Gary Tan reflects on his early career, noting that he repeatedly made the mistake of chasing what was "hot" rather than following his own intellectual curiosity. In 2003, despite being a skilled web programmer, he abandoned the web because the market felt dead, missing the rise of social media. He emphasizes that founders often ask investors "what's hot," which is the wrong question. Instead, they should ask what they are uniquely interested in and what they know that others don't. He argues that "earnestness" is a form of maturity—the courage to trust one's own direct experience over the prevailing market narrative.

The Cult of Truth and Outsider Status

Drawing on his experience at Palantir, Tan notes that every meaningful project in his life has felt like a "cult"—a group of people united by a belief that flies in the face of orthodoxy. He suggests that the most interesting ideas often start as "toys" or fringe obsessions. The key to innovation is finding the edge of the conversation and, more importantly, the intellectual edge. He credits Y Combinator with democratizing access to Silicon Valley by replacing the need for social connections with a meritocratic process where ideas and execution matter more than pedigree.

The New Era of Agentic Coding

Tan highlights a massive shift in how software is built. He notes that "code is no longer precious" and that the barriers to shipping have collapsed. With "vibe coding" and agentic workflows, a single founder can now act as a team of hundreds. He encourages founders to "token max"—using agents to process massive amounts of data (e.g., 800,000+ tokens) to gain a competitive advantage that feels like living in the future. While this is expensive, he argues it is a necessary investment for founders who want to operate at the frontier of what is possible.

Key Takeaways

  • Trust your own experience: Don't let market sentiment or blog posts override what you know to be true from your own work.
  • Find your tribe: Use the internet to find the "weirdos" who share your specific obsession; that is where the most interesting work happens.
  • Don't be a lone dancer: A founder alone is a crazy person; a founder with a co-founder or a small team is the beginning of a movement.
  • Token maxing: If you are building with AI, don't be afraid to push compute and token limits to their maximum to gain deeper insights and faster iteration.
  • Build trivial things: Use new AI tools to build small, low-stakes projects just to develop your intuition and "chassis" for the technology.
  • Avoid the 'scener' trap: Don't waste time trying to crack social circles; focus entirely on building and execution.

Notable Quotes

  • "Everything that's awesome in my life is kind of a cult that starts with some sort of truth or belief that flies in the face of an orthodoxy."
  • "The right question is not 'what's hot,' but 'what are you interested in and what do you know uniquely?'"
  • "If you know something, it sort of doesn't matter if someone contradicts you... because I believe my own direct experience."
  • "A cult where there's only one person and that person can't convince anyone else to join them is worse than a cult—it's just a crazy person."