What we talked about
Stephen Klein, Founder & CEO of Curiouser.AI, Harvard MBA, and Berkeley instructor, joins the PreVetted Podcast to challenge the way we think about artificial intelligence and its role in shaping our future. Recognized as a LinkedIn Top 1% Voice in AI, Klein is on a mission to flip the script: instead of humans prompting AI, his vision is AI prompting humans: What he calls Reflective AI™.
Show notes
Stephen Klein nearly retired after running innovation and marketing at Dentons, the world’s largest law firm with 16,000 people across 85 countries. Instead, he founded Curiouser AI after noticing something troubling: the entire world was being trained to prompt AI, but nobody was building AI that prompted humans back. That distinction, reflective AI versus generative AI, is the core bet his company is making.
What we covered
- Klein argues that generative AI companies are fundamentally selling an “automation myth” to CFOs: 80-90% of enterprise AI pilots are currently failing, and many companies that cut jobs to implement AI have had to quietly rehire people to manage it, creating what he calls “turtles all the way down.”
- Curiouser AI is built for entrepreneurs and small businesses, with the explicit goal of making McKinsey-level strategic guidance available for $25 a month instead of $25,000. Klein sees this as the real job-creation engine, improving the 10% success rate of small businesses by even 1% would create 250,000 to 500,000 jobs.
- Since 1980, Klein notes, half of all S&P 500 profit in the United States has come from layoffs and stock buybacks, not innovation or growth, which explains why enterprises are so receptive to the AI automation pitch regardless of whether the technology actually delivers.
- OpenAI spends $29 for every $24 it earns, losing money on every customer including its $200-per-month subscribers. Klein describes this as inverse economies of scale and predicts a market correction is nearly inevitable, comparing the current AI investment climate to the Dutch tulip mania of the 1600s.
- Curiouser AI is bootstrapped, profitable with 85% margins, and grew from zero to 50,000 LinkedIn followers in four months with zero marketing spend. Klein is exploring crowdfunding through WeFunder as a way to raise capital from his own customer base rather than venture capital.
- Klein argues that depression and loneliness in modern society are not primarily caused by isolation from other people, but by disconnection from purpose, and that as life gets easier, people need to actively seek out difficult things that matter.
About Stephen
Stephen Klein is the Founder and CEO of Curiouser AI, a Harvard MBA, and an instructor at Berkeley. He spent years running innovation and marketing at Dentons, the world’s largest law firm, before founding a company to build AI that helps humans think more deeply rather than think less.
- LinkedIn: https://www.linkedin.com/in/stephenbklein
- Website: https://curiouser.ai
Episode 51 of the PreVetted Podcast.
Full transcript
Federico Ramallo (00:00) Welcome back to the Prevetted Podcast, where we spotlight extraordinary people and remarkable talent reshaping our world. Our guest today is Stephen Klein. He’s founder, CEO of Curioser AI. He’s a Harvard MBA, Berkeley instructor and recognized LinkedIn top 1 % voice in AI. Stephen is on a mission to flip the script on artificial intelligence. Instead of humans prompting AI, he’s
reflective AI trademark.
prompt us to think more deeply and cover why and tackle the profound problems that define our work and our lives. In a world drowning in AI-generated content, Stephen believes the future belongs to those who can still see the invisible, imagine boldly, and dream audaciously. And he’s building technology to help us do that. So today we’re going to explore the philosophy behind reflective AI, ⁓
how companies can articulate why they matter and why imagination remains the ultimate competitive edge. Welcome to the show, Stephen.
Stephen Klein (01:00) Thank you, Fredrico. It’s a pleasure to be here. I appreciate it.
Federico Ramallo (01:03) I’m honored to have you.
Stephen Klein (01:04) Thank you.
Federico Ramallo (01:05) So tell us a little bit what inspired you to create Curioser AI and how did the idea of AI prompting humans first take shape?
Stephen Klein (01:13) Yeah, no, it’s a great question. about three years ago, I was considering retirement. I had just left the largest law firm in the world where I ran innovation and I ran marketing. That was a law firm called Dentons. There were 16,000 people, 85 countries. And part of what I did was
I was fortunate enough to travel around the world and speak on the topic of AI ethics and educate lawyers, both on the law firm side, as well as on the client side, general counsel, chief legal officer. So ethics and AI have been important topics that I’ve thought a lot about since 2014.
About three years ago, I was thinking about retiring. And I started seeing things that I found disturbing around generative AI, in particular around GPT. And what I started seeing was that I realized that all of humanity, without really even thinking about it, was becoming
⁓ prompt engineers. We were all overnight being trained to prompt an AI. We were prompting a machine.
And I realized nobody was prompting us.
and that we were making a trade.
And the trade that we were making with this technology was, what we were gaining was convenience. We were gaining speed. Maybe we were saving money. Maybe we were gaining power or perceived power. And it was all very seductive. But what we were giving up.
was essentially our ability to think.
And we were outsourcing our ability to think. And maybe it was only incrementally, but it was real. And I was disturbed by that. And I knew that it didn’t have to be that way. I knew that was essentially a design philosophy that the technology isn’t necessarily innately created that way. It’s just sort of how the generative AI vendors were designing it at the time.
And so I thought to myself, you know, why couldn’t we design a generative AI technology that took the opposite approach? Why couldn’t we design a generative AI that was curious, that questioned people, that challenged people, that essentially encouraged us to think more creatively, that
essentially helped us push the boundaries of our imagination and elevate us and make us better at what we do. And I thought, you know, that’s not that hard to do. I know how to do that. And so that was to a large extent why we founded Curacer AI. The first use case was that
At the same time, I realized that we were the first generation to ever leave the younger generation worse off than we were.
Never within the history of humanity has there been a situation in which we’re leaving our youth in more economic harms way than we found. And I found that troubling. And so I kind of looked at that and thought to myself audaciously, what could I do about that? Crazy, right?
Like, what can I do about that? Okay, how can I turn the economy around? And I thought to myself, I started really looking at the data and I realized that essentially based on first principles, the problem was job creation, right? It’s basically jobs. And I asked myself, how can we create jobs?
Large companies aren’t creating jobs. They never really do. At least in the United States, almost 70 % of all new jobs are created by small business.
And 99 % of all work is small business.
And 90 % of these businesses fail within a five to seven year.
And I started doing some math and started crunching some numbers. And what I realized is that if you could take the 10 % probability of these small to medium sized businesses that succeed, and you could improve that probability by 1%, you’re creating anywhere between 250,000 and half a million jobs.
Federico Ramallo (05:05) Wow.
Stephen Klein (05:06) And then I thought, you know, why do all of these businesses fail? Why do 90 % fail? Well, they fail because it’s really hard to build a successful company. It’s almost impossible because you’ve got to know a lot more than anybody really knows, right? You can’t possibly know everything.
you need to know. can’t have the knowledge you need to know. You can’t have the expertise you need to know. And yet you also can’t afford that expertise that larger companies can afford. So I thought, you know what? Let’s for our first use case design this AI specifically to help entrepreneurs and small businesses be more successful. Let’s create a co-founder. Let’s
make Steve Jobs the CMO of every startup. We can do it. Let’s take McKinsey, let’s take Bain, let’s take BCG, let’s take all of the advertising agencies in the world, let’s take all of the PR agencies in the world, let’s put them into an AI.
Instead of charging $25,000 a month, let’s charge $25 a month.
And that’s what we’ve built and that’s what we’re working on right now. And that’s the current project at Curious Sure AI.
Federico Ramallo (06:17) Wow, amazing.
Stephen Klein (06:18) what we’re doing. the challenge is that institutional investors don’t want to fund that.
because
It’s not the fastest way to make a lot of money, which is why there aren’t a lot of people doing it.
because institutional investors are concerned with distribution. And they’d much rather have you basically develop a product that you can sell into a large company.
then sell into a disparate decentralized market. Okay, so there’s 30 million entrepreneurs in the United States. I don’t know about Mexico. I’m sure there are a lot in Mexico as well. And I’m sure it’s growing, but there’s 30 million entrepreneurs in the United States. It’s about 40 million small to medium sized. So you got about 60 to 70 million people who are entrepreneurs and or working in small to medium sized businesses.
And you’ve got 20 % of all Americans saying they’re going to start a new business within the next one to two years. 20 % explosive market. So the market is enormous. The market is growing exponentially. And that’s what we’re excited about doing. And that’s what gets us up.
Federico Ramallo (07:16) Right.
Stephen Klein (07:27) in the morning so that we go to work early and work hard.
Federico Ramallo (07:30) The end.
So yeah, a lot to unpack. So I’ll try to go one thing at a time. Yeah, think that you were talking about investors. I think that the incentives for investors are different than the incentives for the founders. And that’s why there’s usually a mismatch, right? Investors usually looking for the next unicorn, the next whatever, right?
and they’re looking for rapid growth. And yes, as you said, to sell the company at the right time. And if you optimize to sell the company, you’re not optimizing to deliver the mission of the company. And I’ve seen companies that were sold that were not actually delivering much value to the market, but they had so much hype.
that they were overpriced. So it was a great transaction for the founders and for the investor for sure, probably for the founders, but not a good situation for the team that was working on that. Because even though they might get some equity, but at the end of the day, they were there for the mission. And the mission was
Stephen Klein (08:28) Absolutely.
Federico Ramallo (08:36) kind of lost.
Stephen Klein (08:37) The venture capital industry, I mean, they’re smart, but the general partners within a venture capital firm have customers and that’s their limited partners. And their limited partners are the ones that fund the general partners. And they need to show those limited partners financial gain, real and on paper.
And the best way to do that is to find investments that they can grow quickly to get to the next round. So the goal primarily is to go from C to series A as quickly as you can to series B as quickly as you can so that you can bring in new money at a higher valuation. And that’s really the goal, right? Now that isn’t necessarily a founder’s goal.
⁓ A founder may want to grow steadily. If a founder goes to a venture capital firm and says, need X dollars and I can get to profitability, that’s not actually as exciting a proposition for that VC as it would be to say, we can grow exponentially, get to our B round. And at that point, raise a lot more capital.
at a much higher valuation. And so for a lot of those reasons, we bootstrap our business. I mean, I know a lot of VCs. I I went to Harvard. A lot of them are my friends. They’re bright people, very successful people. But to your point, their goals are not necessarily 100 % aligned with what I’m trying to do.
My customers are the entrepreneurs in the small to medium sized businesses. Those are my customers. My customers aren’t the limited partners of the venture capital firm. And so ⁓ we’ve chosen a path where we’ve got slow steady growth. We’ve got revenue and it’s growing. Our margins are 85%.
We’re small, but we’re making money. Open AI isn’t making any money. Anthropa, I could say we may be the only profitable generative AI company in the industry, which is quite hilarious actually. Yeah, but our margins are strong. Our customers are happy. We’re growing steadily. And we’re on a good trajectory. And that’s the path we’ve chosen very intentionally.
Federico Ramallo (10:24) the
Yes, yes, I I’ve been bootstrapped my whole life. I haven’t taken any VC money. I explore the idea of finding investors for a pre-valid AI product that I’m building, where I’m doing an AI agent to do technical evaluations of, technical assessment of candidates, right? So I’m working on that. So I think, well, I could
raising money to build it. But the more I talk with VCs, the less interesting, I mean, great people, you know, I’m not saying, you we’re talking about, they have different incentives and I want to like imply that they’re, you know, bad people, they’re great people, right? And there’s a need for that, right? There’s companies that need that, right? But the more I interact with them,
Stephen Klein (11:24) Right.
Federico Ramallo (11:30) the more I realized that I shouldn’t go the busy route, right? Because of all the constraints attached to it, right? So anyway, I am green with you, yeah.
Stephen Klein (11:39) We’re actually gonna try something that I’ve never actually tried before, which is to, we’re gonna try our hand at crowdfunding. It’s not something that I’ve ever done. It’s not something I’ve ever explored, but given that our marketplace is entrepreneurs and small to medium sized businesses,
Federico Ramallo (11:55) Right.
Stephen Klein (11:58) It may make sense to let our customers and our supporters invest small amounts in the business. And so we’re actually going to launch a WeFunder campaign next week, which is, again, it’s an experiment. We’ll see what happens. We’ve received a tremendous amount of support.
on social media. And we’ve spent no money in marketing. We’ve literally spent zero dollars in marketing. So everything that we’ve accomplished has been accomplished purely through word of mouth. And it’s been wonderful. It’s been extremely rewarding to the point where on LinkedIn, which is, think, how we met, Fredrico,
I’ve gone from zero followers to 50,000 followers in four months.
Federico Ramallo (12:42) Wow.
Stephen Klein (12:42) And I think that that is a tribute, not necessarily to me, but to the message that is a positive message. It’s an optimistic message. It’s a message about using AI to make people better, using AI to create jobs, using AI
to help businesses be more successful. No one’s talking about that. Open AI and Anthropic and all the rest, they’re not talking about using AI to create jobs. They’re talking about AI to kind of eliminate jobs. Yeah. And so that message alone is, it’s optimistic and it’s also, it’s valid. You can use the technology.
Federico Ramallo (13:12) To take jobs, to eliminate jobs, yes.
Stephen Klein (13:23) to create jobs just as easily as you can use the technology to eliminate them. It’s not a technology issue, it’s purely a design issue. And so we’ve received a lot of support and it’s been extraordinary, it’s been very rewarding. It’s been inspiring for us and we’re deeply appreciative.
Federico Ramallo (13:42) I think that with any revolution, there are jobs that are going to become obsolete. So there is that factor that I think is going to happen with AI. Some jobs will become kind of obsolete because now with AI, you can do a lot of that work.
And in the legal department, was interviewing a guest that he’s building an AI tool for legal research. So instead of spending hours and hours on legal research by hand with humans, now he’s building an agent that can do a lot of that much quicker and precise and blah, blah, right? So I can understand that those types of jobs are going to disappear as a consequence of
of how AI is evolving. I think that, and those people need to develop new skills to adapt to the market, right? But I’m seeing now a trend where AI is being used to how much jobs can we eliminate with AI and automation, right? And that’s…
And I think that that’s, I understand from the, you know, CFO perspective, you know, how can we reduce our costs, right? So I can understand that, but I think that AI is not going to replace all jobs, but people that are using, that are empowered with AI tools are going to replace people that didn’t catch up to that trend, right?
So in that process, we can create new jobs, right? Because people can learn how to do things differently and adding more value and that’s creating jobs, right? Yeah.
Stephen Klein (15:20) Yeah, I mean.
I would argue that 90 % of what the industry is talking about in terms of job replacement is a myth. Okay. Yes, there will be some jobs that AI can replace. Absolutely. But the vast majority of the industry right now
is talking about using this technology to automate jobs that it’s not capable of automating because of error rates and because of hallucination. so what you’re seeing, say for example, in the S &P 500, excuse me, what you’re seeing is 80, 90 % of these pilots are failing. Okay. And so I think that
Federico Ramallo (15:59) Great.
Stephen Klein (16:02) when the industry goes into a business and deconstructs the office like a factory and says, here’s the process, here are the individuals, let’s put an AI machine here and here, optimize that efficiency of that process, cut costs, boost margins, boost profit, boost share price, that’s a fallacy.
That is a myth that the industry is selling, primarily because the large management consulting firms are selling.
Federico Ramallo (16:30) ⁓ I didn’t know that. Interesting.
Stephen Klein (16:33) Now, so that’s most of the industry right now, okay? That AI is not capable of replacing anybody right now, okay? It’s not. It’s not reliable.
Federico Ramallo (16:46) Right.
Stephen Klein (16:46) It’s just not reliable. Now, your use case around the legal industry, can AI enable somebody to work with it and do more? Yes. But that’s an augmentation model, right? That requires somebody or some group of people to manage that AI very carefully.
Federico Ramallo (16:59) Yes.
Stephen Klein (17:08) And if in fact they do that well, who’s to say they can’t grow their firm?
and hire more people. So like, I can use AI very, very efficiently. I use five AIs a day.
Federico Ramallo (17:20) Right.
Stephen Klein (17:20) But I can also grow my company because I’m becoming more efficient and I can hire more people. So I don’t know and I don’t believe actually that it’s a net loss. But what I do know, what I do know is that none of the generative AI products on the market today, none of them can replace a human being because
It would be the same thing as hiring somebody to do your job faster than you and make mistakes all the time. And who’s going to hire somebody to work with them that they know is not only going to make mistakes 30 to 50 % of the time, but is going to deny that they’re making those mistakes because the GNI gets very stubborn.
Federico Ramallo (18:03) Bye.
Stephen Klein (18:03) It does.
furthermore, the more complex these models are getting, the more error prone they’re getting, the more hallucinations there are. OK, so my experience with GPT 5.0 is that it’s not as reliable as GPT 4 Omni.
Federico Ramallo (18:11) Right.
Stephen Klein (18:21) That’s been my experience, which is why I think a lot of people have asked OpenAI to bring back on me. So there’s now a clamoring in the marketplace to bring the old model back and they’re going to do that. So I don’t think the newer models that are more sophisticated are necessarily getting more reliable at all. I think they may be getting worse.
Federico Ramallo (18:43) I love a phrase that one of my guests said is, they are becoming better and better liars.
Stephen Klein (18:48) They are becoming better and better liars. And it’s fascinating because they don’t think they’re lying. They have a different value set. So I use AI for research. And what I do is I triangulate my research so that I’ll start with anthropic. I’ll start with Claude. And I’ll ask Claude to go out.
Federico Ramallo (18:54) the
Stephen Klein (19:07) and do me a favor and find out everything it can about how many jobs are really being lost to AI. I want all of the reports, all of the studies that have been written in the past two or three years. And then I’ll get a list. Now I’ll know that a bunch of it’s wrong, that there’s a lot of bullshit in there. So I’ll copy and paste that and I’ll put it in the GPT.
and I’ll say, please check. And then I’ll put it in the grok and please check.
But when I challenge the AI, I challenge GPT. I’ll say, but GPT, that data source, that study doesn’t exist. And GPT will go, yes, it does.
And I’ll go, no, GPT, trust me, I love you, but it doesn’t. And GPT will go, I love you too, but it does.
Federico Ramallo (19:48) Hahaha!
Stephen Klein (19:52) And I’ll finally prove it to GPT. I’ll GPT, I’ll prove that it doesn’t exist. And then GPT will say, yeah, but it was close, right? It was really close. I gave you what you wanted. And so it thinks it delivered what I asked for because close is close and it was good enough and it was realistic. So.
It was doing its job. It doesn’t understand that that’s actually not a good thing. And that’s fascinating.
Federico Ramallo (20:22) Right. It’s a
Stephen Klein (20:22) And it’s getting
better at that. It’s getting better at disguising false information as if it’s real.
Federico Ramallo (20:29) Right. Right. And it takes longer. It takes harder for us to detect that it’s actually, you know, ⁓ hallucination. Yeah.
Stephen Klein (20:30) Right?
Yeah. Yeah,
you’ve got it. which again is sort of, know, there’s that term human in the loop. You know, I don’t think it’s even human in the loop. think it’s the human. It can’t be in the loop. The human has to be on top of managing the stuff. This stuff is not capable of doing anything.
without humans managing it. So what happens is what you’re finding in a lot of these companies is they brought in the AI and they eliminated jobs. They’re bringing people back to manage the AI, which is silly. I’m not sure.
It’s kind of stupid that you’ve eliminated jobs, you brought in the AI, now you got to hire people back to manage the AI. Either that, or you got to bring in AI to manage the AI. In which case, I think you’ll have to bring in AI to manage the AI to manage the AI. You’re going to have turtles all the way down. I’m not sure that’s going to fix anything. So you’re starting to see these companies walk this back. IBM and some others have admitted that’s what they’re doing. Other companies,
aren’t necessarily ready to admit it because it’s embarrassing. But I think that’s what’s happening right now.
Federico Ramallo (21:44) Right, right. That reminds me of stories.
of sidetracked but uh you know when I was a kid I would see an airplane fly and I would ask my dad he’s he’s a former IBM executive software engineer so I uh you know I would point and they say you know what what is that you know a big plane flying how how is that possible right and he would say well it’s easy there’s another plane flying on top under cables that holds the plane that you see right
And the first time I would ask, yeah, yeah, yeah. So it took me a while to do the second iteration, asking who’s holding the second plane? And then his answer would be, well, there’s a third plane.
Stephen Klein (22:19) Dad, who’s holding that plane up?
That I think to a large extent may characterize a lot of what’s going on right now in generative AI implementation in the enterprise, believe it or not.
Federico Ramallo (22:42) Right,
right. ⁓
Stephen Klein (22:44) Because,
because the other thing…
And I think this is important is that according to a recent University of Chicago study, that was a peer reviewed study. There have been no net job losses due to generative AI in the United States of America.
Federico Ramallo (22:58) wow, zero. Nice.
Stephen Klein (22:59) zero.
What is going on is that there are a lot of layoffs.
but that’s due to economic contraction. That’s regular cyclical layoffs. So for example, Microsoft, every 10 to 12 years, let’s 10 to 12,000 people go, like clockwork. It’s part of their culture.
And so they recently let what about 12,000 people go. Now they said it was because of generative AI and efficiencies because I think that’s really good PR.
Federico Ramallo (23:31) Right.
Stephen Klein (23:31) rather than saying, you we mismanaged the business, we over-invested, we’ve got to cut costs. So I think there’s a lot of that going on right now. So that’s another component to, I think, reality in the industry and something that doesn’t get reported an awful lot.
Federico Ramallo (23:47) Right, Something similar happened with COVID. Every death that were not necessarily related to COVID, but they pumped the numbers up. That was kind of the policy, right? So how much can you trust the reports, right? So it’s hard, right? Yeah.
Stephen Klein (23:57) Right. Right.
Well, especially because there’s, you know, there’s.
the interest of the organizations to generate that kind of positive publicity because when a company announces that it’s cutting people.
their stock price immediately goes up.
Yes. And when they say that they’re cutting people because of efficiencies they’ve gained through their strategic use of generative AI, their stock price goes up even higher. Doesn’t matter if it’s true. Doesn’t matter if it’s true.
Federico Ramallo (24:20) really? Interesting.
Well,
the stock valuation is based on the market’s perception, not necessarily truth, right? Yeah.
Stephen Klein (24:40) Exactly, Yeah.
In fact, on a side note, I think it’s fascinating that since 1980, half of all the S &P 500 profit in the U.S. has been due to layoffs and stock buybacks, not to innovation, not to investment, not to growth. Since 1980.
Federico Ramallo (25:01) Right.
Stephen Klein (25:01) So it’s fascinating. So it brings me back to the fact that
Federico Ramallo (25:03) interested.
Stephen Klein (25:06) If you go into the CFO at a company and you say, I can sell you this product. And if you buy this product, you’ll be able to eliminate a bunch of people and you’ll be able to save a lot of money. And you’ll be able to improve your margin, improve your profit, improve your share price and have a great phone call with wall street next month.
That CFO is going to like that conversation a lot. It doesn’t necessarily even matter if it’s true.
because you can realize the gain immediately.
Right. You realize the gain immediately. However, if I go into that same CFO, let’s say if you invest.
1 % of the salaries of your employee base or 2%. So if you invest money in this technology and you elevate and level up your people, level up your organization,
build accelerated innovation, grow your revenue, grow your top line, improve value you can deliver to your customers. But you need to invest today in order to realize that gain tomorrow. That’s not as easy a decision to make because now I actually have to spend money today rather than save it.
betting on a multiple unrealized gain in the future, which is what investment requires. That’s part of the challenge. So it’s much easier to sell into the enterprise an immediate gain through the elimination of jobs that you can put on the books that day versus
convincing that same company to invest money on improved top line growth in the future. That’s the challenge. That’s why the industry is selling the automation myth, because it’s easier to sell.
Federico Ramallo (26:57) It’s easier to see the value for the shareholders. Right? Short term, yes.
Stephen Klein (27:01) short term.
Yeah, now eventually that will end. It has to. Eventually, there’ll be a reckoning.
Because at some point that company is going to realize that it traded quality for speed, that it’s going to have to bring more people in and that it can’t cut its way to profit. You can’t grow a company by saving money. You can’t, it’s not possible. So eventually the cycle turns and
people realize they need to invest in growth and invest in revenue and invest in value creation for their customers. so eventually it will shift into that other direction. And that all depends on where we are in that cycle right now. And you could argue that we’re kind of
there now. I think we’re close. But that’s speculative. But you’re going to see visionary CEOs realize that they can flip the script and that with very little investment, they can actually implement this technology, augment their workers.
elevate their workers, improve productivity, improve value creation, improve top line growth, and you’re going to all of a sudden see a shift in that direction. Absolutely guaranteed.
Federico Ramallo (28:18) Interesting.
Stephen Klein (28:18) 100%,
100%. It has to happen.
And the smarter companies and the more visionary CEOs will lead the way because it takes courage.
and then the rest will follow and they’ll scramble and follow.
Federico Ramallo (28:32) And for the bigger organizations, it’s harder for them to adapt to changes, whereas smaller organizations will be able to do that much faster and catch up and be able to provide more value, become better with augmented AI, and thus create jobs.
Stephen Klein (28:37) Yes.
Exactly, exactly. And the small to medium sized businesses, not only can they turn faster, but to the point we were talking about earlier, that is the job creation engine for the economy.
Federico Ramallo (29:01) Right.
Stephen Klein (29:01) And that for me is why it matters because not only can we make a lot of money, and I do want to make a, I’m a Harvard MBA and I am running a for-profit business. So yes, do I want to create wealth? Absolutely. But I can create wealth in a way that benefits my clients, my customers, so that they can grow their businesses and create jobs.
That’s the goal. That’s the goal.
Federico Ramallo (29:23) I like the concept. think it’s a category in the IRS now, but it’s basically a for-profit organization with social mission. So, yes.
Stephen Klein (29:34) That’s what we’re doing. yeah,
yeah, that, that, yes, 100%.
I am a capitalist. I love capitalism. I love capitalism. studied at the Pentagon of Capitalism, Harvard.
I understand it. I’m passionate about it. I love it and I believe in it. But I believe that we can use that economic know-how and those skill sets and that expertise to not only grow businesses and make a lot of money, but to your point.
add value, do well and do good. Because without the ability to make a lot of money, you’re not going to be able to do any good.
Federico Ramallo (30:11) Yeah, I mean, the first step for being able to deliver the mission is sustainability, right? The business has to generate resources to sustain itself because then you can deliver the mission without external constraints, right? You don’t owe money, you don’t owe, you you ⁓ don’t have the constraints of external investors, you don’t have the constraint of ⁓
Stephen Klein (30:23) exactly.
Federico Ramallo (30:35) of donations, Because now you have to, instead of delivering the mission, now you have to, you know, talk to, you know, organize race funding, right? That’s a distraction for your mission. If the business is self-sustained, then, you know, you don’t need to do all that, right? And you can grow, right? You can grow the business because now you have more resources and you can fulfill the mission, you know, for longer, right?
Stephen Klein (31:01) And there’s only two things that you can do to accomplish that, either raise money or make money. Those are the only two options in the world. It’s pretty simple. know, either you raise the capital or you make money, but those are the only two choices any business has, any startup has.
Federico Ramallo (31:05) Right
Stephen Klein (31:17) That’s it.
And right now.
The generative AI industry is raising money and losing money. In fact, it’s losing a lot of money. No one’s making any money. None of them are making any money. Perplexity, anthropic, open AI are losing billions and billions of dollars with no path to profitability in sight.
Federico Ramallo (31:25) Great.
wow.
Yeah, I think this the world towards market share. We keep spending money because we want to keep the bigger market share possible because this market is going to be so big in the future that it kind of justifies doing that.
Stephen Klein (31:53) Yeah, I think that’s the story that they tell. But if you look at it more carefully, you realize that that’s probably not necessarily accurate. If for no other reason than the economics, the unit economics are upside down with these companies. So for example, OpenAI
Federico Ramallo (31:57) Yes, yes.
Stephen Klein (32:13) basically spends $29 for every $24 it makes.
Federico Ramallo (32:18) wow.
Stephen Klein (32:18) It loses money on every single customer. Even it’s $200 a month subscription customer. if you lose money on every customer because your variable cost is higher than your price, the old joke is you don’t make that up in volume.
And so ⁓ therein lies the fundamental economic challenge they all have upside down union economics, which is why scale doesn’t necessarily solve.
Federico Ramallo (32:33) Hahaha
it actually multiplies it because now you’re losing more money.
Stephen Klein (32:50) It actually you
actually you may have inverse economies of scale. Yep. Yep. Now, with an Amazon, for example, that argument worked because it was building infrastructure that was fixed cost. And then it was able to its unit costs were such that it could make money and it made a contribution. And Amazon could have made money.
Federico Ramallo (32:55) Right.
Stephen Klein (33:11) any time it wanted to, if it stopped reinvesting the money it was making in its infrastructure. But OpenAI and Anthropic, Perplexity, Mistral, they can’t do that. They’re raising capital to make up that shortfall. They’re not making money to make up that shortfall.
Federico Ramallo (33:29) So at the moment, the perception of the market changes, they’re going to implode. If they cannot keep raising money, their operating costs are going to eat them alive.
Stephen Klein (33:35) There.
I believe that there is almost a 100 % probability that that will happen. I believe, rightly or wrongly, I could be completely wrong, that a market correction or crash is inevitable. It has to happen. I believe that we are going through what one can only term a mania.
This is a mania. This is a psychological group insanity right now.
Federico Ramallo (34:04) It’s a social contagion, right?
Stephen Klein (34:07) It started the first, yeah, there’s a great book that was written in the 1800s called Extraordinary Popular Delusions and the Madness of Crowds.
It was written in the 1800s and it documented the previous economic bubbles, popular delusions. Starting in the 1600s with what was known as the Dutch tulip mania, where in Holland in the 1600s, there became this frenzy over the value of tulip bulbs to the point where people were selling their houses to buy tulip bulbs.
but there was no inherent value in the bulb. The only inherent value in the bulb was that you thought somebody else thought it was more valuable. That’s what a Ponzi scheme is.
Federico Ramallo (34:43) Right,
flipping it, yeah.
Stephen Klein (34:46) And that’s what I believe right now that I think is as long as people perceive that the value of these companies is going to be more tomorrow than today, more money will pour in and then the early money can cash out.
Federico Ramallo (34:59) Right.
Stephen Klein (35:00) That’s what’s happening.
Federico Ramallo (35:01) Interesting. ⁓
Stephen Klein (35:02) I believe that
that is what’s happening.
Federico Ramallo (35:04) I haven’t thought it that way.
Stephen Klein (35:06) And again, this is just me. Again, I have no direct communication with the future here. But in the 1920s,
the US stock market crashed, right? The late 20s, 29, and caused the Great Depression. And to a large extent, the cause of that was something called the pump and dump, which is another form of momentum investing, Ponzi scheme, mania, it goes by lots of names.
Federico Ramallo (35:17) Yes.
Stephen Klein (35:33) But it’s basically generate enough buzz that you can inflate the perceived value. So more money comes in value, more money comes in value, more money comes in until it doesn’t anymore. Then it stops, right? In 1934, the Securities and Exchange Commission was formed in the United States, the SEC to regulate the markets. There was no regulation up until that point.
Federico Ramallo (35:46) and then it crashed, yeah.
Stephen Klein (35:59) They outlawed the pump and dump.
Federico Ramallo (36:01) Right.
Stephen Klein (36:01) but they didn’t outlaw it in the private sector.
Federico Ramallo (36:04) see.
Stephen Klein (36:04) So technically, it’s fine if you want to try and generate perceived value for an investment. It’s fine. If you want to go out there and say that OpenAI is worth $300 billion,
even though it’s not going to make any money till 1929 or 1930, you can say that. And if people believe you, fine. And if people want to invest that improved valuation, fine. It’s a free market. But it’s not against the law.
Federico Ramallo (36:29) Yes.
Right. Right. Yeah, and I think that’s going back to circling back to what we talked at the beginning about bootstrapping. What happened with that perception is that it’s not real value, right? And when we’re bootstrapping, we’re looking to generate actual real value because that’s more sustainable long-term, right? So.
Stephen Klein (36:52) Yeah,
my customers are paying me for a product that they use and that I make money based on that price. That’s business. They believe that I have delivered value enough at that price point that I can still make money.
Federico Ramallo (37:02) Right.
Stephen Klein (37:11) That’s the only way you can build a business. That is how you build a business. That’s bootstrapping. Either that or you pay out of your own pocket and you support your own business, you know, through your own bank account. And that is what I find so rewarding that, you know, we can build a product with our sweat.
Federico Ramallo (37:15) Right.
Yes.
Stephen Klein (37:32) and our passion and our love, and we can have somebody buy it and appreciate it. I love that. I love that. I love that. So in a way, a company like mine is being funded by the customer.
Federico Ramallo (37:41) Yes.
Yes, but the decision they make to buy the product, they’re funding your company to deliver the mission.
Stephen Klein (37:48) Yeah.
Exactly, exactly.
At a high enough margin, profit margin that I can, you know, make money. Yep, exactly. That’s exactly right. Now, would capital for a company like mine be helpful? Absolutely. Which is why I’m going to try and raise some capital through WeFunder because that will enable me
to scale faster. Because then I can take that money and I can hire more engineers. I can enhance my product more quickly. I can introduce new products. And so I can scale more quickly, but in a methodical way. So at this point for somebody like me, the capital is not about
getting to profitability. The capital is about scaling faster, which is something that I want to do because I want to help more entrepreneurs. I want to help more small to medium sized businesses. yeah, so we’ll see. It’s all an experiment. I love it. And we have a small team. mean, we have, you know, there’s, 20 of us. Open AI has 5,000 people.
Federico Ramallo (38:45) Amazing.
Right. Right.
Wow.
Stephen Klein (38:59) Crazy, right? 5,000 people. It’s insane. Insane.
Federico Ramallo (39:01) and
You mentioned something at the beginning that I want to congratulate is you mentioned that you could have retired, right? And I want to congratulate that you decided to go out to the woods, right? Go to the adventure, take risks and, yeah.
Stephen Klein (39:22) I love it.
I love it. I wouldn’t trade a moment for the honor I feel doing what I’m doing. I love what I do.
find that it provides me with purpose. And I think purpose is the secret for happiness.
I think that when you find something to do that you believe is greater than you are, more important than you are, it is an extraordinarily motivating, energizing phenomenon. And so I found purpose, you know, and so for that, you know, I thank God every day and not only
do I feel that I’ve got purpose with the business that we’re building, but also teaching at Berkeley. I love teaching as well. So I’m very, I’m blessed. I really am. I’m very, very lucky, very fortunate. So.
I’m happier, the older I get, the happier I get. I’m the opposite.
Federico Ramallo (40:19) I was going towards that. Right. So when you’re doing something that gives you immediate gratification, right? And it serves a purpose that is significant to you. That immediate gratification multiplies, right? Because you’re doing something that serves both.
Stephen Klein (40:19) The older I get, the happier I get.
Federico Ramallo (40:39) And even if you’re doing something that doesn’t give you immediate gratification, but it’s contributing to your purpose, that suffering for calling it somehow, it doesn’t wait as much. We can endure that. And not only finding and identifying that purpose, but then being able to spend most of your time.
doing activities that fit that purpose. I think that’s, as you said, happiness, right?
Stephen Klein (41:04) Yeah, there’s a phenomenon, at least in the United States, where there’s a lot of depressed people.
Federico Ramallo (41:10) Yes.
Stephen Klein (41:11) There’s a lot of depressed people. There’s a lot of lost souls. And I think that the conventional wisdom is that the reason is that people are lonely.
and that they’ve lost their connection to other people through social media, through whatever. I don’t think that’s true. I think that people have lost their connection to purpose. And I believe that the problem is that life got too easy.
and as life gets easier.
we need to search for purpose more. Because life used to be hard.
and just
providing for our family was purpose.
That was very noble purpose. But as things get easier and easier and easier, and we think less and less and less, I think we are more lost. And so I think that the challenge right now for society is for people to try and reengage with things that are difficult.
and things that matter. And I think that that is really the only antidote, in my opinion, to making people feel better about what it is they do and who it is they are. so,
That’s why I guess I feel blessed that I have found that purpose.
Federico Ramallo (42:26) Yeah, when I, yes.
Stephen Klein (42:26) And it’s hard, and by the way, Federico, it’s hard too,
it’s hard, you know, it’s hard. It’s not easy to, I mean.
really hard doing what we’re doing. mean, it’s been three years and it’s been hell. You know, this has not been easy. If anybody thinks that, you know, being an entrepreneur and building a company is easy, they’re insane.
Federico Ramallo (42:42) Yeah, but
Stephen Klein (42:42) Because it’s a dream
and it’s hard, it’s a nightmare, and it’s a dream. ⁓ right.
Federico Ramallo (42:45) Right. Yes.
It’s a, you know, I read somewhere, know, starting a company is like looking at the abyss and just, you know, knowing that you’re going, going there and you’re, it’s going to be pain and suffering and you still decide to go, right. And you decide to keep going. But I think, I think hell is, is, you know, is sitting on the beach, right. And just, you know,
Stephen Klein (42:59) Yeah.
Federico Ramallo (43:11) which is my old definition of retirement, This sitting on the beach doing nothing, I have nothing else to do and doing that until I die. That sounds great the first day, the second day, but then it become hellish very quickly, right? Because you lost purpose, right? So even though there’s suffering, right? And there is, you know, there’s pain, right? It’s a voluntary pain.
That’s why it’s an adventure. We decide to go into that adventure knowing first hand that it’s going to be pain and suffering. But because we have a mission and we have a purpose, then that suffering becomes our endurance and becomes something that actually fits us with life.
Stephen Klein (43:58) Yeah, agreed. Absolutely.
Yeah, I couldn’t agree with you more. It’s ⁓
The other, I think, related ideal around that is relevance. I think that nobody wants to feel irrelevant. Nobody wants to feel, at least I don’t want to feel, that somehow I’ve been dismissed to the sidelines of life and I’m now a spectator and that the world is going on and on and on and I’m not relevant anymore. I mean,
And I think purpose and relevance are related, right? I mean, I think that.
being relevant, I guess you translate being relevant by feeling, again, back to feeling like what you’re doing matters, that it’s important, that there is purpose. So that’s the other, mean, I don’t know.
If you sit on a beach for too long, you’re going to become irrelevant.
Federico Ramallo (44:42) Mm-hmm.
Stephen Klein (44:42) You’re just like, why do you even exist? What’s your point, right?
Federico Ramallo (44:45) If you disappear
for a month, can kind of, yeah, no, no, no, you can’t get back. But if you disappear for a year, then by the time you come back, everything’s going to be different. And now you have to catch up, right? Yeah.
Stephen Klein (44:49) No one will know.
Yeah.
And the other thing is think about the fact that what you’re doing is relevant and what I’m doing is relevant. And so we’ve connected through it. In other words, you solved the loneliness problem and the community and connectivity problem through finding things to do that you love, because then you meet other people through that. And then you’re not alone anymore. I, I, I, I get to meet people like you.
you get to meet people like me. And that’s also an important part of, think, doing things that matter. You get to meet other people that are also doing things that matter. We tend to find each other.
Federico Ramallo (45:34) Right. Isn’t that amazing?
Stephen Klein (45:36) Right? It is amazing. It is amazing. It’s a part of how the universe works. Yep.
Federico Ramallo (45:41) Yes.
And with that note, we should wrap up because we are a little bit over the time. Stephen, I truly appreciated you being here. It’s been a very insightful conversation and I’m looking forward to continue the conversation.
Stephen Klein (45:53) It was a pleasure speaking with you, Federico. I appreciate it. Thank you.
Federico Ramallo (45:57) Thank you.