Episode 3: Solo Founders

Lee thinks the number might be ambitious, but the underlying idea is sound. If you can orchestrate the support functions of a business through AI agents while you as a human focus on the core value - the thing you’re actually good at - why can’t a single person build something worth hundreds of millions? James sets out to play devil’s advocate. By the end of the episode, he’s mostly been talked round.

Lee frames the argument around a simple truth: most business roles exist not because the work is complex, but because a single person can’t context-switch fast enough to do everything. A solo founder who’s brilliant at their craft still needs someone to handle accounts, scheduling, lead generation, marketing, operations, and customer service. Traditionally, that means hiring people, which means raising money, which means giving up equity, which means compromising on vision. But what if all of that operational overhead could be handled by AI agents running on compute that costs a fraction of a human salary?

He illustrates this with a snow plower analogy that becomes the episode’s defining example. Imagine someone who clears driveways for a living - a manual, non-technical job. They’re constrained by 24 hours in a day and seven days in a week. They can raise prices, but there’s a ceiling. They can hire people, but that means management overhead, HR, and all the complexity that comes with employing humans. Now imagine AI handles the scheduling, the customer communication, the invoicing, the advertising, the lead generation - everything except the actual snow plowing. The compute cost might be $30,000 a year. Suddenly, instead of needing to make a million dollars to justify a support team, they need to make $150,000 to cover themselves, the AI costs, and taxes. The threshold for a viable business drops dramatically. And if they want to scale, they find other passionate snow plowers, pay them $90,000 instead of minimum wage because the operational costs are so low, and build a team of people who actually want to be there.

James raises the economic counterpoint that gives the episode its tension. If the barrier to entry drops this low, doesn’t everyone start doing it? And if everyone’s a solo founder with AI agents, isn’t it a race to the bottom? Someone sees you making good money clearing driveways, copies the model, undercuts your price by 50 percent, and suddenly the market is flooded with cheap competitors. The overall economy contracts because there are fewer jobs, less spending power, and a growing population of people who don’t have the tech literacy to participate in this new model. Are we heading toward something that looks fundamentally different from the economy we have now?

Lee doesn’t dodge the question. He acknowledges the risk but argues the alternative is worse - big companies dominating the space, laying off 90 percent of their staff, and concentrating wealth even further. At least with a wave of micro-businesses, the value is distributed. He also makes the point that innovation has always caused disruption. The motor industry replaced assembly line workers with robots, and people predicted mass unemployment. What actually happened was a reduction in headcount, yes, but also the creation of entirely new roles that didn’t exist before. It would be bad for humanity, he argues, to still be doing what we did 50 years ago. Now we’re folding proteins and building rockets. In 100 years, you might be on another planet for a weekend holiday.

The conversation gets particularly honest when they discuss the sales challenge. How do you walk into a company as a solo operator and convince them you can do what their five-person team couldn’t? Lee has experienced this firsthand in his consulting career, where he’s had to tell clients he can write the theme tune, sing the theme tune, produce the film, and do the marketing - and they’ve looked at him like he’s mad. The difference now is that AI has become the ultimate justification. You say “I use AI agents for this, I’ve automated this flow, I built this tool” and because the big tech companies have spent billions marketing AI as transformative, the client nods along. The hype, ironically, has done the selling for you. Lee calls it the biggest cop-out defence ever - and he means it as a compliment.

James shares his own version of this, describing situations where he’s been pitched to clients as a one-man army. The scepticism is real - people see a single contractor and compare them to their internal team of developers who can barely keep up. But Lee reframes it: the conversation needs to shift from billing by the hour to billing by outcome. If a project that would have taken six months with a team now takes two weeks with one person and AI, the value hasn’t decreased - it’s increased. The client gets their solution faster and cheaper. The developer makes good money for focused, high-value work. The hourly model just has to die.

They also explore what Lee describes in his book DPR - the shift from using AI to build tools to using AI to replicate roles entirely. It’s not about building a better CRM; it’s about having an AI agent that does what your financial controller does, what your marketing coordinator does, what your customer service representative does. Not perfectly, not for every business, but well enough for a solo founder who understands their domain and can validate the output. The person who benefits most from this isn’t the tech specialist - it’s the generalist who’s been freelancing for years and already does 10 percent of everyone’s job. They know enough about accounts, sales, operations, and delivery to orchestrate AI agents across all of those functions. The gap between “I know a bit about this” and “I can run this function” has been closed by AI.

The episode closes with both hosts reflecting on what this means for human connection. Lee believes that despite all the automation, people still want to engage with people. He envisions a world of passionate solo business owners sharing co-working spaces - not the WeWork model, but genuine communities of people who care about what they do. If you can listen to someone talk passionately about their work - whatever it is - that’s worth something. And if AI can remove the drudgery that stops people from pursuing their passions, enabling more of that in the world isn’t a bad outcome.

James admits he’s been convinced, though he’s still working out how to get on the bandwagon before the economy reshapes itself. His plan? Make his millions and retire to Mars for the weekend. Lee suggests starting with a doorstop business. They agree to keep score of who convinces who across future episodes - because apparently they’re not competitive enough already.

Subscribe to The AI Shift for more honest conversations about AI and the real world.