Paul Graham On Startups, Ambition, and Great Founders
Paul Graham discusses YC's evolution over 21 years, arguing that startups funded today are more ambitious than in the past, and explores how founder ambition, shipping speed, and formidability drive success. He shares perspectives on AI's surprising development as a 'bullshitting' system rather than a perfect-to-complex progression, and explains how YC's batch model has remained fundamentally unchanged despite scaling.
Summary
In this conversation at Y Combinator's Mountain View offices, Paul Graham reflects on his experience directing the 47th YC batch in the organization's 21st year. He addresses the common criticism that YC has "jumped the shark," arguing that this complaint has existed since 2008 and that founders are actually more ambitious today than in earlier years. Graham uses examples like intercontinental ballistic cargo delivery and cancer research startups to illustrate current ambitions, contrasting them with earlier investments like Reddit.
Graham elaborates on the concept of "frightening ambition" he coined in 2012, noting that several ideas from that essay have since materialized, particularly OpenAI as a new approach to search that rendered traditional Google-style web search obsolete. He explains that true founders must possess genuine ambition beyond mere talent, driven primarily by fear of failure rather than dreams of billionaire status. Graham reveals that most founders don't realize their wealth until well into their success, describing how he often has to calculate their net worth for them.
On AI, Graham discusses how its development surprised him. Rather than progressing from perfect simple systems (like a fly's brain) toward human-level performance, AI instead emerged as a full-blown human with significant flaws—comparable to an undergrad bullshitting through a paper. This reversal means AI now progresses toward perfection rather than starting perfectly simple. He notes this explains why AI can solve complex math problems but struggles with basic queries like restaurant hours.
Graham addresses whether lean startup methodologies are obsolete, expressing that he hasn't read the definitive text but maintains that startups can still begin with limited capital by achieving incremental milestones. On AI's impact, he contends that almost everything about startups remains the same, with the primary new cost being GPU and token expenses rather than salaries.
Regarding YC's structure, Graham explains the batch model emerged accidentally from the original angel firm concept. They initially aimed to fund multiple startups simultaneously to learn investing, modeling it as an alternative to summer jobs for college students. The model proved so effective that it became YC's permanent approach. He emphasizes that despite scaling significantly, YC has fundamentally changed very little—individual cohorts function like small autonomous units of the larger organization.
Finally, Graham asserts that the next trillion-dollar company will come from formidable founders rather than specific ideas, and predicts that future founders will look remarkably similar to those of the past 20 years, dismissing speculation about radical changes.
About this episode
<p>YC Visiting Partner Vivian Shen sits down with Paul Graham at the original YC office in Mountain View to talk about startups, AI, ambition, and what makes great founders.</p>
Key Insights
- Graham argues that founders are primarily motivated by day-to-day fear of failure and disaster rather than the prospect of becoming billionaires, which often surprises them years into success when he calculates their net worth.
- AI development followed the opposite of expected progression—beginning as a flawed, bullshitting system comparable to an undergrad trying to pass a paper, then working toward perfection, rather than starting perfect and working upward as AI researchers once theorized.
- Graham claims that shipping speed remains the best predictor of startup success even with AI tools available, and observes that many current YC founders still don't ship fast enough regardless of technological capabilities.
- The YC batch model was discovered accidentally when Graham's team decided to fund multiple startups simultaneously to learn investing, modeling it after summer jobs for college students, yet it revealed structural advantages that became permanent to the organization.
- Graham contends that YC has fundamentally changed very little despite scaling from dozens to hundreds of startups because individual cohorts function as autonomous units similar to the original structure, making the experience consistent for founders.
Topics
Transcript
So we're here in Mountain View today with PG himself at the original offices doing your talk that you do for the YC batch. I looked it up. This is actually the 47th YC batch. You know, I've been trying to calculate that number and I wasn't sure exactly what it was. No wonder I'm not too worried about this talk. You know, I've done it that many times. It changes a little bit. Yes. It's the 21st year also of YC. You know, the weird thing is I make the talk from scratch every time. Really? I probably say the same things over and over, but I always feel like I should, like, I think, oh my God, you…
Full transcript available for MurmurCast members
Sign Up to AccessMore from Y Combinator Startup Podcast
The World’s Largest Electric Aircraft Just Flew
HART Aerospace has successfully flown the world's largest electric aircraft, a 100-foot wingspan hybrid-electric plane designed to reduce regional air travel costs. The company, founded by Anders Forslund, evolved from a 3D-printed model seven years ago to a 40-person team building a commercially viable airliner that swaps jet engines for electric motors.
Max Junestrand: You Need The Willingness To Learn Faster Than Anyone Else
Max Junestrand of Legora describes the company's rapid growth from zero to $100M ARR in 18 months by building an AI operating system for lawyers. He discusses the importance of learning the market deeply, maintaining competitive culture, hiring for growth potential rather than prestigious credentials, and betting on improving AI models without fine-tuning.
Susan Kare: Designing Icons & Graphics For the Original Mac
Susan Kare, the iconographer for the original Macintosh, discusses her journey designing system fonts, icons, and graphics under severe technical constraints (16x16 black and white pixels). She shares design principles learned from mentors like Paul Rand, her experiences across multiple tech companies, and the philosophy that simplicity, metaphor, and meaningful design create universal, memorable user interfaces.
Chelsea Finn: This is the State of the Art in Robotics
Chelsea Finn from Physical Intelligence discusses advancing robotics toward general-purpose models that achieve high reliability through reinforcement learning, memory systems, and diverse training data. She demonstrates how robots can perform complex real-world tasks autonomously and argues that physical AI has reached a ChatGPT-like era comparable to language models, with models now being deployed in real-world applications.
Peter Steinberger: "Fun Is Velocity"
Peter Steinberger discusses the rapid evolution of OpenClaw, the challenges he faced in scaling the project, and the importance of maintaining fun and creativity in software development. He reflects on the frustrations and lessons learned while navigating public attention, security concerns, and team dynamics.