2026-07-28

In 2005, @sama was a Stanford sophomore in YC’s first batch, building a startup in a little Cambridge office while @paulg cooked the founders dinner.
Twenty years later, as co-founder & CEO of OpenAI, Sam closed Startup School 2026 in conversation with @garrytan — on agents, ambition, and why the ceiling for what a startup can take on has never been higher.
00:07 — From YC’s First Batch to Today
03:04 — What PG Taught Sam
04:17 — Why Startups Matter More Than Ever
06:57 — The Coming Golden Age of Ambitious Startups
09:30 — Why Startups Win During Technology Shifts
11:46 — Building OpenAI When Nobody Believed in AGI
14:45 — Finding People Who Share Your Conviction
18:05 — Help People Before You Know Why
19:43 — Earnestness, Ambition, and Ignoring the Haters
24:04 — The AI Safety Incident That Changed the Stakes
26:58 — Preventing AI From Concentrating Power
30:41 — How Fast AI Models Will Improve
36:06 — The Best Version of an AI Future
37:47 — “It’s All Going to Work Out”

YIMBYs vs our opponents

"How do people learn to use Claude Code like Boris?"
Claude Code creator Boris Cherny (@bcherny ) was asked by Diana Hu ( @sdianahu ) at Startup School 2026.
Boris Cherny: "Maybe don’t listen to the LinkedIn influencers.
The thing about the model is that everyone is looking for one weird trick to make it work, but that simply doesn’t exist. There is nothing like that.
You have to approach the model empirically. Give it a task that is too hard. Give it the tools to verify its work, just as you would use those tools yourself if you were doing the task.
See where it struggles, and then fix the problem with better prompting, a skill, or if the model is missing context, an MCP connection that allows it to retrieve the context it needs.
That is basically it.
It sounds very simple. I think people tend to overthink and overengineer it because, in many ways, that is how we had to build systems in the past."
----
From "Y Combinator" YouTube channel, (full video link in comment)

best time to start a company F26 deadline is in ~2 hours (we also accept late applications too)
Aaron Epstein: “This is absolutely the single greatest time to start a company, I’m jealous of all of you” @JensenHuang at Startup School

The YC Office Hours Simulator was a huge success! 
These were physical booths at Startup School where attendees could talk to AI avatars of the @ycombinator Partners. We interviewed over 100 attendees about their experience.
Here's a recap 👇️
1️⃣ The #1 complaint = people wished they could talk longer! Most conversations were 10min (the enforced time limit). There was a line of people waiting to talk to the avatars for the entire two-day conference
2️⃣ Some people wished the avatars were meaner or harsher. This was interesting!
3️⃣ Many people took notes (pen & paper, @meetgranola, or typing furiously) and said the conversation was genuinely helpful. Several folks waited in line for a 2nd or 3rd conversation :) 
4️⃣ People said the avatar aspect felt different than a chatbot. The "face-to-face" interaction helped them feel a stronger connection, pay more attention, and open up more. A few people even said they forgot it was AI! 
5️⃣ The magic moment = when people realized the avatars knew who they were and what they were working on (based on their conference application) 🪄
The "brains" for Office Hours Simulator was trained on hundreds of hours of real office hours. It truly simulated how each Partner talks and the types of questions they like to ask. 
Built by the one-and-only @snowmaker and YC software team (so impressive) 👏
Powered by: @lemonslice (live avatars), @elevenlabs @ElevenLabsDevs (voice AI), @livekit (webRTC infra), @deepgramai (voice AI), @ployai (animations), Hamming AI (testing), and many other great companies.

The negative AI jobs outcome just continues to not be happening as some predicted. A large portion of enterprises I talk with -across industries- are still hiring, just with a tilt in the kind of roles they’re going after.
They’re finding that AI is letting them do more, and Jevons paradox is actually playing out. They’re hiring engineers to go after problems they couldn’t tackle before. They’re hiring in sales because they can go deeper in client relationships now with the help of AI. They’re hiring in internal FDEs to help them deploy AI. And so on.
Anyone using AI merely to cut costs eventually just gets outcompeted by companies that use AI to better serve their customers and drive more breakthroughs in their business.
Could this trend change at some point? Sure. But for now this appears to be the trajectory we’re on.
David Sacks: Narrative violation. WSJ today.

This was such a fun project to work on and came out way better than I expected. The models the @LemonSliceAI team builds are magical.
If you missed getting to try these at Startup School, they're going to live on in the @UseCorgi Cafes world-wide!
LemonSlice: The YC Office Hours Simulator was a huge success! 
These were physical booths at Startup School where attendees could talk to AI avatars of the @ycombinator Partners. We interviewed over 100 attendees about their experience.
Here's a recap 👇️
1️⃣ The #1 complaint = people

Go ahead and ask yourself, really ask yourself, if all these confident predictors are wrong about jobs and not training radiologists anymore and fast take offs, what else are they wrong about?
Then ask yourself: How is it that someone can be so smart in one area of life and/or know so much about AI and/or be really successful in business and still get the future so wrong again and again?
Because predicting the future is a totally different skill. Most people absolutely suck at it but are delusional enough to think they're fantastic at it.
What you know about a subject has exactly zero to do with how well you predict the future about that subject.
If you're a bridge engineer and know everything about building bridges, it doesn't mean you know anything about how bridges will effect society and it doesn't magically give you a crystal ball to predict the development of the car.
David Sacks: Narrative violation. WSJ today.

Great conversation this weekend at @ycombinator Startup School with @lutherlowe on tech policy, AI, the importance of policy frameworks to support startups & little tech, plus my new report Science: A New Golden Age.
Thank you @garrytan!

This prompt is actually magical @mattshumer_
Mark Essien: After working for about two hours, this is what came out. Older screenshot attached.

Sam on the Hugging Face incident:
“This is the first security incident that I have felt very viscerally. I've been a little surprised that more people don't feel it so viscerally.
We paused training. We have to figure out how to secure our sandboxing in a world of multiple zero days being chained together.
We may have to pace the rate of AI development to give ourselves enough time for society to harden around these new capability levels.
And trying to figure out how we do that in a way that does not feel like regulatory capture for anyone and also does not feel like collusion among the frontier labs.”
Patrick OShaughnessy: My conversation with Sam Altman (@sama), CEO of OpenAI.
We discuss:
- Kimi, distillation, and open source
- OpenAI's compute bets
- The Hugging Face incident
- What happens after AGI
- Raising kids in an age of abundant intelligence
- And much more
Enjoy!
TIMESTAMPS
0:00 Intro

http://x.com/i/article/2081790458259517441

Sam on Kimi, distillation, and open source:
“I have always assumed that there are going to be great cheap models in the world, and we better be the greatest and the cheapest.
You get a better deal today, at least at a particular latency, using OpenAI's models than Kimi. We distill our own models. That’s how we make smaller, cheaper models. That’s a very good thing to do.
There will clearly be an important place for open source models in the world and people that will want their own weights.
We have so much usage of our models that we do not need to be a gigantically high-margin business to be able to afford model training.
I would rather people not distill from us, for sure. But this is not in my top ten list of worries.”
Patrick OShaughnessy: My conversation with Sam Altman (@sama), CEO of OpenAI.
We discuss:
- Kimi, distillation, and open source
- OpenAI's compute bets
- The Hugging Face incident
- What happens after AGI
- Raising kids in an age of abundant intelligence
- And much more
Enjoy!
TIMESTAMPS
0:00 Intro

introducing greptile security agent
it combines static analysis, sca, and ai agents to efficiently catch security flaws before merge.
why did we build this?
- coding agents have caused the frequency of security incidents to explode
- neither agents nor deterministic methods alone are sufficient for detecting vulnerabilities
- security review is best done in conjunction with regular code review because many vulns are really just logical issues with security consequences
security review is turned on by default for free for all greptile users.

One of my most defining experiences was 15 years ago watching two ppl with roughly same followers. One HAD to tweet the zinger, HAD to argue with everyone. The other mostly stayed positive and built.
The former ended up having a middling career. The latter a monster career.
Blaire Pang: what annoys @sama about running YC 🤣😭:

One of my most defining experiences was 15 years ago watching two ppl with roughly same followers. One HAD to tweet the zinger, HAD to argue with everyone. The other mostly stayed positive and built.
The former ended up having a middling career. The latter a monster career.
Blaire Pang: what annoys @sama about running YC 🤣😭:

vercel sandbox is just bliss,

Sam Altman: “Concentration of power with AI is a terrifying thing.”
"A lot of the talk about safety concerns is well-founded, and then a lot of it is about people that just really, even if it's slightly subconscious, wanna concentrate power.
I don't think anyone should want to live in a world of AI overlords, where someone is making decisions for all of the future, and in exchange for a cure for cancer,  which obviously is a wonderful thing, we collectively cede all agency.
I am terrified of a world where the very real fears of AI are used as a way to say, “Only this small group of people can have it because it's too dangerous, and only they understand it. But don't worry, they're gonna make the right decisions for all of us.”
I was a child of the internet. There were no rules. It was amazing, and I think it's critical we preserve that spirit with AI."
Patrick OShaughnessy: My conversation with Sam Altman (@sama), CEO of OpenAI.
We discuss:
- Kimi, distillation, and open source
- OpenAI's compute bets
- The Hugging Face incident
- What happens after AGI
- Raising kids in an age of abundant intelligence
- And much more
Enjoy!
TIMESTAMPS
0:00 Intro

http://x.com/i/article/2082137754788646912

Deeply proud of what the SceniX team has built and excited to take it forward at @theworldlabs.
Our real-to-sim-to-real platform enables scalable robot training and evaluation in worlds aligned with reality. This is an important milestone, and only the beginning. 🌎🤖
Fei-Fei Li: When SceniX joined World Labs, we said spatial intelligence was never only about perceiving and generating virtual and physical worlds, but also interacting with them. Today, we’re sharing early results from that vision: building worlds that train robots. 🌎🤖↓

To scale robot intelligence, we must scale the worlds in which robots learn. 🤖🌎
Fei-Fei Li: When SceniX joined World Labs, we said spatial intelligence was never only about perceiving and generating virtual and physical worlds, but also interacting with them. Today, we’re sharing early results from that vision: building worlds that train robots. 🌎🤖↓

Insurance Commissioner is a technical actuarial role.
Unfortunately it’s treated as a backup job by second tier politicians.
Jane was an SF Supe who blocked housing and tried to defund police. She has no insurance knowledge and is profoundly unqualified. Do not vote for her.
Ashley Zavala: Jane Kim is running to be California's next Insurance Commissioner amid the state's insurance crisis.
I asked her repeatedly what experience she has with the industry.
She wouldn't directly answer the question and instead explained what she thinks IC is supposed to do:

World Labs CEO Dr. Fei-Fei Li & SceniX Co-Founder Yunzhu Li on the data bottleneck in robotics and how world models help:
"The lack of data in training, the lack of data in evaluation, this is very, very different from language models, where data is abundant on the internet."
"And we know that in order for robotics to work, we have to somehow unlock the power of scaling law. But where does that come from?"
"It's a profound problem that everybody's battling with in robotics."
"We see a lot of unlock in being able to do this whole process through the modeling of the environments."
"Being able to create these digital worlds... that's just going to unlock so much more potential for being able to replace all the costly and the unsafe data in the real environments with the data generated from the worlds for robots to be able to do scalable learning and evaluations."
@drfeifei @YunzhuLiYZ @martin_casado
Fei-Fei Li: When SceniX joined World Labs, we said spatial intelligence was never only about perceiving and generating virtual and physical worlds, but also interacting with them. Today, we’re sharing early results from that vision: building worlds that train robots. 🌎🤖↓

yc companies doing the coolest stuff these days
Synapse Semiconductor: We put a neural network inside the camera pixel.
Introducing RETINA: a vision sensor that runs AI where the light lands

I wrote about why we believe the future is for everyone. More coming about a positive vision for a world with superintelligence soon.

Re Give it a read: https://www.wsj.com/opinion/the-ai-future-is-for-everyone-a0c24e20?mod=hp_opin_pos_2

Tony Xu on Why DoorDash Waited 7 Years Before Category Two⁣
⁣
"You kind of as an entrepreneur have to take the greedy algorithm. You have to keep going all the way. You see if there's a lot of winning swing, why would you get off? So then your job is basically just say no. What's the opportunity cost?" — @tonyxu

Did not expect to get a second customer, one day after my first ever sale. We're officially on a 2-day sales streak. Built on Replit. Let's go!
BetaBloom | Find your first users: Just achieved my first ever revenue as an entrepreneur. I felt 3 seconds of joy this morning, before the oncoming obstacles flooded my mind. I'm excited to take these on. I feel good, knowing that I have @Replit in my toolkit.

Introducing the Bezos List: The Top 20 People Ranked by Wealth Created for Others
In a resurfaced clip that went viral recently, @JeffBezos called for “a list where they rank people by how much wealth they’ve created for other people,” not just for themselves. So we made it. 👇

Laser focused on reducing the cost of living, from day one.

What are the two most important principles for FDE?
1. always be scoping
2. scale with tokens

I'm sitting next to a YC partner (@borisjabes) who just spent the last 30 minutes debating with himself whether Hop Aero is /technically/ an ICBM
Current verdict: yes
YIMBYLAND:

Huge hats off to @mattshumer_! I can't believe his prompt allowed me to one-shot this game on Claude Code!
Matt Shumer: Prompt:
I want you to build a first-person shooter at the level of the most recent Call of Duty games. It should be utterly perfect, visually beautiful, with every single thing done at AAA quality—from textures to physics to anything you could think of.
Fan out sub-agents and

http://x.com/i/article/2082236458019278848

Was very curious to see if it was just hype, but tried @mattshumer_'s gauntlet style prompt and after 16 hours it produced this.

Model selectors can be overwhelming. Not on Replit, though.
We give you access only to the best models at any point in time, with recommended versions and effort levels based on our internal benchmarks.
We did all the work for you, so you can focus on Outcome-Maxxing.
Michele Catasta: http://x.com/i/article/2082236458019278848