharya:For sure. Yeah, I didn't think I wanted to be an investor because most of my experiences with investors were very poor. I've kind of had every archetype of bad board member. I think I told James this when I met, which is like too rich to care, didn't read the board materials, too many product ideas, and then high anxiety new investor. So those are all the things to avoid. Definitely nodding going on in the room. I think this is relatable. Exactly. Yeah. But after I'd sold my second company, my first company worked, which is awesome. And I think one of the reasons it worked is I had a very authentic connection to the problem I was solving.

2:47Anish Acharya:It was important to me, I think just as you do with ethos. And that's such a great sort of leading indicator of success. With my second company, I just wanted to start a company and be the boss again, and it didn't work. And I think part of it was not having the right intentions. So after four years of Credit Karma, this investor opportunity came along, and it's a point of maturation for me to realize that I just wanted to work on the most important thing that I can make a contribution to, not necessarily be the boss. So I committed, and it's been great. Thank you. And James, tell us a bit about how you two met.

3:16I'd love to hear the origin story of what is clearly a blossoming bromance. So tell us more. First of all, I'd known the Andreessen team for a while because one of the folks on the team, James DaCosta, is actually a very old friend of mine. So met in university. We were at McKinsey for a while together. He became a founder as well. And then over time, went into the Andreessen team. So I'd known the team for a while. And James was always like, you've got to meet Anish. You really, really got to meet Anish. And I was like, ah, okay. And then on a trip to SF, finally got to meet Anish. I think we were talking about this last night.

3:46I think it was like within the first, you said two minutes. I think it was 10. But we had this really electric connection. One of the challenges of this business is that Ethos is trying to build this platform to match people to opportunities with AI. We really believe that one day every person will have a personal AI that deeply understands your expertise, related potential, your aspirations is going to match you to all these things. But the problem is that we see the whole world as like a whole bundle of opportunities, right? So like to us, there's no difference matching you to expert calls or full-time job to an AI model training project.

4:17But every time we meet an investor, they're like, okay, so are you my AI expert research bet or are you my AI recruiting bet? And then some dude inevitably is gonna be like, you're the AI consulting company. You're gonna make it as an AI consulting company, right? And I met Anish and Anish just instantly understood it, right? Anish's instant feeling was, okay, hey, there are so many people with like unique ingredients in their minds of like this latent potential that AI can unlock for them. And that there's real potential here, like over our lifetimes to make a difference. And there was a very, very authentic belief in that that really made me feel, oh my God, okay, we finally found investors to actually get it.

4:53So yeah, so the paid ad is over and my paid incentivized endorsement of A16C is here. But yeah, that was how we first went.

4:59Anish Acharya:Well, I think, you know, what's so compelling about James, and you're already picking up on this, is the sense of inevitability. Like you meet this person and within two minutes, like there's a bunch of yapping going on, but you just get the feeling that this person's going to do what they're going to do and I'd love to be a part of it, but even if I'm not, it's going to happen. So much of the investor-founder relationship starts with the emotional and then builds into the intellectual. And on the intellectual point, if you look at this whole market of data labeling and human data capture, it's sort of nihilistic.

5:27Anish Acharya:It's like, okay, we're going to get all the data from all the people's heads and then make them unnecessary. necessary it's like who wants to live in that world so i think one of the powers well i know we're going to talk about london later of the approach you've taken and the person you are is that you thought about it from first principles and you've taken a much more pro-social view and it's showing up in the business result yeah it's so true i first met james in a founder's retreat there's all these people and just there's something about it i don't want to big you up too much james but there's something about your energy and you just have this incredible infectious energy james And I know everyone in the stream will see it.

5:58And clearly, Anish got it from minute one. And tell us a bit more, Anish, about the actual, like the business itself. Clearly, you were struck and you understood it in the way that James wanted you to. But tell us a bit more about kind of what excited you and your thoughts when you first had that meeting.

6:13Anish Acharya:Yeah, I mean, I guess philosophically, I love the view that human knowledge is highly adaptive. Otherwise, if you think of the most high-powered attorney that makes$1 ,000 an hour and a new graduate attorney, why do they charge so much more than the new graduate? it, they both read the same textbooks. It's because they've all of this sort of tacit knowledge of how things are done, how to build relationships, hold versus to cave. So there's just so much, and a lot of the knowledge is unspoken. I think that's the majority of the knowledge actually in the world. It's also ever-changing because the sort of application of the field is ever-changing.

6:44Anish Acharya:So the idea that actually it's not about human data capture for labs, it's about economic opportunity, and then also unbundling jobs from work. You can do work without it having to be, if you think of the formal construct of a job, traditionally, it was like, you work for 30 years, you get a gold watch, you get a pension. And now we've got people that'll work for four years, the people that'll work multiple jobs, like what is the next sort of level of granularity of work? And it's potentially doing things like the experts are doing in the network today. It's much, much bigger than just human data for labs.

7:16Anish Acharya:Absolutely. Absolutely agree. Like, I think there's a part here where we were talking about, yesterday at a board meeting about how do we get people to stop selling their time and move into position where you own your expertise as an asset and for that asset to be scalable over your lifetime. And we never really had the technology to do this, right? How do we package up what's in your mind, all the tasks and knowledge, and help like many, many more people who need that knowledge, who need that expertise, access it. But where it turns into a form of licensing, a form of royalties, a form of passive income for you, in addition to the thing you spend most of your time on day to day, right?

7:48Like that's the part that we think this will truly unlock over like the next 10, 20 years. And it starts very, very basic in day one, right? In day one, it's basically you get an AI profile. We can match you to jobs with it. We can match you to projects with it. We can match you to expert calls with it. But over time, it's really about creating a new form of income for people. Because if we think about world of work is going, it's essentially that all the people who are selling their time for income have stayed stagnant for 40 years. And And all the people who own assets and capital have made multiples and multiples of wealth.

8:20So we got to figure out how to diversify this ownership of capital. And for the first time, knowledge can become capital. And that's the big transition we're trying to make. Super exciting. I'm conscious we've already got into some of the vision you have for the business, but can we maybe take it back to the origin story? What made you go, this is the problem I want to solve? And how has it evolved since that moment? So I think the fundamental inspiration for it was, how do we build something that can advance social mobility at scale? That was actually the core problem. So actually, when you look at the entity name of Ethos, the U.S.

8:53entity that holds the parent entity is called Somolabs. And what Somo stands for is actually social mobility. And I still own the domain for socialmobility.ai, which one day is going to come back up. So it was really the foundational idea of the company, right? How do we advance social mobility at scale? And the reason why it mattered so much to me is that I grew up in a really weird family where my dad was super successful before I was born. So he was like insurance company CEO and all that. The moment I was born, I was born into this like three-story mansion in Hong Kong, which is crazy in Hong Kong.

9:22And one of my first memories is that I was like driving a little mini Ferrari, like in the house. You know, those like little cars that you just like, it was like completely insane. And then the Asian financial crisis hit. My dad lost his job, made a bunch of very terrible investments. And then suddenly we went from super rich to all the way straight down. So most of my childhood was basically moving from bigger houses to smaller houses. And so we ended up in a flat that was lent to us by one of our uncles. So in a weird way, I was surrounded by nepotism. So I went to one of the best high schools in Hong Kong because the headmaster went to the same class, like high school class as my dad.

9:56So I got in with that. And I was like paying my school bills by like working for my headmaster in the shadows. And then he was basically paying my school fees in the back. It was like completely crazy. So surrounded by nepotism, my dad's friends were all kind of rich dudes with rich kids. Meanwhile, we were in this constant decline. So by the time I went to protesting in Hong Kong and so on, most of my rich friends were not on the streets. They were criticizing all of this stuff. Meanwhile, everyone who's fighting on the streets were losing opportunity and so on. So the fundamental DNA for the last 30 years of my life has been, how do we fix this?

10:29Opportunities are incredibly unevenly distributed. In my first business, what I learned was that it turned out education was not the answer to this. Because most people go, education is the answer. And it turned out education was so reliant on motivation. And the people who need that are not the people who could pay. And so you just fundamentally do not have a business model that can sustain this over time. So it then became a question of what is AI unlocking now that could give us the ability to advance social mobility at scale. And it became increasingly clear to me that actually AI is the first technology that can truly have almost infinite tokens to understand an individual.

11:02We've gone from a position where nobody will ever understand you to the depth that AI can. And at the same time, AI can actually understand the landscape of economic opportunities to a level that no other technology has enabled us to do. And that, in fact, we could match people to opportunities with AI much more efficiently than anything that has ever been created. So that's when it sort of started clicking to me that you could actually start advancing social mobility at scale in this way. And then it became quite clear that we need to build something like this. It's so impressive. I'd love to get into some of that detail.

11:29But also, you've done all this with a team of six people. This is incredibly lean and very exciting. Not great for a headhunter, can I say, but I respect it for sure. Can you walk us through how you've approached growing at such velocity in such a lean way? Because I think the founders in the audience, the founders listening to this, We'll really appreciate hearing kind of why you chose that approach. Yeah, for sure. So the latest numbers, by the way, is we're now 84X'd since the series A. We're well into nine figures, annualized eight figures per month. The full-time employees are eight people now instead of six.

12:04The fundamentals of this was we were trying to think from first principles, how do we build a company in the age of AI? And it became quite clear that we have to build almost like a centralized intelligence of a company. So when we first started building a company, we built this knowledge graph of people, companies and products, right? So we bought around 500 million profiles, 60 million companies. We sort of figured out, OK, how do we build a graph of this? Then we basically loaded a bunch of agents on top to like operate on top of this. And every single thing we then did, right, our Slack, our linear, which we eventually dropped because we don't need it anymore, like Notion, our CRM, our internal management of the company, so on and so forth.

12:37All of those became vibe coded tools, actually. So we just built on top of our own system. We basically stopped like outsourcing any of this. And we plugged all of that context into like the group of agents that are working for the company. Then as the sort of models got better and better, we started plugging, you know, essentially the agents into Slack for the whole company to use, right? So sharing this yesterday, but literally our entire non-technical team are in Slack every day talking to Cursor. Literally, they're just like, I need to build this into the product. I need to build this into the support system.

13:06I need the admin panel to change in this way. And at the same time, for the experts that we work with, the support agent is fully automated, but it's not just like automated and fixed data, right? So as the project evolves, as people talk on the system, all of that context is being ingested and then leveraged by the support agent to support the experts with live information of what's going on. So essentially, we built this centralized intelligence that allows us to use agents to do most of the work. And for the company, then we were very focused on having an extremely, extremely high quality bar.

13:38And our thesis there was that the sort of human coordination was actually the biggest sort of cost to the company. So like if you have lots of people in a hierarchy and they're debating and they're going, I need gated approvals to do this and that, you slow down the company immensely. But in our company, we flattened it entirely. And in engineering, Daniel is basically at the same level as like the other engineers. Everybody can just make decisions autonomously and go do something. Everybody's been given like a very high level of ownership, comparable companies elsewhere as well. And for me with the ops team, it's like at this point, Ashini on our team is basically running the company.

14:12And like Sam is running the company as well. each of them own like functions that they're just driving and they don't really need my approvals for anything right so we are almost like a collective running the company but what that relies on is extremely high quality bar for talent that is not just about their experience and their skills but it's also about their personality about their ethics about their conviction of the mission all types of things one reason why like every engineer took roughly five to six months to hire so far and we're just going on and on just looking for people and we've met so many people from top companies.

14:40I'm still continuing this work. Yeah, I love that. Keeping that super high, but you've created such a strong business and culture already that you don't want to drop that at all. That makes total sense. Anish, how does that for you, seeing this in London must be exciting. I'd love to know how that compares to what you're seeing in the Bay Area and get your thoughts as an investor. What gives you the confidence that this sort of velocity at this sort of lean structure is sustainable?

15:05Anish Acharya:So that should be our collective aspiration, right? Like we don't want to be the best X for London or for Europe or for the rest of the world. We just want to be the best X. I think that's a really important part of all of our stories when we go and talk to investors. I think on one hand, all the best companies that have this explosive growth have a similar shape, right? Which is a lot of tenacity and high agency sort of generalists in the early days, because it's hard to hire the specialists. And typically the work is so overlapping that there isn't really room for that specialization. And when there is, of course, will come and work with you and it'll be amazing.

15:37Anish Acharya:But up until that point, you just need high agency generalists that can get shit done. I think the magic of the era we're living in now, we were talking about loops and automations, what you're describing with customer support, like there's a huge amount of loops and automations you can build and code. There's a bunch of stuff you can do around business processes broadly. And then I think the most satisfying thing for us as founders is that there's no excuse to not build all the software you want to build now, right? Because of coding agents, product velocity is just an expectation and a baseline.

16:06Anish Acharya:And the beauty is you don't have to be like, oh, I'll hire five engineers to build this feature and we'll do it in three months. Like you're able, there is really no ceiling on software ambition. So the company shape, I think is a lot more compelling to investors, but it's also just a lot more fun to work at. Whereas the old days of being a founder was all your time was hiring instead of all your time was building. So true. Thank you very much for sharing. James, you're operating in a really crowded market. But how do you think about differentiation? And how do you ensure you can continue to execute fast and kind of beat out that competition?

16:37Yeah, for sure. So I think one of the key observations from the beginning of the company is that the whole spectrum of economic opportunities can actually be facilitated by one profile. So we shouldn't think of it as building an AI recruiting company or an AI market research company, an AI model training company. But instead, think of it as how do we build the best possible profile for an individual expert that will deeply understand them in every way and match them to opportunities continually. And then every act of matching will improve the data that we hold on these experts and basically help them eventually structure this into an asset that we could scale.

17:10So that was super important from day one. And I think it's a differentiation inherently, because once you do that, suddenly the people who are coming in because they could do a thousand dollar per hour expert call, who might be a senior executive who's more used to that type of thing, is getting cross-sold like, hey, by the way, you You could do like$200 an hour, but for 40 hours a week or 30 hours a week or 10 hours a week for like AI model training. And by the way, there's this advisory project that's super interesting to you that you might want to do. By the way, like there's lots of market research companies that want to understand like your income profile, so on and so forth.

17:40So these cross benefits that continually develop in a way that each of those vertical players are unable to access. And the sort of fundamental observation making is that no matter whether it's recruiting or model training or expert calls, People just want a warm network of the most incredible people in the world and an understanding of what they know that is deeper than anywhere else they could search for and get them quickly. So you can actually build a way warmer network with way more depth by being cross vertical. So that, by the way, to this moment, even after all the growth and all that, remains a contrarian thesis.

18:13If I talk to 10 VCs, nine of them will tell me I am distracted. I'm like working on too many things at once. This thing is not going to work. There's so many players that are like dominating each of those verticals. You get killed by all of them. And I get it. But this is the foundational piece of the company. We're going to keep going at it.

18:30Anish Acharya:What I love about what James is saying here is I think there's something really important to observe here, which is the importance of the authentic connection to the problem. Because if you were sort of casually looking at this market, you'd be like, holy shit, super competitive. It's very crowded. There's all these statements. So like, what are we actually going to do here? But instead, if you actually approach it from first principles, because you have that connection, things like differentiation and moats reveal themselves. There's, I think, a thing that's very easy for us to do as founders is want to have this perfect plan.

18:57Anish Acharya:And it's sort of, forgive me, but it's like a strategy masturbation where you're like, okay, I need to have the perfect plan because I have the perfect plan. It's all going to work and investors will like me. And it's like, no, you know, moats and differentiation is something that's discovered through shipping, not something that you actually have an epiphany about in the middle of the night. Yeah, very well said. Anish, I guess you'll have seen companies have this sort of exponential explosive growth many times over. We always like to talk about the harder parts and the challenging parts of scaling journeys.