Andrew Wilkinson on Business Ideas, AI Automation, and the Limits of Money

Andrew Wilkinson on Business Ideas, AI Automation, and the Limits of Money

transcriptlennys-podcastentrepreneurshipbusiness-ideasai-automationbootstrappinghappinessadhdtiny

Andrew Wilkinson on Business Ideas, AI Automation, and the Limits of Money

Source: Lenny’s Podcast Speaker: Andrew Wilkinson Date: ~2025 Link: https://www.lennysnewsletter.com/p/ive-run-75-businesses-andrew-wilkinson

Key ideas

  • Fish where the fish are; boring beats glamorous. Charlie Munger’s heuristic applied to startups: the best opportunities are where everyone else isn’t looking. Cafes and restaurants attract maximum competition because everyone imagines the lifestyle; pest control, government form-filling software, and pressure washing attract almost no one. First-time entrepreneurs should start with the lightest possible business (web design, not AI company) to build narrative and muscle before attempting heavier lifts. The niche market’s ceiling is often much higher than it looks once you can hire away from the work.
  • Moat over growth rate. Tiny buys profitable bootstrapped businesses and holds them forever. The acquisition filter: brand, network effects, or switching costs — something that makes the business hard to mess up and hard to compete with. Letterboxd (largest film social network — network effects), AeroPress (brand), Serato (DJ software — deep hardware integration + passionate user base). No venture capital: if the moat is real, the numbers compound naturally without burning cash on customer acquisition.
  • AI as personal chief of staff. Andrew runs a suite of Lindy agents across his inbox: auto-archives threads he doesn’t need to see (−20% inbox volume), flags time-sensitive emails, surfaces simple decisions as multiple-choice prompts so he can reply with a single number, auto-syncs contacts to CRM, and briefs him 30 minutes before any meeting via a Perplexity deep dive. Also: Replit (Claude 4) for vibe-coded web apps, Limitless wearable for commitment tracking and relationship coaching, Gemini 2.5 for long-document tasks. “It’s like having the world’s most reliable employee who costs $200 a month and works 24/7.”
  • We are in the Palm Treo phase. The current AI agent toolchain is the Palm Treo before the iPhone: powerful for those with time to build, but not yet self-assembling. The real disruption comes when an AI can interview you, identify your role, and spin up a complete agent stack without setup — the “Her” moment. Until then, getting fluent with current tools is a compounding advantage. Job displacement is already real for researchers, admins, translators; the question is how far it extends into knowledge work when models reach “smarter than any PhD.”
  • Money doesn’t quiet the anxiety loop; medication did. Scaling from $20M to $300M in revenue produced identical anxiety. Multi-billionaires still covet the next yacht above them. The intervention that worked where wealth couldn’t: SSRI for anxiety (took the volume down on the internal critic) and ADHD medication (Times Square brain became a quiet library). DNA testing (23andMe + pharmacogenomics) let Andrew pick an SSRI he could metabolise without side effects. Approximately 30% of entrepreneurs have ADHD vs 5% of the general population.

Overview

Andrew Wilkinson — co-founder of Tiny, the bootstrapped holding company behind Letterboxd, Dribbble, AeroPress, Serato, and 40+ other businesses — covers how to find great startup ideas (fish where the fish are), when not to raise venture capital, how he’s replaced his EA with Lindy agents, and why none of it solved his anxiety. Unusually candid episode on the limits of money, the utility of SSRIs, and the reality of ADHD entrepreneurship.