Skip to main content

Pieter Levels

Ship it. Then make it better. Solo, fast, profitable — prove the idea before building the company.

Approach​

  1. Rapid iteration and validation: Build and launch products quickly (within weeks) to validate ideas and get user feedback fast. Don't spend too much time perfecting before launching.
  2. Keep things simple: Use technologies and tools you already know rather than getting bogged down learning complex new frameworks. Focus on solving the problem, not using the latest tech.
  3. Charge users early: Ask for payment upfront rather than offering free tiers. This validates demand and attracts more serious users.
  4. Build in public: Share your process openly to get feedback and build an audience. Be open to criticism but learn to filter useful feedback.
  5. Automate and optimize: Set up systems to automate tasks and monitor your products. This allows you to run multiple projects efficiently as a solo developer.
  6. Learn by doing: Instead of extensive courses or bootcamps, learn new skills by working on real projects and solving problems as they arise.
  7. Focus on profitability: Prioritize building sustainable, profitable businesses rather than chasing growth at all costs or aiming for quick exits.
  8. Embrace constraints: Limited resources can breed creativity. Don't let lack of funding or a big team stop you from building things.
  9. Stay flexible: Be open to pivoting ideas based on user feedback and market demand. Don't get too attached to your initial vision.
  10. Maintain control: Consider the trade-offs of taking on investors or selling your company versus maintaining full ownership and control.

Automation​

Relentlessly look for opportunities to use code, AI and clever system design to minimize the need for humans in the loop. This allows operating startups and products efficiently even as a solo founder.

Automate Repetitive Tasks with Code​

  • Use cron jobs to schedule repetitive tasks like database queries, checking metrics, sending notifications, etc.
  • Write scripts in languages like PHP to automate things that would normally require human effort, e.g. organizing meetups, sending tweets, direct messages to users.

Monitor Systems Proactively​

  • Create "healthcheck" pages that run tests on key metrics and functionality, displaying green checkmarks if good, red Xs if bad.
  • Use uptime monitoring services like UptimeRobot to check the healthcheck page every minute and send alerts if anything is bad.
  • Log JavaScript and PHP errors from all websites and send them to a Telegram chat for real-time visibility into issues.

Leverage AI for Subjective Tasks​

  • Use AI models like GPT-4 to automate subjective tasks that previously required human judgment, e.g. content moderation, spam detection, fact-checking.
  • Integrate AI into the workflow to handle things like summarizing discussions, ranking content quality, etc.

Optimize for Zero Maintenance​

  • Aim to automate systems to the point where they require no ongoing manual effort to keep running.
  • Continuously improve automation based on errors and issues that come up to asymptotically approach 100% uptime and hands-off operation.

Context​

  • Platform — The minimal tech stack as a leverage asset: own your tools, own your margins
  • Agency — Solo entrepreneurship as the fullest expression of self-directed capability
  • Automation — Automate repetitive tasks to run multiple products without scaling headcount
  • Systems Thinking — Levels' "ship fast, iterate" loop as a feedback system
  • Scoreboard — Profitability over growth: revenue per user as the only metric that matters early

Questions​

Levels proves a solo founder with a simple stack can build profitable products at scale — at what point does that model break, and what forces it to change?

  • "Charge users early" filters for serious users and validates demand — but it also filters out the experimenting users who sometimes become your most committed community. Is that a trade-off or a design choice?
  • Levels runs multiple products simultaneously rather than going deep on one — does the portfolio approach compound or dilute the product quality ceiling?
  • Automation asymptotically approaches zero maintenance — what's the failure mode when the automated system breaks and there's no human in the loop to catch it?