Mike Conover Transcript

Clint Betts

Mike, thank you so much for coming on the show. You are the CEO of Brightwave and the founder of that company. You have a long and rich history in artificial intelligence, which is the topic of the day and year and probably next decade and the decades to come.

Mike Conover

Yeah, millennium. Yes.

Clint Betts

Tell us about Brightwave and how you got there.

Mike Conover

Excellent. Yeah, of course. So Brightwave is an investment research platform that uses AI to expand the limits of human cognition. So, if you think about the job of an active asset manager, it's to know something about the market that nobody else has seen to identify a mispriced asset. Our view is that it's not clear that this is a task that is well suited to human intellectual attention span.

When you think about the body of material that is relevant to an investment decision, including SEC filings, earnings call transcripts, sell site analysis, breaking news content from the public internet, podcast transcripts, and the Congressional Record, the volume of relevant information exceeds what a person is able to keep in mind simultaneously. We've built a system that is able to connect the dots across huge bodies of content to accelerate investment research workflows.

So, a lot of that just bread and butter, fact finding, question answering, but also up levels human analytical capabilities by identifying fact patterns across many, many different documents that support and investment decisions. So finance professionals are able to accelerate workflows and make decisions with greater conviction and insight.

Clint Betts

What do you do when everybody's using Brightwave? And it's just this incredible: anybody can day trade, and they're all making the correct decision. Do you ever think about that? That's just the first thing-

Mike Conover

Sure.

Clint Betts

... that's going to take forever, but it's really interesting if everybody starts using this technology. I guess you can put different qualifiers and things like that, but what if everybody just wins all their trades? I guess that's [inaudible 00:02:05]

Mike Conover

So, this actually came up a lot when we were pitching the company originally. I think if we live in a world where Brightwave has created a perfectly efficient market, it's a very different world that we will inhabit from the one we inhabit. But I think more pragmatically, there is always going to be information that has not been digitized, and there's always going to be intelligence that these models do not have access to.

So, we've created a system that centers the analyst and the researcher in the decision-making process, and we allow a human to direct the attention of this very powerful system. But language models and AI technology are not a replacement for human judgment, and so we see that taste making process, "That's an interesting fact. I want to know more about, I want to know more about that. You should run that down for me."

I think that is where you are going to get very, very heterogeneous behaviors from different teams that have different theses, have different views on macro global. It's just still very much a taste-making process. I would make the comparison to accountants in the 70s.

So VisiCalc, the first computational spreadsheet was released in 1979, and if you were to speak with an accountant in 1978 and say, "What's your job?" They would say, "I run the numbers. I'm the numbers guy. Write it out on a piece of paper in a grid. And it's cognitively demanding. It's time intensive and it's important to the business. It is a real important job." And now nobody wants to do that work.

And it's because it's not that we no longer have a finance function; it's rather that the sophistication of the analysis that we can bring to bear on a problem is just much expanded. And I think that we're seeing a similar moment for qualitative and language-oriented reasoning.

Clint Betts

Yeah, that is super interesting. What made you see the opportunity here? Because you worked for incredible companies, you've had an incredible career in this. You have a Ph.D. from Indiana University in Complexity Science-

Mike Conover

I do.

Clint Betts

... you have a lot of background here. What made you decide, "All right, I'm going all in on Brightwave?" And just can you just go a little bit deeper on the opportunity that exists there in the investing space?

Mike Conover

Yeah, thanks for a thoughtful question. I have been fortunate to work in environments where I've had access to data sets that describe the true structure of society. So, my Ph.D. dealt with the structure of how information spreads through very large groups of people on Twitter.

We had very early access to the Twitter data set, and I worked at LinkedIn, where we had this high-resolution graph, this network structure describing the global economy, who works, what skills they have, and where they are located. At Workday and SkipFlagma, my previous startup, I just have had a front-row seat to data sets that describe the systems that govern the course of our lives and have the privilege of seeing the statistical reality of...

You think about traffic, the economy, or pandemics and how these processes unfold. Climate is a good example. It is something that is larger than ourselves, but that has a material impact on our lives and that has a real quantitative mathematical structure, but we can't see it. And it's only in the past 25 years that computational technologies have made those systems studyable even to us and with language models...

So prior to founding Brightwave, I was at Databricks where I created one of the first language models to demonstrate that open source technologies could exhibit the kinds of behaviors we've come to expect from ChatGPT. And that work in language model engineering really helped me understand how... You look at the glide path for this technology, it is expanding in its ability to process large volumes of information simultaneously.

And you connect that fact with regards to speed and volume of content processing to the reality that there are systems that govern our lives that we can't see, and the market's being the ultimate manifestation of those systems. And it seems like a very natural decision to build a technology that accelerates that process.

Clint Betts

I'm fascinated by the debate between open source and closed source and AI and you got Zuckerberg, on the one hand, who is all in on open source, which is really fascinating. That seems like a kind of a new thing for Zuck. And then you've got Sam Altman who runs OpenAI, but it's not actually open, it's closed AI.

Mike Conover

Sure.

Clint Betts

What is your take on the difference between these different paths in AI, and which one do you think eventually wins out? Or, I mean, in software, they both exist, obviously. So, maybe they just [inaudible 00:07:27]

Mike Conover

I think that what Facebook is doing, I think that axiom is commoditizing your compliment. And so many of the things that they do when you think about newsfeed, content generation, or chat, this language being the substrate that Facebook operates on, in addition to images, all of that is better and more usable with something like language models. And so, being the de facto standard for open-source language models and having a robust developer ecosystem around their technology makes their business run more effectively and efficiently. Like you think of digital avatars as they move into BR and AR to BR and AR language models. So, I do not think it is pure generosity of spirit that leads to this decision, but I think it contrasts with some of the commercial offerings or these closed source models, not for nothing, running a 400 billion parameter model, like Llama 3 45B, is, which is their largest and most powerful model from Meta. There's a fair amount of engineering overhead associated with that.

And there's some comparability among the current frontier models, whether it's from Anthropic or OpenAI or Meta. But goodness, it is convenient not to have to run it all yourself. And so I think it's really about more just classical trade-off management. And so I think it's just choose the right tool for the job, is the approach that we take and also recommend that our customers take.

Clint Betts

And again, this is probably the question of the millennia: What is the future of AI? How do you see this playing out as you think about over the next five to 10 years? I mean, what is this thing going to be capable of? And I know that's hard to answer, but just give us your broad strokes of that.

Mike Conover

I think I can say that in the next several months and years, we're going to see an increase in what would be described as agentic tool use. So right now, these models, it's out of the box, you get a ask a question, get a response. But there's not a lot of planning work that goes into that. So, if you ask it to achieve a complex goal, it can often get tripped up.

So at Brightwave, what we're building towards are systems that unfold a sequence of decisions like an analyst would make in pursuing an investment thesis that allow us to control the direction of how the model makes decisions, so it doesn't get caught in factual misstatements or running down dead ends, or if it runs down a dead end, it bails on it as quickly as possible.

And I think that right now the real product form factor has been chat. That is what most people think of these systems is doing, and we're seeing a move towards more asynchronous, longer-running, more intensive workloads, whether that's, "Hey, I'm going to give you seven to nine minutes to write a fairly sophisticated piece of software and with complex structure," for example.

Or, "I want you to take time to assess, critique, and refine this thesis, and then when you have considered all of the relevant factors, report back to me with a comprehensive brief on the topic." Rather than having me at every turn pushing you to say, "Answer this question, answer this question, answer this question," and that agentic behavior, where that has access to tools like market intelligence, like database lookups, like calculation capabilities. I think that's the future of this technology in the six to 12-month time period. What's going to happen in five years is...

Clint Betts

I know.

Mike Conover

I wouldn't even know to speculate.

Clint Betts

We may all just be unemployed and serving the robots at that point, which [inaudible 00:12:00]

Mike Conover

That is actually where I disagree. I think that the parallel to... Think about the amount... Here's a concrete example. We're going to a big conference, we're looking at the attendee list.

Historically, you would've had to sort that attendee list into the investing strategies of all of these individuals who are attending. You would've done that by hand, and there was no alternative to a person doing it. Now, you can have a system do that, and it would've taken three hours, and now it takes a matter of minutes. I think there's a fair amount of work in our professional lives that we don't notice how tedious it is because it's all we've ever known.

I like to think that as these technologies get more sophisticated, we will be freed to focus on the things that are enjoyable and actually value and require the most of what we have to offer, rather than the tedium similar to those accountants in the 70s.

Clint Betts

Yeah. Yeah, I definitely think that's right. This is a question where we began, but going a little bit deeper, what is the future of trading platforms in the age of AI? What does a Charles Schwab, for example, think Brightwave is doing?

Mike Conover

Well, I wouldn't be able to comment on any individual platform. But I think that what you can say is that retail trading is above 30% of the total transaction volume as of last year. And as with all things, there's this democratization of technology, where you used to book a travel agent, and now everybody can plan incredible trips.

And historically, it used to have a broker who you would call to place a trade, and the first wave was now we have democratized access to low cost trade execution. I think in the individuals having excellent access to the kinds of information that historically has been the exclusive purview of extremely well capitalized firms, and not just information but technology to enhance decision-making where if you believe...

If you have a thing that you believe about the world, whether it's the future of drones or if it's how Africa is on the rise or the political situation in South America, and you just want to articulate that by putting your money where your mouth is, I think that individuals will be able to do that in a way that goes far beyond just tilting at windmills.

Clint Betts

What do you make of the rise of retail trading? Obviously, one of the reasons why it's happening, as you just explained, is the advances in technology, and that's only going to get better in the amount of resources and knowledge and information that retail investors are only going to get greater, and that gap between retail versus more accredited investors is only going to get smaller and smaller.

But is there anything else to it? The GameStop thing doesn't seem like it was that [inaudible 00:15:21], various things, was there more to why this might be happening?

Mike Conover

I'm a fairly uninteresting investor. I bet on the progress of technology very secularly, and so I just want to represent faithfully my areas of expertise. Candidly, I think part of it is probably a sense of disillusionment with the American bargain, at least domestically, and a feeling that, like, geez, I've got to be trying something. I want a piece of the action. And I think that is a factor here. I think you saw that in crypto as well, where people were just trying something and trying to get ahead through any means possible. Now, is this the right way to manifest that urge or instinct? That's an open question. I would, again, be above my pay grade. So, I think there's the democratization of the tools, but I also think there's this sense that it is harder to get ahead in American society now than it was when I was a kid. And I think people are responding to that, and whether this is the correct response to that fact is unclear, but I would say that's probably a co-factor. What do you think?

Clint Betts

As you see the distrust in all institutions go down seem, it makes a lot of sense that the rebellion would come up at that same time, and so I think that's a good piece of it. And I think the disillusionment is a great way to describe what's happening there. I want to ask you this as CEO of Brightwave, what does a typical day look like for you?

Mike Conover

Goodness gracious, it is all over the map. We move very, very quickly. We are extremely responsive to feedback from our users. So, it's a mix between product design, just reviewing user interface. I mean, the design problem we're solving is how do you reveal the thought process of something that's considered 10,000 pages of text to a person in a way that's useful and understandable?

It's not a design problem anybody has ever solved before. So, we are at the forefront of user experience and user interaction design. It is not a subject in which I have classical training, but likewise, marketing and sales. I mean, we are constantly thinking about how to cut through and deliver our message in a...

People are so saturated. Goodness, I get dozens of AI-generated sales outreach emails in a week, where clearly a model has read something from our website and generated sales collateral that, and how do you communicate your message in that environment? I think there are problems that we think about.

So, there's a lot of commercial activity, there's a lot of product design activity, but the bread and butter of the work is just fundamental systems engineering and machine learning engineering, where we are creating a system that solves a very very complex problem in a way that has never before been possible. One of the advantages that Brightwave has relative to some of our competitive sets is that it's founded and staffed by extremely accomplished technical minds.

How do you direct in a reliable and repeatable way the decision-making process of this agentic research assistant, is effectively an open research problem. And so a lot of the work is looking at the numbers, understanding the limits of the technology and designing algorithms that allow us to accomplish our goals.

Clint Betts

What do you read? What reading recommendations would you have for us?

Mike Conover

Do you mean technical nonfiction?

Clint Betts

Just general-

Mike Conover

[inaudible 00:19:39]

Clint Betts

In general, what do you read?

Mike Conover

That's a great question. I would say that, I'll give you a personal answer. My inclination is always to read non-technical work, or rather, is always to read non-technical nonfiction. And I have to overcome that urge because you can't be running at 120% all the time. And so lately, I've been reading the Expanse scifi series.

Clint Betts

That's great.

Mike Conover

Actually, when we were doing the work with Dall-E, which was the language model we created at Databricks, I reread Neuromancer and Snow Crash, which are sort of both classics of the cyberpunk AI gone wild genre and things like this.

Clint Betts

I like that. I wonder, I've been thinking about this a lot lately and I wonder what your take would be on it as a CEO and a founder and somebody who's building a company. I live in Utah. We have this really strong tech and [inaudible 00:20:51] community here.

Mike Conover

It's a beautiful state; I'm from Arizona.

Clint Betts

Thank you. We got to get you up here for Silicon Slopes Summit in January.

Mike Conover

Love it.

Clint Betts

So I'm here, and we really have this mental health crisis in our community that I'm seeing. It's not unique to Utah, where entrepreneurship is so lonely, but being a CEO can be lonely at times. I am just wondering how you work through that. Do you have any tips or tricks, or things that you do to stay balanced?

Mike Conover

For loneliness in particular?

Clint Betts

That, and just general mental health? It seems like it's degrading a little bit, and that's not a unique thing to entrepreneurship, or it just [inaudible 00:21:30]

Mike Conover

Sure. So, that's a great question. I really think that the work that you do to understand where you stand, if you have a foundation that you believe in, makes all other decisions much clearer. And for me, it's relationships, natural beauty, and personal growth. So, constantly reflecting on how I am investing my time and how it connects to the things that I want to achieve in my life in a wraparound sense helps me feel like the decisions I'm making are lining up with the larger objectives.

I would say, additionally, that my family is a great source of strength and the reservoir from which I draw my sense of purpose in the world candidly. And so I think the last thing I'll say about this is that relationships are that first pillar. I have a community of men in my life where I am candid and vulnerable and friends with whom I can really break it down and who understand. I think all of these circles around the natural world and human connection are so different.

And I think that the tonic that is the salve for these modern times is not forgetting the archetypal rewards that are baked into us at a neurological level. And no phone is going to replace that circuitry.

Clint Betts

Do you think as AI accelerates, and we can do more and more on our own and all, don't you think human connection becomes even more valuable, not less?

Mike Conover

Yes.

Clint Betts

I see that as becoming the gatherings and this... Because we're very social animals as human beings. Don't you see human connection as being like that's going to be even more valuable in the future?

Mike Conover

Absolutely. And when I think about the unlock when it comes to our professional lives, where are we connecting? How are we connecting? How are we investing our time so that it's not the rote tedium that is automatable but instead interpersonal work, and business deals will still always be relationship-driven?

And I think those are the extent to which people have more time to spend with their family and friends, to the extent that they're able to focus on the professional relationships that matter. I think technology is unlocking that dimension.

Clint Betts

What do you think about creating, defining, and building a culture inside of your company?

Mike Conover

So we have three core values at Brightwave, ownership, humility, and craft, and I can speak to each of them in turn. So ownership is this idea that, have you ever read Man's Search for Meaning by Viktor Frankl?

Clint Betts

Yes, yes.

Mike Conover

Yeah. So, Holocaust survivor reflecting on what is it that allowed some people to endure the horrors of the Holocaust, and the lesson that I took away from that book is effectively that you can't control your circumstances, but it's inalienable to you to choose how you respond to your environment. And that idea of being a high-agency human who does not make excuses but rather says, "I take responsibility for how things unfolded," is a very, very powerful way to be in the world.

Humility is borne out of this scientific method. But also, just as a principle, there's so little that we actually know about the world, what other people think, and what's actually going on. And it's a real source of strength to recognize that and just accept that you are just this moat of dust floating through space.

There's a lyric from a rap group I really like, Run the Jewels, which is roughly, "You can't crush me on dirt." And it's just a real powerful, "What are you going to do?" If I'm wrong and you and I disagree about something, it's probably because you have the information I need to get.

And so my job should be to elicit what you are seeing that I'm missing. Because I believe you're a reasonable person, as a default. And that often leads to better outcomes because you're more interested in doing the right thing rather than being right. And so that final piece is just craftsmanship, which is the pride of work that accompanies meticulousness and just really focusing on the details.

Technically, this is extremely important in the sense that if you have a, you have a mechanical system that has a joint with a little bit of play in it, you may not notice that when there's only one joint. But if there are eight connections in serial now, you suddenly have a lot of play in that system. And with language models, there are so many places that the system can exhibit a little bit of play.

So, being able to rigorously and methodically execute a technical agenda and building things right the first time, even if it takes a little longer, allows us to move faster in the limit. And so ownership and humility and craft are three of the core values for our teams.

Clint Betts

Ownership is a great one. I don't know, but I really hear that one very much as a value. I love that. I think that's kind of fascinating. How do you stay motivated, just motivated to keep, I mean, that actually [inaudible 00:27:39] question. How do you stay motivated?

Mike Conover

So a couple things, these are great questions by the way. I appreciate it. So I am a very relationally driven person. I think that the sense that I am coming into a workplace... We work in person five days a week. I'm coming into a workplace surrounded by people who are intense, who are extremely capable.

We've got folks from Meta and Instagram and UBS and McKinsey and Wharton. It's just a stacked bench where it's like I have to earn the respect and trust and time of the people around me all the time. And I find that extremely energizing to just be surrounded by people who cause me to achieve more. And additionally, the problem we're solving is just intrinsically interesting.

It's no less than understanding the true structure of the global economy with artificial intelligence. Boy, howdy. If that's not enough to keep you busy, I don't know what it is. So I just really enjoyed the scope of the problem that we were solving, and for whatever reason, I found the day-to-day and technical details, as well as the commercial details, very, very interesting.

Clint Betts

What made you decide to go all in person five days a week versus a lot of people doing the hybrid versus the work from home? What went into that decision?

Mike Conover

So, it's actually our stated policy is three days a week. It's just the pleasure of being around great people. So it was not a decision, it's just this has happened organically. I mean, listen, when you're at a big company and whether you're there or not, doesn't matter that much. People notice that whether or not they can articulate it, they feel it. So I think a lot of the, why am I commuting? There's a direct through line to, "If I wasn't here, it wouldn't really matter that much." But at a startup, it matters. And we have offices in New York and Boulder and we have a team in California. We'll open an office in California.

So there's a remote culture, but we are... I don't know what the model for it would be because we're not remote. It is centered around these hubs, and it's probably just going to be those three sites. It's not going to be 12 different locations. So, being together in person, we must understand that we can work remotely and do work remotely. I think it's a new way of working.

Clint Betts

Two more questions for you.

Mike Conover

Sure.

Clint Betts

How are you thinking about 2025? Obviously, we have an election coming up here in the United States very soon, which is an interesting thing because CEOs nowadays are being asked to comment on things that have nothing to do with their companies or what they're focused on. I'm not asking you to comment or don't tell us who you're voting for or anything.

Mike Conover

Sure.

Clint Betts

But how are you thinking about how that shakes out and going into 2025?

Mike Conover

Capital efficiency is just having enough cash on hand to weather whatever may come, whatever regulatory regime, and whatever geopolitical uncertainties. I think I wouldn't call it a fortress balance sheet, but you want to have enough money to... Money is fundamentally optional, so being extremely judicious about how you spend and preserving optionality in a period of change, I think, is always a good policy.

Clint Betts

Finally, we end every interview with the same question, and that is at CEO.com, we believe the chances one gives is just as important as the chances one takes. When you hear that, who gave you a chance to get you to where you are today?

Mike Conover

Who gave me a chance? That's a great question. There are so many people. I think that I would say... I paused for a moment there because so many people came to mind. I would say that the... I will thank our investors. This is not a twee answer, either. Their slogan is, "Believe in technical founders and invest early." And I can't believe I'm thanking our investors and recognizing them, but I actually feel this way. They are people who make early bets on teams, just on the basis of conviction in the absence of other data, really.

We set out to raise capital and we met them, and it was just, we hit it off immediately, and they just got long on what we were doing, very assertively and very confidently, and I think that is... It just created an opening for us to do so, and we built something. We knew what was going to happen, but having investors who share the enthusiasm and the vision and just believe, I think, has created the opportunity for all of this to exist.

Clint Betts

Who are your investors? If you want to shout them out, just so other entrepreneurs know who they are.

Mike Conover

Decibel, Moonfire, and one of the largest asset managers in the world.

Clint Betts

Mike, thank you so much for coming on. Seriously, congratulations on everything you're doing. Man, what an opportunity and a category you're going after. So, best of luck with everything.

Mike Conover

I appreciate it, Clint. This was truly a pleasure.

Edited for readability.

Podcast Newsletter

Every Episode

Never miss an interview or episode of The CEO.com Podcast.