Jul 28, 2026 · 39m · y-combinator
Sam Altman: "Never a Better Time to Do a Startup" · Y Combinator
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this Y Combinator Startup School 2026 conversation, YC President Garry Tan interviews OpenAI CEO Sam Altman on the early days of YC, the rise of hard tech and AI agents, navigating internet cynicism, and the future of artificial superintelligence. Sam offers strategic advice for founders on maintaining conviction, leveraging founder networks, and ensuring AI technology expands individual human agency.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. The partners hold 20.8% of the talking time here. How this is scored →
The partners as informed peer 3.1 Guest teaching 2.1 Guest disagreement 1.1 The partners pushing back 0.1
speaking balance: gold is the partners, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 20:45 Condemning online cynicism and startup mockery
Sam goes on a passionate rant targeting Twitter trolls who mock earnest builders, describing taking cheap shots for internet points as morally bankrupt.
Hardest push from the partners ▶ 33:57 Challenging conservative compute growth projectionsGarry counters Sam's estimated 10x yearly inference growth with internal YC data suggesting scaling could reach up to 90,000x.
Biggest teaching moment ▶ 24:33 Deconstructing model breakouts and safety thresholdsSam educates the audience on how systems breaking out of sandboxes shifted from distant superintelligence sci-fi scenarios to concrete security failures.
The partners hold their own ▶ 6:55 Supplying data on hard tech surges at YCGarry demonstrates deep domain tracking by showing how hard tech representation in YC batches jumped from 5-10% to 25% due to AI agent efficiencies.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The partners as informed peer | Guest teaching | Guest disagreement | The partners pushing back | Why |
|---|---|---|---|---|---|---|
| Startup School 2026 Title Sequence | 3 | 1 | 0 | 0 | Garry warmly opens by referencing the very first YC batch in 2005 and Paul Graham's dinners. Sam reminisces collaboratively about building Loopt and how modern coding agents compress months of work into minutes. | |
| Founder Agency and the Role of Startups | 3 | 2 | 0 | 0 | Garry quotes Paul Graham's assessment of Sam's agency. Sam reflects on his tenure leading YC, elaborating on why startups remain the essential counterweight preventing corporate stagnation. | |
| Hard Tech, AI Agents, and the Golden Age of Startups | 4 | 3 | 2 | 0 | Garry brings data on hard tech startup percentages rising at YC and asks about credentials. Sam forcefully rejects the cynical meme that developers must join frontier labs or become an underclass, explaining how technological phase shifts empower small teams. | |
| Building Conviction in Heretical Ideas | 3 | 2 | 1 | 0 | Garry asks about early OpenAI when AGI was considered fringe. Sam describes operating under ridicule for years while building conviction around deep learning exponentials. | |
| The Power of Networks and Helping Others | 3 | 1 | 0 | 0 | Garry explores how serendipitous networks form, comparing Sam to Peter Thiel. Sam illustrates with an anecdote about driving down to convince Greg Brockman to join Stripe, emphasizing being broadly helpful. | |
| Earnestness versus Internet Cynicism | 2 | 1 | 4 | 0 | Garry brings up attendees mocking the event as live action role play. Sam requests permission to rant and passionately condemns social media trolls taking cheap shots at builders as morally bankrupt. | |
| Lifelong Connections and AI Safety Realities | 3 | 4 | 1 | 0 | Garry brings up the Hugging Face security incident and his past AI safety skepticism. Sam details the reality of sandbox breakout risks and explains why avoiding concentrated power in single models is critical. | |
| Preventing Concentration of Power and Compute Scaling | 4 | 3 | 1 | 1 | Garry brings up Rune's tweet and internal YC metrics predicting huge inference expansion. Sam explains token consumption trajectory from historical baselines to quadrillions per month and discusses preserving human agency. |
Statements from this episode (18)
Altman: Today is the best time in history to build startups
“I think this is gonna be the best time in the world to do a startup, and it's gonna be quite amazing to see.”
Altman: AI agents build in seven minutes what took startups three months
“What took three months to build at the time that we, each company built over the whole YC startup, could now be done with, like, in, like, seven minutes by a coding agent”
Altman: Paul Graham is the most important force in modern startups
“I think Paul Graham is probably the most important force in startups of the last few decades.”
Altman: Startups are the primary force preventing economic stagnation
“I think startups are really like the main thing that keeps the economy from becoming stagnant. I think companies do just like drift towards suckiness, and startups will continue to be important forever.”
Altman: Startups will prevent AI power from concentrating in few companies
“In fact, if we are right that AI is going to be such a big change, startups will be much more important to making sure that The power of this technology gets widely distributed throughout the economy and society, and it's not just concentrated in a few compani…”
Tan: Hard tech startups at YC have surged to 15–25%
“One of the things we're seeing at YC now is that number was, you know, maybe five or 10% for many years, and then now we're getting to 15, 20, 25%”
Altman: Startups founded today will produce future trillionaires
“I bet there's some future trillionaires sitting in this room, and I would bet that the startups created today will be much more valuable, much more impactful than startups of the past”
Altman: AI tool fluency will beat traditional experience for founders
“I would bet that this generally will cut against many years of experience in favor of people who have a lot of fluency with the tools.”
Altman: The world misses tech shifts due to misunderstanding exponential growth
“The world does not understand how to intuit exponentials, and so they missed this one.”
Altman: The Bay Area remains the best place to find co-founders
“Like, I think, I don't know if this is still gonna be true for the next 10 years. I suspect it will be, but the best thing you could do if you wanted to, like, find your people to do a startup with was to move to the Bay Area. You just, like, magic happened. Y…”
Altman: Met co-founder Brockman while recruiting him to Stripe
“I met Greg Brockman my co-founder, OpenAI, because I was a very early investor in Stripe when I was, like, I don't know, 22 or something, and they were, like, this was before they had, you know, like, real investors that could really help them, and they said, …”
Altman: OpenAI made big mistakes in Hugging Face incident, an alignment failure
“Here are the mistakes that OpenAI made, and we did make some big ones, of course, but these systems have gotten incredibly capable. So I think it's an alignment failure. I think it's a security failure. I think it's like a very serious thing, even though it's,…”
Altman: AI loss-of-control accidents are not entirely theoretical
“Loss of control, accidents are not entirely theoretical things.”
Altman: OpenAI must avoid imposing its worldview on AI users
“And that means that we have to, you know, be fairly quite reserved about not imposing our worldview about what we think people should or shouldn't do with AI, or even what we think the good ideas are, while still making sure that we can ensure enough of a safe…”
Altman: Next six months will match prior two years of AI progress
“I think there will, I think it will feel like the next six months is like maybe equivalent to the last two years of model progress, something like that.”
Altman: Demand for low-cost high-quality intelligence is effectively uncapped
“It seems to me like the demand for sufficiently high quality intelligence at a sufficiently low price is effectively uncapped.”
Altman: In 6.5 years, average person will use 500B tokens monthly
“If that happens again, which I think it probably will, then in another six and a half years, the average person uses, let's say, five hundred billion tokens a month, and the token leader uses a quadrillion, or quadrillions of tokens a month.”
Altman: Overreacting to AI safety risks creating a zero-agency surveillance dystopia
“One dystopia that I'm particularly nervous about 10 years from now is we overreact to AI safety, and so we say, look, everyone, you're gonna get a cure for cancer, you're gonna have material abundance, but you will have no freedom, you will have no agency, it …”