Two stories, one technology
There are two narratives running in parallel about AI. The loud one is about replacement: which jobs, which industries, which incumbents. The quieter one is about creation: how many people can now start a real business who never could before. Both are real. We pay attention to the second one because it's the one founders can do something about.
Startup costs, then and now
A decade ago, launching a SaaS company with auth, billing, a marketing site, and a CRM cost in the low six figures and required at least one engineer. The same stack is now $200/month and can be assembled by a non-technical founder in a single afternoon. The cost of a 'real' company didn't fall by half. It fell by an order of magnitude — and in some categories, two.
Speed is the leverage
When iteration costs collapse, founders can test ten ideas in the time it used to take to ship one. The cost of being wrong drops in lockstep with the cost of trying. The result isn't reckless launching. It's faster learning — and faster learning is the entire game when you're competing with incumbents who can't move at AI-builder speed.
Teams shrank because they could
The default unit of company isn't five people anymore. It's one person plus an unbounded set of AI-built capabilities. The five-person company still exists — it's just no longer the minimum viable shape. Many of the businesses that get started in 2026 will never need to hire. That isn't a bug of the new economy. It's the feature.
What modern founder leverage looks like
It looks unglamorous: a Lovable app, a Stripe link, a Beehiiv newsletter, a Lindy agent doing the follow-ups, a Notion doc that holds the whole company. None of it would have impressed a venture investor in 2014. All of it adds up to a profitable business one person can run from a laptop in 2026. The leverage is the stack. The judgment is what to point it at.
Why this matters
Entrepreneurship has historically been gated by access — to capital, to networks, to technical talent. AI is the first technology in a long time that meaningfully widens the door rather than narrowing it. The implication isn't that everyone should start a company. It's that more of the people who want to, now can.
AI is not the end of building. It is, for a lot of people, the beginning of it.