The Lean AI Stack for a Solo Founder
You don't need fifteen AI subscriptions to run a one-person business. You need one tool per role — the jobs you'd otherwise hire for. Here's the named stack I'd actually start with.
- #entrepreneurs
- #founders
- #tools
Sam Altman has a running bet with other tech CEOs on the year one person builds a billion-dollar company — something he's called unimaginable before AI. You don't need to chase that headline, but it points at something real: a solo founder can now cover the roles they used to hire for. The catch is the stack. Collect tools and you get a zoo; pick one per role and you get a company.
Think in roles, not products
A startup — even a one-person one — has four jobs that never stop: decide what to build, build it, get the facts right, and keep the lights on. Map one tool to each and the stack picks itself.
The stack I'd actually run
- Decide & draft — Claude or ChatGPT. Your frontier-model co-founder: strategy calls, drafting, thinking a hard decision through out loud. It's your most-used hire, so pay for the top tier and don't ration it. (Anthropic even ships a Claude for small business setup aimed at exactly this.)
- Build & ship — Claude Code or Cursor (code), Canva (visuals). Whatever produces what you sell or market. Claude Code and Cursor turn intent into working code; Canva's Magic Studio gets you to "good enough" graphics without a designer.
- Get it right — Perplexity. Citation-first research for the market and competitor questions you'll have to defend. When being wrong has a cost, use the tool that shows its sources.
- Keep the lights on — Notion AI + Stripe. Notion AI for docs, SOPs, and recurring admin; Stripe to ask revenue questions (MRR, churn, failed payments) in plain English. The boring half of the business still has to run.
Four roles, four or five tools. That's a complete stack — stop there.
Where AI doesn't replace you
It won't decide what's worth building. Sam Altman's Startup Playbook has aged well here: talk to users, build something people want. That judgment matters more when AI makes building cheap — when anyone can ship in a weekend, taste about what to ship is the moat. Keep that job for yourself.
Start here this week
Don't just "sign up" — run one real task through each starter tool so it earns its keep on day one:
The co-founder (Claude / ChatGPT):
→ Paste your biggest current decision and ask it to argue both sides,
then recommend one. Ship the decision by Friday.
The maker (Claude Code / Cursor, or Canva):
→ Build the smallest real thing you sell or market — one working
page, script, or graphic — start to finish, today.
This week
Write the four roles and the one tool you'll use for each. Start with two — the co-founder and the maker (run the tasks above) — and add the rest only when a real need shows up. A lean, deliberate stack beats an impressive-looking pile every time.
Sources
Keep reading
More guides like this one.
ChatGPT Work vs Claude Cowork: the answer is in your admin console
Two workspace agents shipped 48 hours apart. The thing that actually decides which one your team runs isn't the model — it's a toggle an admin flipped before anyone got to try either. Here's the seam, and my routing rule.
A 3-Step AI Cost Audit: Seats, Tokens, and the Hours Nobody Books
Two thirds of companies have an AI dashboard. Barely a quarter can actually see what AI costs them. Here is the 3-step audit I run on a tooling budget: what to list, which number to pull for each line, and the threshold that means kill it.
97% vs 47% — Your Job Title Says Nothing About Your AI Exposure
Anthropic's index found 97% of tasks are automation-dominant through the API and 47% in chat. Same model, overlapping tasks. Exposure follows the deployment, not the job title — so stop forecasting the fog and run the migration test instead.