Key points
- Sam Altman tells Ti Morse that AI has made small teams faster and more powerful, so startups gain their biggest edge when the ground shifts most.
- Altman says most founders chase easy wins with AI agents in enterprise verticals, and he is surprised more people do not attempt ambitious projects that assume scaling continues.
- Altman names AI authoritarianism as one of the biggest risks and says OpenAI wants power to stay decentralized, with no desire to build every vertical startup on its platform.
Sam Altman says AI gives small startups their biggest edge
AI has made the startup edge larger, because small teams can now do far more and the ground keeps shifting. Altman thinks most founders settle for easy wins while too few attempt projects that assume scaling continues. OpenAI follows the same logic by dropping projects to back what works, while arguing that the resulting power should stay spread out.
Takeaways
- AI widens the startup edge. Altman says AI has changed what a small team can do and how fast, so a ten week old startup now looks very different. Startups hold their biggest edge when the ground is shifting most.
- Most founders aim too low. Most are chasing easy wins with today's AI agents for enterprise verticals. Altman is surprised more people do not attempt ambitious projects that assume scaling continues. He says many hold rigid worldviews and end up chasing trends.
- Resilience comes only from lived experience. Altman says operating in chaos cannot be taught, and young founders handle it poorly because they lack career experience. He finds it easy to be grateful for bad days, since the opposite of a bad experience is no experience.
- AI power must stay decentralized. OpenAI's mission is to make AI abundant, cheap and powerful, and to put it in everyone's hands. Altman names AI authoritarianism as a top risk and says OpenAI has no desire to build every vertical startup on its platform.
- OpenAI cut projects to back what worked. OpenAI shut down projects such as robotics and Sora to focus on what was working, including Codex. Sora would not have been super successful, so compute and energy went to coding agents. Codex began as a bet against Claude Code that a team accepted.
- Transistors are the bottleneck, then electrons. Altman says transistors come first and electrons second. OpenAI aligns its suppliers by showing them upcoming models and research. He treats electron production as more of a commodity that does not need to be brought together with OpenAI.
- Small wins build self-belief and ambition. Altman says small repeated wins build self-belief. New YC founders from corporate backgrounds often lack ambition, because people at big companies are punished for thinking too big or wanting too much.
What comes next
- Altman forecasts a third wave of persistent agents.
- In a post superintelligence world he expects busier lives.
- Human activities like cooking should continue, since betting against evolutionary biology is usually a bad bet.
- The four hour workweek promised by technology never arrived at mass scale, and he does not expect AI to change that.