Stephen Sikes on Why Public's Agentic Brokerage Changes Everything

Transcript

All right guys, welcome back to AI Supercycle. This is episode number 35 powered by our friends at Near with their Near AI product suite, agent marketplace, and more. I've got Stephen Sykes, the COO of public.com. He leads brokerage operations, crypto, institutional expansion, alternatives, customer support, legal, compliance, risk, a former US Army veteran, cavalry officer, engagement manager at McKinsey, and more. Guys got it all. Welcome to the show. >> Thanks for having me. Yeah, for sure. I'm looking forward to talking. >> Steve, good good view, good day in the city?

>> Totally. Totally. Great great book over my shoulder. >> Yeah, exactly. Exactly, man. Intelligent Investor. Well, I mean, where else to start other than this kind of convergence between financial markets and agentic finance, right? I mean, more than ever are we seeing this transition of portfolio management, workflows, operations, decision-making happen from a LLM first kind of like mindset from from humans, right? Even so much so putting their own capital in and wanting higher levels of intelligence to be managing their capital, managing their emotions, etc. Maybe just let's just start high-level with your kind of current view on the state of this agentic investing world from from what you you're doing up at at Public and

kind of we'll start there. >> Yeah, so I think like helpful to put where we as Public are in in context. Overall, I think it'll kind of help explain like how we're thinking about the world. So, I think we see ourselves and we branded ourselves as sort of the platform for serious investors and serious traders. That's led us on a journey to build out sort of the best platform for asset class exposure, right? So, serious investors and traders need access to equities, fixed income, corporate bonds, treasuries, crypto, options, and you know, you can keep going and we're going to keep going into, you know, futures, prediction markets, and so on. But, like we first viewed our job as like make sure we have the platform that can do everything that serious investors and traders want. The

second thing we found serving that customer base is like they need the best tools. And that's like data, content, uh and increasingly it's AI tools. And so we've been on a journey to serve this serious investor and trader with AI tools literally going back to summer 2022 even before, you know, when it was just uh GPT-3 via API to help support our customer support team all the way through ChatGPT launching later that year in 2022 to our first product uh which was uh you know, still was in the app. It's called Alpha. It's our AI research assistant, which made a ton of sense, right? I think in 2023 we were looking at a world where there's never been better content, information, market data for retail investors. There's never been more of it, but wading through what was most relevant from a research

perspective was really hard for people that weren't professionals and weren't doing it full-time. And so Alpha was great. It helped sort of filter out filter the universe for what's interesting, what they should know. We built a product called Why Is It Moving? So we found like most people like what do they show up to Public to do? They look at their portfolio. What's up? What's down? When things are up or down by more than sort of the market, they want to know why. AI is phenomenal at figuring Even back to 2023 and 2024, AI was phenomenal at figuring it out. But that was like the first sort of research summarization phase. Uh then what we saw people using it for and and started to build was like how do I build how do I use this sort of filter on the world of information to build better

portfolios? So how do I take a theme or a cut you know, a custom point of view, a theme, a strategy and build that into my own sort of direct index, my own custom index. And so we built generated assets uh in 2025, released that to help people build a broad portfolio. Awesome way if you have a point of view and you just want, you know, 20, 30, 40, 50 individual names in a long-term portfolio professionally managed, you know, by our system. Awesome awesome product. But like, you know, that was 2025. Yeah. Now we're in 2026. What did we see at the tail end of 2026 or 2025 was the emergence of this agent, right?

>> Yeah. We saw We saw Claude Opus 4.5 make a huge jump in terms of what it could do from a, you know, logic, software development, and just agentic process technology. And that jump led us to like, you know, just start exploring. Like, what can we do with it? Like, I think, you know, all of us in the you know, in sort of the senior team at Public had open Claude and Mac minis on our desks within, you know, a few days of that catching fire and we've been playing with it really since sort of, you know, again, late last year. And the very first thing we started working on was, all right, now we've got research, we've got portfolio management. What's the third lever, which is like basically full automation on top of your brokerage account using agents to be able to execute trades

again across the full platform that we have, money movements, trades, alerts, whatever you want, um our sort of agentic brokerage can do it. Right now, uh we've rolled it out to, I don't know, somewhere between 10 and 20,000 of our customers off the wait list. We have tens of thousands, maybe almost 100,000 more people still waiting on the wait list to get off, but uh we're rolling off, you know, 5, 10,000 a week at this point, just bringing people into the product. And, you know, it's awesome. It takes the best of sort of the agentic world with all of the sort of safety, security, and control of what you expect from Public um within sort of, you know, our UI. Now, you asked the question, which is like, what is my point of view overall, right?

I had to tell that story, I think, to to get everyone in the audience up to where we are. But, I think we're in the middle of like a really important platform shift, right? I think our industry of investing has been through many of these, right? Like, the first platform, you know, or like the most, you know, call it going back to like the the time we might remember, like, the first platform shift was from like showing up at a broker to calling your broker. Then it went from like calling your broker to like doing it online on a web view. Then it went from like web to mobile. Now it's going from mobile to AI. And like, sure, dude, we still have people calling in trades at brokers. So, it's not like the prior versions are going to disappear, but we know that this AI, you know, use cases,

AI UI, this AI UX for investing is the next platform shift, and we're really excited about it. Well, and I mean, I described our agentic brokerage, but we've made two other big bets, which I think will, you know, be really important and like the first is our sort of direct-to-consumer API. So, we have basically our entire brokerage stack, our entire investing stack available via API. You can sign up for Public and you know, deposit money, move your account over, and grab your API keys in 3, 4, 5 minutes. You can take those API keys, you can give them to Open Claw, you can do what you want. Great. So, like, you can roll your own agent, use your own API keys, and execute sort of whatever you need to against your Public account.

It's all there. The second thing is like, you know, the middle version between, you know, you roll your own with our API keys, or you use the Public UI with our own agent inside the Public app. The middle version is like, okay, you want to use Claude, you want to use OpenAI, and you want to use an or Perplexity, and you want to use an MCP behind it, right? Um, we have that as well, right? So, we've done sort of all three of these bets, cuz the truth is, we see the platform shift, but we don't really know where user behaviors are going to go. Are more people going to want direct-at-the-metal API key access, roll their own agents, third-party open-source agents? I don't know. More people going to want to use their own sort of Claude UI, Claude Co-Work, and an MCP behind it? I don't know. More

people going to want to use the Public, uh, you know, UI and experience that we've built to help, you know, work with agents. And the truth is like, it's probably going to be a bit of all three, and right now we've got all three built, and they're all humming, and they're all great businesses for us, and, you know, we're excited to drive adoption across the board. >> Yeah, I think like this entire shift that you described from phones to digital, like the next big shift is kind of like this like prompt shift. Sort of like this like kind of like And also, the big breakthrough with Open Claw was previous kind of iterations of LLM's or models were very like um you had to prompt them extremely well to get the result and they were they were far they were far less prescriptive than Open Claw was. Open Claw came to you and

was like prescribing a solution to whatever you were saying and suggesting an action and like then doing the action versus like kind of like having this back and forth to get to your end state. It was like no no here's your end state. Do you want to do you want me to you know, take you there. And so when it comes to investing that like promptless or that prescriptive unlock of Open Claw or of this, you know, next era of intelligence is going to allow such a breadth of of prompts or ideas to go in and for the feedback to come back and to change the entire scope of thinking for any given investor.

>> Totally. And I mean I think it was that that change was like you know, the ability for these models to sort of iterate inside themselves, right? Ask is this the right thing and like be able to iterate on their own prompts and make sure they're really identifying your goal and be able to agentically achieve sort of the goal or intent or strategy that you wanted. And like it's still not perfect, right? We all sort of even with like Open Claw Open Claw 4.7. No, you we're still iterating as humans on top of it. But it's way you know, we know we're sort of interacting with a real intelligence. Again, it's not the same as a human. It's not there yet, but I think this is the worst it's ever going to be.

>> Right. >> And it's kind of wild to look at it. And so I think you know, >> And the most expensive it's ever going to be as well for this level of intelligence. >> Totally. I totally agree with that. Which is also insane. Maybe. I mean again I'm like yes, it is the You don't agree? I think what's going to happen I mean again, I think we will see the state of the art intelligence. My guess is it actually probably gets more expensive than the state of the art intelligence today. >> Really? >> Again, I just think we have more people chasing more finite commute compute. Like I don't know. We're still [clears throat] in like early adopter days of this stuff, right? I still think we got a lot of especially when you get into the agentic world.

You know, our our, you know, dads, moms, grandmas aren't really doing the things probably that you and I are. And like, you know, we haven't really crossed that chasm to like mass adoption yet. And when we do, again, it takes I don't know if it's 6 months, 18 months, 24 months for really that to happen. You know, the agentic the volume of demand for intelligence is going to grow beyond even what our computer I mean, like I don't think our compute demands are you know, are constrained right now, right? We see Claude with, you know, latency issues and and trying to trying to talk us into doing simpler queries all the time. Like, you know, I don't know if that what that's related to, but in general, it's like,

you know, I I don't know. My guess is the state of the art's actually going to get more expensive. I think you're right for the current level of intelligence, that will probably get cheaper over time. >> Yeah. But there'll be like this like god tier like if you want to access the best of the best. >> Yeah, it's going to get even more expensive, my guess. >> Right. Right. Um, because again, the use cases are that valuable. And I think that's Again, I'm like, you know, speculating here, but that's that's sort of my guess. >> Yeah. So so then, let's kind of think think about this agentic workflow product suite that you've laid out, right? And think about it from like a practical use case. I've got an open claw. I use all different types of of LLMs.

Um, you know, I would say like the like the consumer demand is I think there's a way to make better investment decisions based on a intelligence that is 24/7, that does not have human emotions, human biases. >> Totally. >> And I want to tap into that. And so like, how how how how would you envision somebody going from like not using AI in their financial investing kind of career to like just starting to like dabble a bit? Like, what does that process actually look like? >> Yeah, I mean, I think they're like, I don't know, we'll take it in two parts.

I think there are two things. One, um, like the real hardcore algorithmic trading strategies. Strategies that are only accessible by software, right? Which means you have to write your own software to be able to actually execute that strategy profitably or with the latency required, etc. That's like one end of the spectrum. By the way, that is now unlocked for a universe of people that it was not. But I think where you're going is actually the other side of the spectrum, which is regular people, right? I think what we've found is we really do have the smartest, most sophisticated, most serious retail investor population we've ever had, right? Again, I said earlier like we've never had

better access to like content, analysis, data, educational materials for the markets for retail investors. We've never had it better. And as a result, we have entire generations of investors who like really know this stuff. Now, they, again, not being professionals, understand it at the level of a strategy or an intent. Now, how you execute on that strategy and that intent, like we sometimes see it get a little suboptimal or off the rails, >> Right. Of course. >> And like I think it's like, I don't know, take the take one of my um favorite ones to talk about, like a a very classic covered call strategy, >> Okay.

>> right? Um you know, you own a big position in Nvidia or Micron cuz you're way smarter than I am and you bought that stuff 2 3 years ago. As you said, I Again, we have a lot I mean, we have a lot of members who have made a lot of money. But now they're sitting on these huge positions >> Yeah. >> um and they're willing to like, you know, both like they want to take some income, they also want to hedge a little bit of risk. Like a covered call strategy can be helpful in those in those veins. And they know they want to execute a covered call strategy, but that's about where it goes. I want to generate some I don't actually know, should I execute you know, how far into the future? How many days to expiry?

>> you just know you want to take some risk off or that you're feeling a little uneasy and you just kind of want to do something about it. >> And I think we have a lot of people that really understand that those are things you can use options for, >> Right. >> But, the specific implementation of that really matters a lot. And there's optimal, sub-optimal. There are different strikes >> Yeah. >> in the option chain where there are tighter spreads, aka like lower cost to enter. There's, you know, different levels of volatility depending on where you're at in the day days to expiry. And like, you know, sitting down with Public Sagint brokerage and having that conversation, "Here's my intent, you know, what are How should I think about this? Where on the option chain should I look?" And then, like you can actually do that sort of education about

the strategy directly back and forth with a really smart, you know, trading AI. You can then filter and screen sort of the options chain for the stuff that you now think is the right decision. And then you can take the third step, which is you can implement it and you can automate it. And you can, you know, not worry about every Friday, "Do I need to be on my phone or my laptop at 10:30 cuz I think that's the optimal time to roll my strategy." No. Don't worry about it. Like, write it up. Our Basically, the way our agent works is it takes your intent, it takes your strategy, and then it commits it to a deterministic workflow. It just builds a step-by-step trading strategy or step-by-step, you know, transaction stream. If this, then that. All fully deterministic. No AI

decisions happening once you hit the authorize or submit button. And there you go. And you're off and running. And so, like you've got the security of like knowing what you just said you want to happen is what's going to happen. You don't have to worry about anything downstream of that. And you get the the convenience of it being fully automated. So, like, you know, that's sort of like the use case that we Again, a good example of a use case we've seen um is just again, people that are really smart, serious, understand these strategies, and might even have an idea how to implement it, but aren't totally sure if they're in the optimal place. >> Yeah. Yeah. And I think, you know, that goes for a lot of people who have an idea of what they would like to get out of the market and you know, yeah, as you

said, they don't want to do this complex strategy or execute this intense kind of research to figure it out. You know, I'm curious when it comes to this workflow. Today, Robinhood announced this kind of MCP uh workflow as well. Yesterday, I saw like a more crypto uh crypto-native company called Liquid. They also announced like an LLM front end basically to trade perps on Hyperliquid. Um What is like obviously this trend is happening, but like let's get to the bottom of kind of like what's actually happening. These agent workflows using MCP, like what does that even mean? And and how is that going to change my interaction with like Publics AI, Robinhood AI? Like like what does that

even look like? >> Yeah, so I mean I think the the what we're seeing and again we're seeing this a very similar from the Liquid and the Robinhood teams today. Um we're seeing the model where you can go into Claude, you can add a custom connector is what they call it, right? Which is a custom MCP somebody else has built that gives your instance of Claude in the UI that you're used to talking to access to sort of the APIs or some subset of the APIs of those providers, right? Um you know, I think I haven't really gone deep on the Robinhood one yet. I'm not actually sure if it was working. I was I saw some 404 reports uh on on X as I was coming over just trying to see what the latest state was. Um but what we've heard from their announcement

was it's equities only and you have to have a separate account just for this thing, right? So it's not like able to help you with your current positions. You're not actually, you know, really using it to ask about your current uh account. You're not able to do transactions. There's no options. There's no crypto. So it's a more limited implementation. Sure. Now again, I said earlier we at Public believe in sort of this model. Custom connectors, MCPs on sort of, you know, your favorite LLM interface and we've done the same thing. But, you know, the difference is we've built and allowed the sort of full power of of the Public platform. All the asset classes, all the transactions, you know, with great documentation that allows the LLMs directly to sort of work within our API stack to do what you want. Again, it can

interact with your individual account and you can you know, sort of do similar transactions. Again, it's it's what I said. There's a platform shift coming. I don't know, you know, where we think we're going to see a bunch of investors that want to use the public experience where it's more controlled and trusted. You know, we also think we're going to see a lot of investors that want to roll their own via API. We also think there's a middle way where people are really comfortable again using Cloud and using the the sort of UI that they have there and just pulling in sort of an MCP or custom connector in order to do that.

>> Yeah. Okay, so I have Open Claw running and I don't have a public.com account. I set up a public >> We should do that right now. >> up a public.com account and let's say I put in a grand and then I can essentially talk to the public.com API through my Open Claw and dictate the movement of my funds or give permission of the investment of my funds based on a given like whether it's risk tolerance strategy asset class whatever they you know, there's a maybe some questionnaire that happens effectively.

And and I can control all of that from my Open Claw. >> Yeah, totally. And so like there are a couple ways you can do it, right? There is an existing sort of Claw Hub Open Claw skill that we've built that allows Open Claw specifically to deal with it. But if you have again really any third party harness that you want to use grab the API key, grab the docs, give it to the model and they're smart enough now that they can do it and like honestly tell it to build its own skill. Hey, this is what I want you to focus on. Again, I think where we are now in the agentic world, it's all about giving it the right primitives which is, you know, a fully featured API and then full documentation for it to understand sort of what what what it does

and any of sort of the harnesses that are out there are going to do quick work. Again, the API by design so called first party API, you have it you have complete control of your account, your API keys it can and sort of your entire all of the accounts within your own user and you you know, you can basically do any transaction that you can do via clicking buttons and >> Multi-asset strategy that >> It doesn't really matter. It can do equities commodities crypto bonds. >> And it's all it's all like KYC'd and custodial. So it's just kind of like it's not the same as actually like So that seems easy to like knock down, right? Because it's like you guys have a back end you guys are able to kind of like have this AI manage this back end.

It's like credits and debits and it's kind of like this entire kind of system of data and then if you actually do that though on chain like on the blockchain where the agents are actually moving assets and buying and selling on a protocol or buying and selling on chain sending there's a whole another kind of like world of of like requirements there. >> Totally. >> Which is going to be like a I think a lot more difficult to figure out but also probably a lot more longer term fruitful for payments, for micro payments, for transactions, etc. 24/7 optionality.

I'm curious like do you guys see any transition to bringing in more like on chain offerings for public like just give us like your blunt view. Actually hold on. Just give us like public's blunt view of crypto. It's in the gutter right now. Like perspective is its perception is terrible. Like like what is public's view on crypto? >> Um you know, I listen. Like I said, we exist as an investing platform. >> Right. >> Right. Um you know, what >> You don't have a view. You have to be you have to love all your children. >> No, that's I mean that's right. Like again, we see crypto as an asset class first and foremost. It's not really like a um you know, not a method of payment.

Right? Not a currency. We see it as an asset that people want to buy and have exposure to as an asset class. Um and so all of our sort of decisions around crypto flow from there. Um you know, we've invested in uh you know, making it really easy, making it really secure, making sure we have all the order types that you might want, integrating it with all of our awesome AI tools, um building tax-advantaged crypto accounts, right? So, we have a crypto IRA. Basically, you can own spot crypto of I think we're up to like in, you know, 70-something tokens within your IRA, right? Great way to buy and hold any sort of spot crypto, and you know, we're one of the few platforms that can do that in probably the only one that can do that in sort of the broad-based platform uh that we have today. And so,

we've been invested in that as a, you know, viewing it as an asset class. Now, I do think we see over time as you get to more maturity in true DeFi opportunities where, hey, there might be individual opportunities that haven't crossed into the more centralized uh world, you know, we've dabbled in those, right? I remember a few years ago we were like, you know, we did some experimentation in NFTs. You know, we had a platform where um we actually offered sort of securitized interests in NFTs if you wanted them. Now, that world came and went pretty quickly, but we're always staying ahead of like, where are there interesting asset classes or interesting emerging asset classes from the crypto rails that we think we can make use of. Again, thinking about things like um

you know, tokenized securities, be those public companies uh that are tokenizing their own equity, or, you know, third parties tokenizing private company equities. Like, you know, we have sort of a dim view on that second one, but we'll we'll follow where the sort of uh capital markets infrastructure goes, um but again, always focused on it as asset class first and foremost. >> Yeah. Yeah, no, it's just like like like the world of using AI to kind of uh manage a portfolio is like is like one thing, and then the world of using AI to like manage an on-chain environment of like payments and transfers and and uh settling uh cross-border things, or doing trades 24/7 is like kind of like a whole new uh endeavor as well. When it comes to

actually like the portfolio management side, Um, what what is the most sought-after kind of uh or like what's what's been the most successful sought-after use case for AI? Is Is it just as simple as this kind of strategy that we talked about? Oh, I want to you know, execute this kind of more complicated strategy. Um, like what are some of the the highest uh success rates or what are some of the most uh used AI strategies? >> Yeah, so I think they're like, you know, um two categories of things that we're seeing most prevalently. The first is what you said. It's hey, I have a strategy I have an automated trading strategy. I either am executing and it takes me a lot of

time and I find it very annoying and I can just automate it. Or hey, I have one I would like to do, but it's so annoying I wouldn't want to do it on my [laughter] own. So like we do see again, a lot of it is that sort of automated trading and it's things that are like regular, you know, sort of regular always-on trading strategies. It's buying the dip, it's buying at a certain level, selling at a different level, RSI-based, you know, trigger-based, indicator-based. Uh it's options strategies where there's a defined time horizon and you need to be able to roll those and manage those on an ongoing basis. Like I think that's a really big one. I think the other thing I'm seeing a lot of is just um using an agent to to to find this sort of infinite customization of like a feature that already exists, right? So

like a um stop loss or take profit, right? Um you know, you can any trigger, any set of rules that you could imagine you could implement now. Dividend reinvestment is another one that um we're seeing a lot, which is like, hey, if I get uh my Apple dividend, I would really like to reinvest those proceeds in Google. And like you might be like, oh, why would you do that? Mm. That doesn't make any sense. But like people want to do it. I don't know.

>> Yeah, yeah, yes. Yeah, yeah, yeah. It's infinite optionality. Totally. And it's like >> Like whatever your heart >> desires. >> basically like all of these The way I think about it is like somebody that's been building this business over many years, you know, prior to the agentic world, any individual thing we wanted our customers to be able to do, >> Yeah. >> we had to have a product manager, a product designer, and a bunch of software developers build it. >> And it took a while. >> Totally. You want to build like a um you know, you we wanted to build one of my favorite order types and this is this is one that I always wanted, but I was like a man on an island screaming at our product team was like um you know, what what you call like a walk a walk the bid order or a walk order,

right? So I want to buy something and it's got a wide bid offer spread. Um I want to move up my limit price one penny every 10 seconds, every 15 seconds, every minute, whatever uh until I get executed right? >> Yeah. >> I want to just like walk it on in until I get the price that I want. Um and like, you know, we again, I'm one guy. We didn't get a lot of demand for it. It's not a big request. It was the thing I always wanted. And so like now, that's like the easy like tell the agent, >> That's what I want to do. >> you know, this is the option I want, you know, walk the bid walk the bid in in five cent, you know, in five cent increments every 15 seconds, you know, for the next 20 minutes.

>> Yep. >> Boom, done. >> Boom. >> And like there's stuff like that where it's like basically all these features it used to require a software you know, multiple software developers to build any of these features. Now it's like no, we actually just build like a real thin layer of like orchestration on top of all the features we already have. And it's just a new way to build features and a new way to help, you know, bring the power of all the infrastructure we've built to our members uh that isn't like a human building a big blue button where they can press it, you know. >> So you seem obviously very forward-thinking, innovative when it comes to using these tools. Kind of zooming out from the public.com product and more to the actual company and the people, you know, we saw Jack Dorsey lay off a

ton of employees from Square. we saw Coinbase, I just had a great conversation with the Kraken co-CEO. He talks about how right now they're undergoing a massive kind of shift and focusing on AI proficiency within their company as a core driver of whether they'll be at the company in the next 6 months to 12 months. You know, when it comes to being one of the executives at public.com, a company with hundreds of people, hundreds of people, how do you guys how are you guys reacting to the AI proficiency shift in the in the world to keep up with your competitors, to be lean, to be efficient? How do you manage your team in an era of AI proficiency being not a nice to have, a must have?

>> Yeah, I mean for us like I I I think there are kind of three big things. First, we are in um a space that we love because there's literally 10 years of work ahead of us to build. We could build, you know, there are so many products and features across sort of the capital markets in the investing space that we could continue to build into. And like we have to make hard decisions about what hits our road map every quarter. And if we get a few more things every quarter cuz AI makes us more productive, amazing. Amazing. Now, we've always run, second point, we've always run like a lean shop because we want to be able to deliver the best possible outcome to our customers, to our you know, to our members, and to our like our shareholders and investors. So, we've always run pretty lean and like hiring

people is never the first answer whenever we're trying to solve a problem. We don't throw people at a problem in that way. And like, you know, we are continuing to take all the amazing tools that are here, you know, that they're that exist here in order to give them to our awesome, you know, high talent density team members so that they can do more. And the last thing I'll say is the current state of uh the coding like that's the current state of the coding ability of these LLMs today is good. It's good for like internal tools, it's good for like QA and bug bashing. It's not good for trading systems, right? I think we've seen a few competitors from the crypto world let go a large portion of their engineering

team and maybe say things about vibe coding infrastructure. And like that is not advisable, right? And like we take very seriously our obligations to build really serious back-end infrastructure for accessing the capital markets. And like you can't vibe code that stuff. There's a level of detail, a level of expertise required. And like there's a level of trust, right? Like we have people trusting public with their life savings. Like I'm not going to risk losing that on, you know, something some code that Claude built. Now, it does make our our team members more effective and efficient, but like we're not going to get to a place, at least in the short run, where we're truly like, you know, letting Claude build an entire back-end feature really integrate with, um, you

know, major sensitive, uh, trading infrastructure. Now, this is probably the worst the technology will ever be. Maybe 2 years from now, uh, we will find it's capable of doing that and we will stay ahead of ahead of it again cuz we got so much to build. But, for now, you know, we don't think that's, you know, the best use of sort of these these, uh, these coding tools. >> Okay. And and so then I guess just to be very blunt, your competitors will use this technology, they will scale their companies with it, and they will try to take shortcuts, do things better, do things faster, you know, go to market faster, launch products faster. They will cut their costs more. Um,

but you're not at all looking like people are worried about layoffs. >> Yeah. >> I I don't want people in your company to be worried about getting laid off because of AI. >> No, no. And and again, I don't think >> I mean? >> Yeah, that's not how we're like we are in the stage where we will take every bit of technology and every bit of help to do more. >> Yeah. >> Like we have a long track record of outshipping our competitors despite having much smaller teams and we do that by having phenomenal people and giving them the independence to deliver, right? Um you know, we sort of organize our engineering team in a bunch of atomic units. We keep those units and teams as small as we can so they you know, so they can each build awesome individual

features and so every quarter we're getting six, seven, eight major features. Maybe 10, 12 and really effective and like we're going to see that ramp up and we're actually going to add more people because the ROI on each person goes up as you've got your best developers now not just operating with the you know, the sort of uh leverage of their own fingers writing code but they can actually start to leverage multiple other you know, sort of agents underneath them. So no, I think if anything it's going to probably lead to a growth in team but like I'm not worried about uh competitors Vibe coding uh trading software.

>> Yeah. >> Uh or or passing us in you know, again, we're 120. We have 120th the size of the engineering team as many of our our competitors and um we still outship them. >> outship them. >> And we still outship them and we will, right? And we're not you know, we we sit in a wonderful spot where we just have such good people and uh such a great a great org design and structure where we're just an execution machine and like that's going to that's not going to change. It's only going to get better with AI. In fact, it's only going to allow us to invest in more people and more teams to continue to be more effective and my hope is that more of our competitors do Vibe code stuff because that level of reliability that we you know, that we deliver while

having uh a high a high pace of uh delivery um is unmatched. >> Yeah. >> You know, all we need is uh more of our competitors to to fall on their face from a reliability perspective and you know, we'll be we'll be welcoming it. >> Yeah. Yeah, so there's a there's a kind of broad you know, meta question going on here which is like this AI displacement you know, that's been happening and um in in in in in DeFi we're seeing exploits to the fullest extent. You know, tens of thousands, millions, tens of millions, hundreds of millions of dollars. Some of them linked to some sort of like AI honeypot software. Uh in the traditional markets we're seeing memory stocks, um we're seeing chip stocks, we're seeing manufacturer stocks, energy, everything

just fly. Um we're seeing layoffs. So, like the AI displacement is already happening based on inc capital gains inequality, market inefficiencies, layoffs. You know, when it just comes like holistically as we kind of wrap, what is your current view on AI displacement for human labor? I think there's something to the tune of $11 trillion on a yearly basis of kind of human labor as as part of a capital force. How does this economic force shake up going forward? What is your like broad view holistically on AI displacement?

>> I again comes back to like sort of how I what I see at Public. Um we are a growth-oriented firm. We will take every bit of advantage we have, every bit of new technology to grow. >> Yeah. >> Um we will take every dollar of savings that we get, every incremental unit of productivity, and we will reinvest it in the business because we see such a big opportunity out of us. Um and I think that is I believe that that mentality that we have is actually shared across the economy. I believe there are thousands of companies that have the same sort of mindset as us, and they see such big opportunities in their own individual segments that they're going to keep investing behind it. And that means, sure, they're going to spend more money on tokens, but they're also

going to spend more money on people as that unlocks more new businesses for them to be in, more new opportunities for them to go after, more sales people to sell the new products that they need, right? Again, I don't think Claude's picking up the phone and calling and and executing, you know, major sales. So, again, I think like listen, like there are sort of I don't know, generally speaking, there are two parts, you know, there are two styles of business. There's one that exists to sort of build profitability and return profits to shareholders. Those firms might take the efficiency, they might let people go, they might, you know, push those, uh, you know, efficiency improvements down to the the bottom line. Um, but on the other side, there's the growth-oriented firms that are going to take every bit of that efficiency to grow faster and

build new businesses. And the experience of the American economy is that the second one always wins. The growth The growth-oriented firms always take more share, always build bigger businesses, and always consume more of that resource and labor that's let off from the profitability focused side to continue to grow uh the economy over time. And like I think that's going to continue to be true. Um, you know, time will tell if this is this time is somehow different, but again, I'm going to I'm going to trust in like what I see in the way we approach the work every day, and the way I see all the companies we work with driving that sort of growth. Um, I'm going to trust that that exists and that we're all going to be really excited to build more new cool stuff,

more new awesome businesses, and honestly, make more money together over time, rather than sort of take these like narrow profits without getting the growth in top line. >> Yeah. Very interesting, Stephen. Very, very interesting times. Quite a timely show today as well. Thanks for Thanks for joining us. Great points. Uh, stoked to see the success of public.com. Going to try it out. I I've been dying to get my hands on open claw to finance kind of movement happening. So, stoked to try it out. >> For sure. For you and the audience, man, I'm an easy guy to find uh on X, it's Sykes, s i k e s underscore. And then we have our own subreddit, uh, public app, r/publicapp. I'm a mod there and very active. So, come find me if you're interested in public. Happy to help you

get set up or uh set you up with one of my teammates who can get you uh everything you need to to get your first agentic trading strategy rolled out. >> Yeah. See if it'll outperform me. Ladies and gentlemen, that's been AI Supercycle episode number 35.