Sam Altman's 10 Rules for the AI Era
OpenAI is signing 20-year infrastructure contracts while most founders are still debating whether to use AI at all.
I watched the full two-hour Sam Altman and Patrick Collison conversation at Stripe Sessions so you do not have to.
Here are the 10 that matter.
1. The One Call That Got AI Out of the Lab and Into Your Hands
Iterative public deployment was the contrarian call. The consensus inside the field was against it.
“It is extremely important that we avoid that kind of power concentration and that we build this for the world.”
The dominant view before ChatGPT launched: keep it inside the lab. Control who touches it. Frame it as safety. The actual structure was power concentrated in fewer hands than any prior technology in history.
Altman rejected it on one specific ground. Lock the technology inside a research cohort and you have created exactly the thing you claimed to be preventing.
Every useful AI product built by founders today is a direct downstream consequence of this single call. It was a closer decision than most people realize.
2. 8 Months of Collective Psychosis: What Real Technological Belief Costs
Most founders claim they believed before others did. Altman describes what belief without any external feedback actually feels like.
“We’d walk the halls sometimes. Are we engaging in collective psychosis? Have we gotten totally whipped each other into this frenzy? And there was no feedback to keep us in check or sane from the outside world.”
OpenAI finished training GPT-4 eight months before launch. Everyone inside used it daily. Zero external signal. Zero market validation. Just conviction and a question that had no good answer yet.
The craziest OpenAI story Altman chose to tell was not the board crisis. It was eight months of living inside a reality the rest of the world had no access to.
This is what genuine technological discontinuity feels like from the inside. Not triumphant. Vertiginous.
Founders who have experienced a real threshold moment will recognize this immediately. Prepare for it before you are inside it.
3. The Idea-Only Founder Goes From Punchline to Term Sheet
The most important founding filter just shifted for the first time in 20 years.
“There was a time when we used to make fun of the idea guy... it would be like saying, I have a great idea for a song. I just need that guy with the guitar to make it for me.”
For two decades, no technical co-founder meant no deal. Altman now explicitly wants to fund founders who understand their users deeply and cannot write a single line of code. The model is the guitar player.
What still filters:
▫️ Deep user understanding. Still the first non-negotiable.
▫️ Problem clarity. You have to know exactly what you are solving and why.
▫️ Distribution instinct. The build is easier. Getting people to use it is not.
A large cohort of operators, clinicians, lawyers, and domain experts who were previously unfundable just became fundable. Most of them do not know it yet. That is an arbitrage with a short window. Study what top VCs look for in 2026 before that window closes.
4. The Shopify CEO Playbook: Why Tobi Lutke's Personal Rule Became the New Executive Standard
“It was not like a token leaderboard. It was not some other kind of gamified hackable thing. It was just like the CEO of the company said, we are now going to put AI into everything we do.”
Tobi Lutke was the first CEO Altman saw do this correctly. No incentive programs. No gamified pilots. Lutke built AI automations himself, held the standard personally, and told the team: this is how we operate now.
OpenAI is now testing a program where they embed a full-time employee directly with a CEO to automate that CEO’s own workflows first. The fractal effect follows naturally. The behavior propagates when the leader goes first. It stalls when the leader delegates.
Build one workflow automation yourself this week. Your team is watching to see if you mean it before they believe it.
5. A Decade of Science in 12 Months: The Bet No Portfolio Is Pricing Correctly
This is the most consequential claim Altman made in two hours. It received the least discussion in the room.
“If we can start doing a decade of science of what it would have taken us in the old world in a year, the compounding effect there and what we’ll be able to do and discover will just be extremely great.”
Three areas Altman flags specifically:
▫️ Biology at the foundation model level. The analogy to language models is direct. The training data exists. The architecture already works on protein structure.
▫️ Complex disease cures for conditions that have resisted drug development for decades. AI hypothesis generation changes the search space entirely.
▫️ Materials science. Altman calls this massively underrated. New materials unlock physical limits in energy, computation, and manufacturing simultaneously.
Most investor attention is on apps and agents. The civilizational leverage is in science. Most portfolios are making only one of the two bets on the table.
6. The Next Category After Coding Is Every Hour You Have Never Thought to Audit
“The degree to which most people will realize they can sit back and watch an AI do most of their drudgery is going to surprise people.”
The unlock is not a new industry. It is the sudden recognition of how much time people spend on tasks that should not exist in their workday at all.
Copying between apps. Formatting documents. Responding to messages that were themselves routine. These tasks are invisible until the moment you have something better to do with that time.
Altman reports this firsthand. Automating his own drudgery did not just save hours. It changed the subjective quality of his work. The market has not priced this correctly because the category stays invisible until it disappears.
If you want to see what this looks like in practice, the Claude Code chief of staff system is the closest real-world implementation of what Altman is describing.
7. OpenAI Chose to Be Stripe, Not Google
“I’d be happy for us to be a forever low margin as long as we can be huge and growing fast business, and I would like us to supply kind of an intelligence meter.”
The model Altman admires is Stripe. Pure infrastructure. Revenue aligned with customer success. Stripe gets bigger when its customers get bigger. That is the structure OpenAI wants.
Two mechanics make this stable:
▫️ Switching costs in AI tools are lower than most people assume. Lock-in through capability is the only defensible position long-term.
▫️ As models improve across the industry, margins compress for everyone. Volume and customer alignment become the durable position.
Founders building on OpenAI should take the stated intent seriously. An infrastructure provider that competes with its customers destroys the trust that makes the infrastructure worth using.
8. The One Management Skill Behind Every OpenAI Breakthrough (A Biographer Named It. Altman Could Not Predict It.)
Altman did not identify his own most important contribution. Someone writing a book on OpenAI did.
Altman did not identify his own most important contribution. Someone writing a book on OpenAI did.
“You figured out how to get a lot of people who all thought they were the only capable or most capable person, and everything had to go their way, to work together long enough to figure out the breakthroughs. That was the magic of OpenAI.”
3 conditions that made it work:
▫️ Concentrated resources on one direction. Not spread thin across competing bets.
▫️ Shared conviction on the mechanism. Everyone believed scaling laws were real before the market confirmed it.
▫️ A goal worth the cost. The mission made the pain of working together functional instead of fatal.
For founders managing elite technical talent, the forcing function is shared conviction strong enough that the pain of working together is worth it. Lose the belief and the pain ends the company.
9. OpenAI Is Signing 20-Year Contracts: The Demand Logic That Makes the Math Work
“In some sense, I think demand for intelligence at a low enough price is effectively uncapped.”
OpenAI is signing 20-year power and land agreements. Efficiency gains on a per-GPU basis are ahead of internal projections. Revenue is ramping to match the capital commitment.
The demand-side logic is what makes it rational. Every time the price of intelligence drops meaningfully, demand rises more than proportionally. Use cases appear that did not exist at the previous price point. That dynamic runs opposite to most industries, where lower prices compress the total opportunity.
The contracts are signed. The land is committed. The only remaining question is what you build on top of the infrastructure that is already arriving. Study what VCs are betting on in AI to understand where the capital is already pointed.
10. Your AI Company Passes Exactly One Test. Run Every Product Decision Through It.
“You as a business want to be on the side of hoping that AI gets smarter.”
Run this test on every product decision and every investment thesis:
▫️ Pass: Your product gets meaningfully better every time a new model drops. Your users notice and stay.
▫️ Fail: A model capability improvement removes the reason your product exists. You were renting a gap.
▫️ Red flag: You are quietly hoping capability stalls. That is the clearest signal you are on the wrong side of the curve.
This framework resolves every “is this a real company” debate faster than any due diligence process. Run it on your existing portfolio today.
The Infrastructure Playbook
The through-line is simple. AI is a utility being built at a scale the world has no comparison point for.
For founders: Deep user understanding plus problem clarity is now a fundable combination without a technical co-founder. The window is open now. Get in before the market prices the shift. Sharpen your pitch deck around user insight, not technical architecture.
For investors: The science acceleration bet is underpriced relative to apps and agents. Biology foundation models, automated labs, and materials science are the positions most portfolios are missing while they compete for the same agent infrastructure deals. Rebalance before the pricing catches up. Run your portfolio through the VC return model with this lens.
For executives and operators: The Shopify playbook is simple. The CEO goes first, personally, with no leaderboard and no pilot program. Build something yourself this week using Claude or ChatGPT. Your team needs to see you mean it before they believe it.
Build on the side of hoping intelligence increases. Every other position is a slow retreat.
The contracts are already signed. The only open question is which side of the curve you are on when the price drops.
If this breakdown saved you an hour, share it with one founder or investor who needs to see it. They will thank you later.
The infrastructure is being built whether you position for it or not.
The contracts are already signed.
The only open question is which side of the curve you are on when the price drops.
Full interview:
If this breakdown saved you an hour, share it with one founder or investor who needs to see it. They will thank you later.