This conversation at theCUBE and NYSE Wired Mixture of Experts examines agentic artificial intelligence and retail investing. Stephen Sikes of Public joins hosts to discuss AI-driven portfolio construction, trading automation, persona creation and community-driven research. Sikes describes Public's approach to observable deterministic agent workflows that remove AI discretion once a user approves parameters; they highlight the company's progression from large language model-powered research to an AI-managed generated assets portfolio and toward agentic brokerage.
Sikes notes Public's reliance on the Anthropic Claude family, including Opus, Sonnet and Haiku, while testing other models. They outline a three-pronged roadmap: API, MCP and native agentic brokerage. Sikes also advises retail investors to consider float and market mechanics when engaging with new issues and underscores the importance of observability for compliance and risk management in agentic systems.
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
theCUBE + NYSE Wired: Mixture of Experts Series. If you don’t think you received an email check your
spam folder.
Sign in to theCUBE + NYSE Wired: Mixture of Experts Series.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open this link to automatically sign into the site.
Register For theCUBE + NYSE Wired: Mixture of Experts Series
Please fill out the information below. You will recieve an email with a verification link confirming your registration. Click the link to automatically sign into the site.
You’re almost there!
We just sent you a verification email. Please click the verification button in the email. Once your email address is verified, you will have full access to all event content for theCUBE + NYSE Wired: Mixture of Experts Series.
I want my badge and interests to be visible to all attendees.
Checking this box will display your presense on the attendees list, view your profile and allow other attendees to contact you via 1-1 chat. Read the Privacy Policy. At any time, you can choose to disable this preference.
Select your Interests!
add
Upload your photo
Uploading..
OR
Connect via Twitter
Connect via Linkedin
EDIT PASSWORD
Share
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
theCUBE + NYSE Wired: Mixture of Experts Series. If you don’t think you received an email check your
spam folder.
Sign in to theCUBE + NYSE Wired: Mixture of Experts Series.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open this link to automatically sign into the site.
Sign in to gain access to theCUBE + NYSE Wired: Mixture of Experts Series
Please sign in with LinkedIn to continue to theCUBE + NYSE Wired: Mixture of Experts Series. Signing in with LinkedIn ensures a professional environment.
Are you sure you want to remove access rights for this user?
Details
Manage Access
email address
Community Invitation
Stephen Sikes, Public
This conversation at theCUBE and NYSE Wired Mixture of Experts examines agentic artificial intelligence and retail investing. Stephen Sikes of Public joins hosts to discuss AI-driven portfolio construction, trading automation, persona creation and community-driven research. Sikes describes Public's approach to observable deterministic agent workflows that remove AI discretion once a user approves parameters; they highlight the company's progression from large language model-powered research to an AI-managed generated assets portfolio and toward agentic brokerage.
Sikes notes Public's reliance on the Anthropic Claude family, including Opus, Sonnet and Haiku, while testing other models. They outline a three-pronged roadmap: API, MCP and native agentic brokerage. Sikes also advises retail investors to consider float and market mechanics when engaging with new issues and underscores the importance of observability for compliance and risk management in agentic systems.
>> Palo Alto Studio Connection, Silicon Valley, and Wall Street. I'm John Furrier, host of theCUBE, here with Dave Vellante, my co-host.
Gemma Allen
>> Welcome back to theCUBE Studio here at the New York Stock Exchange. I'm Gemma Allen with NYSC Wired's Mixture of Experts. And joining me now for a conversation on all thing retail investment and the world of AI is Stephen Sikes, COO of Public. Welcome, Stephen.
Stephen Sikes
>> Yeah, thanks for having me. Very excited.
Gemma Allen
>> So Public claims to be the world's first agentic AI platform.
Stephen Sikes
>> Yeah.
Gemma Allen
>> Break that down. What exactly does that mean?
Stephen Sikes
>> Yeah, so we've been on a journey, I think, over the last three or four years, call it since mid 2022, to figure out how really serious investors, people that are focused on building their wealth, having great outcomes in the market, either through long-term investing or even trading, how they're using AI. And I think we started in post-ChatGPT era with sort of investment research, and we realized, "Hey, that's really useful." AI and LLMs can be a great way to summarize the market, take all the myriad information, of which there's more than there ever has been before, and synthesize it into something that's useful for investors and traders. Then what we quickly realized is people were asking the AI and the LLM to do more stuff. And so we started to build in that direction, towards what we saw people interacting with AI and wanting from sort of AI within that investing context. And that led us to our Generated Assets platform, which is effectively a portfolio manager driven by AI. So you tell the Generated Assets platform what theme, metrics, et cetera, you'd like to invest in, it'll build you a long-term diversified portfolio and manage that for you over time, which is really cool. Then what we saw coming out of the back of that was, okay, we've done this research, we've done portfolio management. We really saw a heavy demand, especially in the wake of the agentic revolution that started in November around Claude Opus 4.5 being released. We saw a bunch of people coming to us and saying, "Hey, how do I automate more of my trading, more of my transactions? How do I use these sort of generalizable LLM capabilities on top of the full breadth of the platform that Public offered? So how do I automate my money movements? How do I automate my trading strategies? How do I automate my options selling strategies? How do I automate sweeping funds into treasuries? Because that's something that I like to do." And it became very clear that's where we had to go. And so literally, I think in early December, we started writing code on our own internal harness, our own internal agents for the purposes of investing and basically using the full breadth of the Public platform to invest straight on using an agent or sort of an LLM.
Gemma Allen
>> So let's talk about what's happening in the industry broadly, right?
Stephen Sikes
>> Sure.
Gemma Allen
>> You mentioned it before. We also had OpenClaw, all of the excitement around that. Let me maybe share a story and you tell me your observation. We had a visitor here at the NYSC a while back who was visiting from university, and he shared with us that he was using OpenClaw to do college applications for him-
Stephen Sikes
>> Sure....
Gemma Allen
>> while he was in New York. And we said, "Are you using it for anything else?" And he said, "Actually, yeah. My granddad's a big investor. He doesn't invest in crypto, so I have it investing in crypto for him." It's connected to his Fidelity account, right?
Stephen Sikes
>> That's cool.
Gemma Allen
>> And I thought, "Wow. That's cool, but also kind of scary." Okay?
Stephen Sikes
>> Totally.
Gemma Allen
>> What are your thoughts on that story? If you were talking to that young man, what would you say?
Stephen Sikes
>> I mean, I think zooming out, the reality is there are many different use cases and many different styles. There will be some people, like the gentlemen you discussed, who are, what it sounds like cool, giving an agent permissions to go trade on their behalf according to some structure. Now, I don't know that that's the majority of people. And we have a platform that lets people do that. We have a fully featured and fully documented API that you can drop those docs, you can give access to those docs and that API. You can give your API key to OpenClaw and let it go wild. I don't know if that's the best model. I know that's what most people are doing. But for people that are sort of on that leading bleeding edge of AI usage, we are seeing some of that. I think where we think more people are right now, and I think where we see more opportunity, is building sort of those controls and that experience within the four walls of a brokerage app or a brokerage experience. And that's what we've invested in in agentic brokerage. And so what we've found for most people is they want the trust and credibility of a platform, plus they want a human in the loop. They want a deterministic outcome of the sort of strategy they're working with AI on. So the way our platform works is you will discuss with our agent what you're interested in, give it your intent, give it your strategy. "Hey, I want to run a covered call strategy on top of a big holding, a position of mine. Hey, I actually want to hedge this position using puts. Hey, I want to sweep... Like I said, I want to sweep funds at the end of the day into treasuries."
What you'll do is you'll have a conversation with our agent about how to do that. You'll discuss the specific parameters, if-this, then-that style. And then when you're ready, you'll commit that into a deterministic workflow where you'll see every step that'll be resolved deterministically. So once you get into that workflow, the AI doesn't have any more judgment. The AI doesn't have any more discretion. It's just your partner and building sort of the workflow using its amazing tools for doing so. And so you get there, you hit the button to authorize or to submit your agent and start it running, and then everything runs according to the pre-planned workflow that you've described. So no opportunities for hallucination, no potential for it trading without your consent or without your understanding. And then you can see on the back end every single not just transaction that gets submitted, but every step in that if-this, then-that sort of flow, be it alerts or triggers that you put in place. And so we've built, again, with that maximum of observability plus sort of the deterministic nature of what you would expect from a sort of trading strategy engine.
Gemma Allen
>> I want to get into the observability and also the-
Stephen Sikes
>> Sure....
Gemma Allen
>> persona creation. But before I do, I also want to ask you about something that I don't think gets asked enough, and that is the profile of investors-
Stephen Sikes
>> Sure....
Gemma Allen
>> in this world of AI, right? We know that there has been a mass portion of society that were very left behind by the previous index cycles and ways in which the Robinhood era where folks were suddenly able to make some clever moves with their savings, et cetera.
Stephen Sikes
>> Right.
Gemma Allen
>> Public was set up around that, right? Democratizing access to investment. What sorts of new profiles are you seeing? Or if we're being honest, is it really the same type of individuals, the same type of risk-takers from-
Stephen Sikes
>> Yeah....
Gemma Allen
>> 2019 who were truly leading the charge here? Break it down a bit.
Stephen Sikes
>> Yeah. I mean, I think we are... The reality is I will say two things. One, we are certainly still in the early adopter phase, right? I think the people who are using AI to invest are the same sort of risk on people. Again, I think it's perceived to be more risky. I think what we have a job to do as an industry, and we at Public take very seriously, is starting to use these AI tools to make more advanced, higher-quality strategies and executions available for the masses. I think there are a bunch of investment strategies that maybe were inaccessible to investors or very expensive that now an AI can do on your behalf quite easily. I'll give you a good example, a very vanilla glide path asset allocation strategy. Say you're a long buy-and-hold investor. I just want to put money in every time, and I want something to manage my portfolio according to my age and risk tolerance over time. That sounds like something you would typically pay a robo-advisor or a financial advisor to do. Now, with an agent sitting on top of your brokerage account, you don't really need that, right? Working with Public's agent, you can very easily give it the parameters, "Hey, I'm this old. Hey, I have this risk tolerance. Hey, I have this time horizon. Build me a well-diversified portfolio of ETFs. And then every day, week, month, quarter, check my asset allocation and rebalance it accordingly." And you know what? Does a pretty good job, right? Not to say that there aren't opportunities for humans to add value in that loop or really expertise to matter, but if you sort of know and understand that you should sort of own something well diversified over time, the agents can really help you get to the best implementation of that, again, according to your intent and strategy.
Gemma Allen
>> And when you are thinking about your competitive place in this market, in attention and loyalty and trust, which is a big, big part of this-
Stephen Sikes
>> Totally....
Gemma Allen
>> how are you thinking about using Public's option versus, for example, people who are just using a $200 prescription for Perplexity or-
Stephen Sikes
>> Totally....
Gemma Allen
>> Claude or... And I mean, let's be frank. A lot of folks probably don't know how to go about setting up a Mac Mini, right?
Stephen Sikes
>> Totally. Yeah.
Gemma Allen
>> There is that side of it, too. It's not as simple as it sounds.
Stephen Sikes
>> Right.
Gemma Allen
>> But how are you guys thinking about that? Obviously, what is the core USP?
Stephen Sikes
>> Yeah. The reality is we are seeing major demand for investors and traders using AI and who want to use AI and who want to use it in their investing processes, be it research, portfolio management, or actually trading and execution. We're not totally sure how people are going to want to do that in the long run. And so that's why I said earlier, we've built the API version. We actually have an MCP that's sort of two clicks directly integrated within your Clause or ChatGPT experience, and then we have the full sort of native agentic experience. If you ask me to bet, I think you're going to see more people wanting the native agentic experience in the trusted, highly credible, deterministic platform, but I don't know that to be true. I think a lot of that comes back to what is the level of trust that investors have in sort of the foundational models? Because if we find, hey, Claude continues to improve at the rate that it has, in three years from now, it might be better than just about anything else and fully trustworthy and never makes mistakes. If we get to that world, I think we might see more of the trading volume and more Claude actually eating more of the industry there, but we don't know. And so I think we're sort of building our business in a way that allows for sort of all of those futures to play out and for us to be well positioned to take advantage of any of them.
Gemma Allen
>> Hedging your bets, as you well should.
Stephen Sikes
>> You know.
Gemma Allen
>> You mentioned Claude a lot there.
Stephen Sikes
>> Sure.
Gemma Allen
>> Talk to me about... Let's go under the hood a second for what you guys-
Stephen Sikes
>> Sure....
Gemma Allen
>> have built in house, right? That tech, that stack. What are you seeing? I presume you're probably playing around with a whole lot of different models. Are you using open weight models? Are you heavy in on the Anthropic journey-
Stephen Sikes
>> Yeah....
Gemma Allen
>> which a lot of enterprises seem to be? Why, if so?
Stephen Sikes
>> Yeah.
Gemma Allen
>> Give me your thoughts.
Stephen Sikes
>> We are. I would say, very reasonably and fairly, we are very heavy in the Anthropic universe. I think Public's in-house agents are all built on top of the Claude family, some mix of occasionally Haiku, often Sonnet, and very often Opus. And again, we've got our own orchestration and harness layer there that has a bunch of logic on top of them. Now, we have done some experiments. And every time there's a new release from a major model firm, we are testing it against our own evaluations to see, "Hey, where's the cost-benefit trade-off? Are there opportunities for this to improve over time?"
And we have seen ChatGPT 5.5. It's quite high quality. It is more expensive than sort of Opus 4.8, but I think you can see a lot of replaceability there. And so we do think there is an opportunity for OpenAI and ChatGPT to take the crown again for these agentic workflows. But at the moment, on a cost and quality access, I think Opus and the Claude family of models are still the best.
Gemma Allen
>> And when we think about the access to data information on intelligible signals that existed five, six years ago, right, like Bloomberg Terminal, for example, there's a reason that institutional investors are really high-end family , et cetera. They had access to data because they bought that data, right?
Stephen Sikes
>> Right.
Gemma Allen
>> That is... You know?
Stephen Sikes
>> Totally.
Gemma Allen
>> Let's be frank. How do you see that changing from the perspective of AI? Do you see this mass opportunity for more availability of data that's actually intelligible and can be used for the right purpose and signal for all?
Stephen Sikes
>> Totally.
Gemma Allen
>> Or do you think we have the same-
Stephen Sikes
>> Totally. And I don't know. You know, I mean, I think there's one... Right now, I guess for retail investors right now, we've never had better and more access to market data, fundamental data, analysis, content, right? I think, again, this is sort of like internet 2.0 stuff, right? We've seen this massive explosion of all this information available for investors, more than most people could consume and orders of magnitude more than might have been available 20 years ago. I remember when I started my investing journey, like getting real-time data, not a delayed feed, but true real-time quotes was like a thing you had to deposit a lot of money or even pay for at your broker. Now that's like you can go anywhere on the internet and find real-time quotes. And it's been a wild evolution even just to see that basic thing, but I think we have seen that explosion in information. Now again, I think that's mostly the force of the internet more so than AI. What I think's going to be interesting from here is a lot of the magic of Bloomberg is actually the sort of hard work they've done to manually gather information across sort of the markets and across the economy and pulling that into structured data sets within the Bloomberg terminal. It's things like 13Fs and going and understanding shareholders and who the actual owners of an end firm are and pulling in all of this myriad information out of SEC reports. And a lot of that was done manually, right? There were teams upon teams upon teams within Bloomberg whose job it was to go and grab all this data structure and make sure it was available within the terminal. That's a job AI is great at. And so I think we've never had a better set of technologies for competing with Bloomberg. And now listen, I don't think there's actual competition of Bloomberg. Bloomberg has its own sort of differentiators that are above and beyond even the data layer, but I do think we will see things that approximate that data layer or even match sort of the Bloomberg data layer that are available to retail investors because it's easier to create that using AI than it has ever been before.
Gemma Allen
>> Talk about where these conversations are happening and how that's changing, right? You mentioned 10 years ago, a lot of folks used Reddit. We had the whole crypto punk way for folks. Got a lot of great tips on X, for example, on these different platforms which aren't necessarily Bloomberg, right?
Stephen Sikes
>> Right.
Gemma Allen
>> But people made a lot of money. That world, too, is changing. And yeah, there's so much more data, but there's also potentially a risk with verification, right?
Stephen Sikes
>> Right.
Gemma Allen
>> A risk with false data or misleading data, et cetera. What are you seeing? Where do you think retail investors are going to talk these days?
Stephen Sikes
>> Yeah, it's interesting. I think there are two main sources right now. One, I do think Reddit is still the most vibrant community for investors across all of the myriad channels that are there that match every sort of style of investing you can imagine. Companies like Public, we have our own subreddit that we interact with our members. And it's one of the places I spend the most amount of time interacting with members. I think that's still an amazing and probably the best community of people. I think beyond that, you have Discord, which I think has its own connotations as being closely connected to the crypto universe. But I think we do see a lot of the trading communities and the real fast-paced conversations happening within Discords, which is interesting. And it's a place where we've invested time in getting to know the individual moderators and the individual communities and where we need to be present to meet our members. And then the third, obviously, still X, Twitter. Still a lot of phenomenal conversations. Definitely more on the sort of analytical side and slightly longer form than sort of the chat styles that you see in Discord or the more community forum style that you see in Reddit. But I think those are the three platforms where we definitely see the most conversations from retail investors.
Gemma Allen
>> Let's talk about what's happening in the market for a second.
Stephen Sikes
>> Sure.
Gemma Allen
>> I mean, we had SpaceX took a bit of a-
Stephen Sikes
>> ....
Gemma Allen
>> dip yesterday. I haven't seen it yet today, but it's... Actually, I'm looking right now. It's down. Insane story, though, right?
Stephen Sikes
>> Right.
Gemma Allen
>> When we think about, again, 10 years ago versus now, people say it's a good time to buy boring stocks. I mean, from your perspective, what are you seeing and hearing from retail investors? Where is the appetite or even the opportunity or the access? What are your thoughts? Do you think the Mag Seven is exactly the most it has always been?
Stephen Sikes
>> Yeah. I mean, listen, I try very hard not to make predictions about investments or stock prices because I'm usually wrong. Every once in a while I'll do it.
Gemma Allen
>> You and me both.
Stephen Sikes
>> Somebody will pin me down on something, and I'll have to watch that video getting reposted to my timeline every few months. No. I mean, I think what we're seeing from retail, SpaceX specifically, tremendous amount of interest in demand. I think it's unique in that we have an entrepreneur, a founder, an operator, an executive in Elon who has done such an amazing job for retail investors over the last 15, almost 20 years in sort of running and building Tesla. Tesla has always been well-owned by retail, and retail investors have been wildly rewarded for that loyalty. And I think we're seeing a lot of those same people invest heavily in SpaceX again on the same sort of thesis and on the trust that Elon will do well for them over time. It certainly is a different valuation profile from even Tesla, which is quite healthily valued at the time, so we'll see. I think my caution always to retail investors in this situation, especially with new issues and IPOs, is to be thoughtful about the float relative to the volume that may be coming on over time, right? SpaceX came out with a relatively low float. I think we see a lot of unlocks coming over the next 60, 90, 120 days. And even with the options opening, last week we started to see a little better price discovery. And so I'm always counseling people to pay attention to things like that, that it's not necessarily shown on the chart, but it's something you sort of have to understand about the market overall and how float and flows can have a major impact on price movements and waiting for sort of the float... More of the float to be in the market, I think, is always prudent.
Gemma Allen
>> Love it. Okay. So Stephen, what's ahead? Talk to us. Last question, what does the next two years look like for you guys? Are you continuing to build proprietary technology? What's the product roadmap? And what are the big bets?
Stephen Sikes
>> Totally. I mean, I think for us, again, the biggest thing is continuing to capitalize on this AI transition, right? We saw prior generations transition from calling your broker to making your trades online, from making your trades online and on the web to doing it on your phone. Now we think you're going to move from doing it on your phone to doing it with AI in the loop and having that sort of AI-intermediated experience. And so we're going to continue to invest in building more and more tools on all three of the pillars that we have, both our API, our MCP, and then our agentic brokerage that's sort of native to Public. And again, continue to build more and more features into each of them, making them higher quality, easier to work with, and, again, more observable and deterministic in their execution I think are all sort of where we're going to be spending our time.
Gemma Allen
>> Well, we certainly look forward to watching. Stephen, thank you so much for coming on theCUBE.
Stephen Sikes
>> Thanks for having me.
Gemma Allen
>> I'm Gemma Allen here at theCUBE Studio at the New York Stock Exchange. This is NYSC Wired's Mixture of Experts. Thanks for watching.
>> Palo Alto Studio Connection, Silicon Valley, and Wall Street. I'm John Furrier, host of theCUBE, here with Dave Vellante, my co-host.
Gemma Allen
>> Welcome back to theCUBE Studio here at the New York Stock Exchange. I'm Gemma Allen with NYSC Wired's Mixture of Experts. And joining me now for a conversation on all thing retail investment and the world of AI is Stephen Sikes, COO of Public. Welcome, Stephen.
Stephen Sikes
>> Yeah, thanks for having me. Very excited.
Gemma Allen
>> So Public claims to be the world's first agentic AI platform.
Stephen Sikes
>> Yeah.
Gemma Allen
>> Break that down. What exactly does that mean?
Stephen Sikes
>> Yeah, so we've been on a journey, I think, over the last three or four years, call it since mid 2022, to figure out how really serious investors, people that are focused on building their wealth, having great outcomes in the market, either through long-term investing or even trading, how they're using AI. And I think we started in post-ChatGPT era with sort of investment research, and we realized, "Hey, that's really useful." AI and LLMs can be a great way to summarize the market, take all the myriad information, of which there's more than there ever has been before, and synthesize it into something that's useful for investors and traders. Then what we quickly realized is people were asking the AI and the LLM to do more stuff. And so we started to build in that direction, towards what we saw people interacting with AI and wanting from sort of AI within that investing context. And that led us to our Generated Assets platform, which is effectively a portfolio manager driven by AI. So you tell the Generated Assets platform what theme, metrics, et cetera, you'd like to invest in, it'll build you a long-term diversified portfolio and manage that for you over time, which is really cool. Then what we saw coming out of the back of that was, okay, we've done this research, we've done portfolio management. We really saw a heavy demand, especially in the wake of the agentic revolution that started in November around Claude Opus 4.5 being released. We saw a bunch of people coming to us and saying, "Hey, how do I automate more of my trading, more of my transactions? How do I use these sort of generalizable LLM capabilities on top of the full breadth of the platform that Public offered? So how do I automate my money movements? How do I automate my trading strategies? How do I automate my options selling strategies? How do I automate sweeping funds into treasuries? Because that's something that I like to do." And it became very clear that's where we had to go. And so literally, I think in early December, we started writing code on our own internal harness, our own internal agents for the purposes of investing and basically using the full breadth of the Public platform to invest straight on using an agent or sort of an LLM.
Gemma Allen
>> So let's talk about what's happening in the industry broadly, right?
Stephen Sikes
>> Sure.
Gemma Allen
>> You mentioned it before. We also had OpenClaw, all of the excitement around that. Let me maybe share a story and you tell me your observation. We had a visitor here at the NYSC a while back who was visiting from university, and he shared with us that he was using OpenClaw to do college applications for him-
Stephen Sikes
>> Sure....
Gemma Allen
>> while he was in New York. And we said, "Are you using it for anything else?" And he said, "Actually, yeah. My granddad's a big investor. He doesn't invest in crypto, so I have it investing in crypto for him." It's connected to his Fidelity account, right?
Stephen Sikes
>> That's cool.
Gemma Allen
>> And I thought, "Wow. That's cool, but also kind of scary." Okay?
Stephen Sikes
>> Totally.
Gemma Allen
>> What are your thoughts on that story? If you were talking to that young man, what would you say?
Stephen Sikes
>> I mean, I think zooming out, the reality is there are many different use cases and many different styles. There will be some people, like the gentlemen you discussed, who are, what it sounds like cool, giving an agent permissions to go trade on their behalf according to some structure. Now, I don't know that that's the majority of people. And we have a platform that lets people do that. We have a fully featured and fully documented API that you can drop those docs, you can give access to those docs and that API. You can give your API key to OpenClaw and let it go wild. I don't know if that's the best model. I know that's what most people are doing. But for people that are sort of on that leading bleeding edge of AI usage, we are seeing some of that. I think where we think more people are right now, and I think where we see more opportunity, is building sort of those controls and that experience within the four walls of a brokerage app or a brokerage experience. And that's what we've invested in in agentic brokerage. And so what we've found for most people is they want the trust and credibility of a platform, plus they want a human in the loop. They want a deterministic outcome of the sort of strategy they're working with AI on. So the way our platform works is you will discuss with our agent what you're interested in, give it your intent, give it your strategy. "Hey, I want to run a covered call strategy on top of a big holding, a position of mine. Hey, I actually want to hedge this position using puts. Hey, I want to sweep... Like I said, I want to sweep funds at the end of the day into treasuries."
What you'll do is you'll have a conversation with our agent about how to do that. You'll discuss the specific parameters, if-this, then-that style. And then when you're ready, you'll commit that into a deterministic workflow where you'll see every step that'll be resolved deterministically. So once you get into that workflow, the AI doesn't have any more judgment. The AI doesn't have any more discretion. It's just your partner and building sort of the workflow using its amazing tools for doing so. And so you get there, you hit the button to authorize or to submit your agent and start it running, and then everything runs according to the pre-planned workflow that you've described. So no opportunities for hallucination, no potential for it trading without your consent or without your understanding. And then you can see on the back end every single not just transaction that gets submitted, but every step in that if-this, then-that sort of flow, be it alerts or triggers that you put in place. And so we've built, again, with that maximum of observability plus sort of the deterministic nature of what you would expect from a sort of trading strategy engine.
Gemma Allen
>> I want to get into the observability and also the-
Stephen Sikes
>> Sure....
Gemma Allen
>> persona creation. But before I do, I also want to ask you about something that I don't think gets asked enough, and that is the profile of investors-
Stephen Sikes
>> Sure....
Gemma Allen
>> in this world of AI, right? We know that there has been a mass portion of society that were very left behind by the previous index cycles and ways in which the Robinhood era where folks were suddenly able to make some clever moves with their savings, et cetera.
Stephen Sikes
>> Right.
Gemma Allen
>> Public was set up around that, right? Democratizing access to investment. What sorts of new profiles are you seeing? Or if we're being honest, is it really the same type of individuals, the same type of risk-takers from-
Stephen Sikes
>> Yeah....
Gemma Allen
>> 2019 who were truly leading the charge here? Break it down a bit.
Stephen Sikes
>> Yeah. I mean, I think we are... The reality is I will say two things. One, we are certainly still in the early adopter phase, right? I think the people who are using AI to invest are the same sort of risk on people. Again, I think it's perceived to be more risky. I think what we have a job to do as an industry, and we at Public take very seriously, is starting to use these AI tools to make more advanced, higher-quality strategies and executions available for the masses. I think there are a bunch of investment strategies that maybe were inaccessible to investors or very expensive that now an AI can do on your behalf quite easily. I'll give you a good example, a very vanilla glide path asset allocation strategy. Say you're a long buy-and-hold investor. I just want to put money in every time, and I want something to manage my portfolio according to my age and risk tolerance over time. That sounds like something you would typically pay a robo-advisor or a financial advisor to do. Now, with an agent sitting on top of your brokerage account, you don't really need that, right? Working with Public's agent, you can very easily give it the parameters, "Hey, I'm this old. Hey, I have this risk tolerance. Hey, I have this time horizon. Build me a well-diversified portfolio of ETFs. And then every day, week, month, quarter, check my asset allocation and rebalance it accordingly." And you know what? Does a pretty good job, right? Not to say that there aren't opportunities for humans to add value in that loop or really expertise to matter, but if you sort of know and understand that you should sort of own something well diversified over time, the agents can really help you get to the best implementation of that, again, according to your intent and strategy.
Gemma Allen
>> And when you are thinking about your competitive place in this market, in attention and loyalty and trust, which is a big, big part of this-
Stephen Sikes
>> Totally....
Gemma Allen
>> how are you thinking about using Public's option versus, for example, people who are just using a $200 prescription for Perplexity or-
Stephen Sikes
>> Totally....
Gemma Allen
>> Claude or... And I mean, let's be frank. A lot of folks probably don't know how to go about setting up a Mac Mini, right?
Stephen Sikes
>> Totally. Yeah.
Gemma Allen
>> There is that side of it, too. It's not as simple as it sounds.
Stephen Sikes
>> Right.
Gemma Allen
>> But how are you guys thinking about that? Obviously, what is the core USP?
Stephen Sikes
>> Yeah. The reality is we are seeing major demand for investors and traders using AI and who want to use AI and who want to use it in their investing processes, be it research, portfolio management, or actually trading and execution. We're not totally sure how people are going to want to do that in the long run. And so that's why I said earlier, we've built the API version. We actually have an MCP that's sort of two clicks directly integrated within your Clause or ChatGPT experience, and then we have the full sort of native agentic experience. If you ask me to bet, I think you're going to see more people wanting the native agentic experience in the trusted, highly credible, deterministic platform, but I don't know that to be true. I think a lot of that comes back to what is the level of trust that investors have in sort of the foundational models? Because if we find, hey, Claude continues to improve at the rate that it has, in three years from now, it might be better than just about anything else and fully trustworthy and never makes mistakes. If we get to that world, I think we might see more of the trading volume and more Claude actually eating more of the industry there, but we don't know. And so I think we're sort of building our business in a way that allows for sort of all of those futures to play out and for us to be well positioned to take advantage of any of them.
Gemma Allen
>> Hedging your bets, as you well should.
Stephen Sikes
>> You know.
Gemma Allen
>> You mentioned Claude a lot there.
Stephen Sikes
>> Sure.
Gemma Allen
>> Talk to me about... Let's go under the hood a second for what you guys-
Stephen Sikes
>> Sure....
Gemma Allen
>> have built in house, right? That tech, that stack. What are you seeing? I presume you're probably playing around with a whole lot of different models. Are you using open weight models? Are you heavy in on the Anthropic journey-
Stephen Sikes
>> Yeah....
Gemma Allen
>> which a lot of enterprises seem to be? Why, if so?
Stephen Sikes
>> Yeah.
Gemma Allen
>> Give me your thoughts.
Stephen Sikes
>> We are. I would say, very reasonably and fairly, we are very heavy in the Anthropic universe. I think Public's in-house agents are all built on top of the Claude family, some mix of occasionally Haiku, often Sonnet, and very often Opus. And again, we've got our own orchestration and harness layer there that has a bunch of logic on top of them. Now, we have done some experiments. And every time there's a new release from a major model firm, we are testing it against our own evaluations to see, "Hey, where's the cost-benefit trade-off? Are there opportunities for this to improve over time?"
And we have seen ChatGPT 5.5. It's quite high quality. It is more expensive than sort of Opus 4.8, but I think you can see a lot of replaceability there. And so we do think there is an opportunity for OpenAI and ChatGPT to take the crown again for these agentic workflows. But at the moment, on a cost and quality access, I think Opus and the Claude family of models are still the best.
Gemma Allen
>> And when we think about the access to data information on intelligible signals that existed five, six years ago, right, like Bloomberg Terminal, for example, there's a reason that institutional investors are really high-end family , et cetera. They had access to data because they bought that data, right?
Stephen Sikes
>> Right.
Gemma Allen
>> That is... You know?
Stephen Sikes
>> Totally.
Gemma Allen
>> Let's be frank. How do you see that changing from the perspective of AI? Do you see this mass opportunity for more availability of data that's actually intelligible and can be used for the right purpose and signal for all?
Stephen Sikes
>> Totally.
Gemma Allen
>> Or do you think we have the same-
Stephen Sikes
>> Totally. And I don't know. You know, I mean, I think there's one... Right now, I guess for retail investors right now, we've never had better and more access to market data, fundamental data, analysis, content, right? I think, again, this is sort of like internet 2.0 stuff, right? We've seen this massive explosion of all this information available for investors, more than most people could consume and orders of magnitude more than might have been available 20 years ago. I remember when I started my investing journey, like getting real-time data, not a delayed feed, but true real-time quotes was like a thing you had to deposit a lot of money or even pay for at your broker. Now that's like you can go anywhere on the internet and find real-time quotes. And it's been a wild evolution even just to see that basic thing, but I think we have seen that explosion in information. Now again, I think that's mostly the force of the internet more so than AI. What I think's going to be interesting from here is a lot of the magic of Bloomberg is actually the sort of hard work they've done to manually gather information across sort of the markets and across the economy and pulling that into structured data sets within the Bloomberg terminal. It's things like 13Fs and going and understanding shareholders and who the actual owners of an end firm are and pulling in all of this myriad information out of SEC reports. And a lot of that was done manually, right? There were teams upon teams upon teams within Bloomberg whose job it was to go and grab all this data structure and make sure it was available within the terminal. That's a job AI is great at. And so I think we've never had a better set of technologies for competing with Bloomberg. And now listen, I don't think there's actual competition of Bloomberg. Bloomberg has its own sort of differentiators that are above and beyond even the data layer, but I do think we will see things that approximate that data layer or even match sort of the Bloomberg data layer that are available to retail investors because it's easier to create that using AI than it has ever been before.
Gemma Allen
>> Talk about where these conversations are happening and how that's changing, right? You mentioned 10 years ago, a lot of folks used Reddit. We had the whole crypto punk way for folks. Got a lot of great tips on X, for example, on these different platforms which aren't necessarily Bloomberg, right?
Stephen Sikes
>> Right.
Gemma Allen
>> But people made a lot of money. That world, too, is changing. And yeah, there's so much more data, but there's also potentially a risk with verification, right?
Stephen Sikes
>> Right.
Gemma Allen
>> A risk with false data or misleading data, et cetera. What are you seeing? Where do you think retail investors are going to talk these days?
Stephen Sikes
>> Yeah, it's interesting. I think there are two main sources right now. One, I do think Reddit is still the most vibrant community for investors across all of the myriad channels that are there that match every sort of style of investing you can imagine. Companies like Public, we have our own subreddit that we interact with our members. And it's one of the places I spend the most amount of time interacting with members. I think that's still an amazing and probably the best community of people. I think beyond that, you have Discord, which I think has its own connotations as being closely connected to the crypto universe. But I think we do see a lot of the trading communities and the real fast-paced conversations happening within Discords, which is interesting. And it's a place where we've invested time in getting to know the individual moderators and the individual communities and where we need to be present to meet our members. And then the third, obviously, still X, Twitter. Still a lot of phenomenal conversations. Definitely more on the sort of analytical side and slightly longer form than sort of the chat styles that you see in Discord or the more community forum style that you see in Reddit. But I think those are the three platforms where we definitely see the most conversations from retail investors.
Gemma Allen
>> Let's talk about what's happening in the market for a second.
Stephen Sikes
>> Sure.
Gemma Allen
>> I mean, we had SpaceX took a bit of a-
Stephen Sikes
>> ....
Gemma Allen
>> dip yesterday. I haven't seen it yet today, but it's... Actually, I'm looking right now. It's down. Insane story, though, right?
Stephen Sikes
>> Right.
Gemma Allen
>> When we think about, again, 10 years ago versus now, people say it's a good time to buy boring stocks. I mean, from your perspective, what are you seeing and hearing from retail investors? Where is the appetite or even the opportunity or the access? What are your thoughts? Do you think the Mag Seven is exactly the most it has always been?
Stephen Sikes
>> Yeah. I mean, listen, I try very hard not to make predictions about investments or stock prices because I'm usually wrong. Every once in a while I'll do it.
Gemma Allen
>> You and me both.
Stephen Sikes
>> Somebody will pin me down on something, and I'll have to watch that video getting reposted to my timeline every few months. No. I mean, I think what we're seeing from retail, SpaceX specifically, tremendous amount of interest in demand. I think it's unique in that we have an entrepreneur, a founder, an operator, an executive in Elon who has done such an amazing job for retail investors over the last 15, almost 20 years in sort of running and building Tesla. Tesla has always been well-owned by retail, and retail investors have been wildly rewarded for that loyalty. And I think we're seeing a lot of those same people invest heavily in SpaceX again on the same sort of thesis and on the trust that Elon will do well for them over time. It certainly is a different valuation profile from even Tesla, which is quite healthily valued at the time, so we'll see. I think my caution always to retail investors in this situation, especially with new issues and IPOs, is to be thoughtful about the float relative to the volume that may be coming on over time, right? SpaceX came out with a relatively low float. I think we see a lot of unlocks coming over the next 60, 90, 120 days. And even with the options opening, last week we started to see a little better price discovery. And so I'm always counseling people to pay attention to things like that, that it's not necessarily shown on the chart, but it's something you sort of have to understand about the market overall and how float and flows can have a major impact on price movements and waiting for sort of the float... More of the float to be in the market, I think, is always prudent.
Gemma Allen
>> Love it. Okay. So Stephen, what's ahead? Talk to us. Last question, what does the next two years look like for you guys? Are you continuing to build proprietary technology? What's the product roadmap? And what are the big bets?
Stephen Sikes
>> Totally. I mean, I think for us, again, the biggest thing is continuing to capitalize on this AI transition, right? We saw prior generations transition from calling your broker to making your trades online, from making your trades online and on the web to doing it on your phone. Now we think you're going to move from doing it on your phone to doing it with AI in the loop and having that sort of AI-intermediated experience. And so we're going to continue to invest in building more and more tools on all three of the pillars that we have, both our API, our MCP, and then our agentic brokerage that's sort of native to Public. And again, continue to build more and more features into each of them, making them higher quality, easier to work with, and, again, more observable and deterministic in their execution I think are all sort of where we're going to be spending our time.
Gemma Allen
>> Well, we certainly look forward to watching. Stephen, thank you so much for coming on theCUBE.
Stephen Sikes
>> Thanks for having me.
Gemma Allen
>> I'm Gemma Allen here at theCUBE Studio at the New York Stock Exchange. This is NYSC Wired's Mixture of Experts. Thanks for watching.
John Furrier
Gemma Allen