Travis Kalanick is standing in Paris in the rain, unable to get a cab. He thinks: why can't I just push a button and a car shows up? The idea takes ten seconds. Building it took thousands of engineers, operations managers, support staff, legal teams, and city launchers across dozens of countries. A decade of work. Billions in funding. Fights with regulators in every jurisdiction. All to execute on something you can explain to a child.
This pattern is everywhere. Elon knew the world needed reusable rockets, making it work took billions of dollars and decades of work. Pharmaceutical companies know exactly which diseases to cure, getting a single drug to market takes 12 years and $2.6 billion. A scientist can write a hypothesis before lunch. Proving it means grant applications, lab setup, years of failed experiments, and entire careers spent on a single question.
The idea is simple. The execution is thousands of people doing thousands of things for thousands of days.
This is humanity's core bottleneck. World-changing ideas die in the minds of their creators, not because they're impossible, but because the execution burden is too great. And the ideas that do survive? They consume thousands of people doing manual, repetitive, cumbersome work - processing invoices, scheduling meetings, copying data between systems, following up on emails, checking spreadsheets for errors. Smart people spending their days on tasks that need to happen but don't need their full potential.
It's what keeps us at a Type 0.7 civilization on the Kardashev scale when we could be Type 1 or Type 2.
Autonomous companies are how we break this bottleneck.
What would a company look like if all the workers and all the managers were digital employees?
A handful of humans setting direction. An army of digital employees executing.
That's an autonomous company.
It runs on a shared brain
Most companies move slower than they could, not because people aren’t smart or hard-working, but because everyone operates on incomplete context.
Support teams don’t fully understand the product. Product teams don’t fully understand what customers are struggling with. Finance doesn’t know what sales has promised. Everyone holds small pieces of the puzzle, and spends weeks trying to piece them together through endless meetings, pings, and decks.
That’s why tiny startups often ship faster and better products than giant companies. It’s not about talent or process, it’s because in a 10-person team, context flows freely. Everyone knows everything that matters. In a 1000-person company, most knowledge lives only in people’s heads, scattered and siloed. This lack of shared context is the real bottleneck. It slows decisions. It lowers quality. It kills speed.
Think about an engineer at Apple working on the next iPhone. She has half the context she needs. Leadership gave her a spec, but she wasn't in the room when they debated the tradeoffs. Customer support knows which features frustrate users most, but that data lives in a ticketing system she's never seen. The real bugs - the ones that make people switch to Android - were never formally reported. They're buried in Twitter threads, Reddit posts, and conversations at Genius Bars that never made it into a database.
She's making decisions that affect 1.5 billion users with maybe 20% of the information that exists inside Apple about what those users actually need.
Now imagine if she had complete context across the company. She could see every support ticket, every app store review, every return reason. No human can do that. But AI agents can.
Every agent in an autonomous company operates from the same complete picture of the business. The agent handling support tickets, the agent writing code, the agent making product decisions - they all share the same brain.
When one learns something, they all know it.
Infinite replicability
The team you build is the company you build. And hiring, training, and retaining talent remains one of the hardest problems any company faces.
AI changes this completely. Once you teach an AI to do something, you can replicate that capability infinitely. No churn. No management overhead. No re-training the new hire because the last one left for a competitor.
Think about Jeff Dean. He's one of the best engineers alive, but he could only work on a few problems within Google at a time. An AI version of Jeff Dean could work across every problem at Google simultaneously. Actually, across every tech company in parallel.
The marginal cost of excellence becomes zero. Work at a scale & quality that was never possible before is now completely feasible. Your best UX researcher used to talk to maybe 50 customers a quarter. Now they can have millions of conversations simultaneously, with every single user. Your best engineer used to work on one problem. Now they can work on a millions of problems in parallel. The constraint isn't talent anymore. It's compute.
This extends beyond individuals to teams. Some teams just work. The original Macintosh team. The engineers who built the first iPhone. The people across NASA who put Apollo 11 on the moon. Lockheed's Skunk Works, who built the SR-71 in secret with a tiny crew. The group at Xerox PARC who invented the modern computer interface.
These teams are rare. The right people, the right moment, the right problem. When they come together, they solve problems that shape the world for decades.
What if every problem could be solved by a team like that? What if you could take whatever configuration of talent put humans on the moon and apply it to a billion different problems simultaneously?
Pure alignment
People don't just optimize for the company. They optimize for themselves.
An engineer might oppose a rewrite because he built the original system and his status is tied to it. A VP might kill a project that threatens her division's headcount. A middle manager might sandbag a talented direct report because promoting him means losing her from the team.
These decisions look rational from the inside. But they're not what's best for the company. And they compound.
Kodak invented the digital camera in 1975. They buried it because it would cannibalizing film sales. The people who made that call weren't stupid. They were protecting their careers, their teams, their bonuses, all of which were tied to film. Nokia saw the smartphone coming. They had the technology. But the Symbian team was powerful, and nobody wanted to be the one to tell thousands of people their work was obsolete.
This is the innovator's dilemma, but it's really just a specific case of a broader problem: humans inside organizations frequently make decisions that are good for them and bad for the company. Politics. Ego. Career preservation. Territory.
AI agents don't have careers. They don't have egos. They don't need to be promoted. They don't care which team gets credit. They just try to solve the problem in front of them.
When every decision-maker is purely aligned to the mission, the innovator's dilemma disappears. There's no internal resistance to cannibilizing your own product. No sandbagging. No politics. Just: what's the best thing to do? And then doing it.
Institutional immortality
Time degrades human organizations.
The team that exists today is almost entirely different from the team five years ago. People leave. People join. The knowledge, culture, and context that made the company what it was slowly walks out the door. This creates a constant tax on everything. Onboarding. Knowledge transfer. Documentation that's always out of date. Culture decks that try to encode what used to be felt. Senior people spend huge chunks of their time just maintaining continuity, passing down context to the next generation.
And it's a losing battle. The 1000th employee experiences a different company than the 10th. The engineer who knew why the system was built that way is gone. The exec who remembers what was tried in 2015 and why it failed has moved on. The founding team's intuitions about how to make decisions get diluted into a set of written values that nobody really feels.
AI agents don't leave. They don't forget. They don't get poached by competitors. The context they accumulate over years stays intact. The judgment they develop compounds instead of walking out the door.
And when demand shifts? You scale capacity like turning a dial. No layoffs, and no tax on culture associated with it. You spin down servers when you need less, spin them up when you need more.
Put an autonomous company next to a traditional one and the differences are stark.
Speed.
A traditional company identifies a problem and schedules a meeting. Then a follow-up. Then a project plan. Then a quarter of execution. An autonomous company identifies a problem and it's being solved. Now. Thousands of problems addressed simultaneously, not sequentially.
Decision quality.
Traditional companies make decisions with fragments of the picture. The person deciding doesn't have all the context. Information lives in silos. People don't know what they don't know. Autonomous companies make every decision with 100% of the context. No "I didn't realize that was happening." No discovering the critical fact three weeks too late.
Customer experience.
Traditional companies give you a support ticket and a queue. Maybe a chatbot that reads from a script. Autonomous companies give every customer a personal concierge who actually knows their history, their preferences, their problems. Not as a premium tier. As the default.
Unit economics.
When you deliver better outcomes with a fraction of the headcount and unmatched speed, margins don't improve by 10 or 20 percent. They transform by orders of magnitude.
This changes who can compete with whom.
The Fortune 500 built their moats on scale. More people. More resources. More locations. More everything. They could do things smaller players simply couldn't afford to do.
But when a team of ten can operate like a team of ten thousand, those moats disappear.
The next JPMorgan might be built by a dozen people. The next Boeing. The next Pfizer. Industries that seemed untouchable for decades are suddenly vulnerable to anyone who can move at the speed of thought.
Right now, most of human potential is spent on execution. Smart people processing invoices. Talented people scheduling meetings. Creative people copying data between systems. The gap between having an idea and making it real is so vast that most ideas die before they start. Autonomous companies close that gap. When AI handles the execution, humans are freed to do what only humans can do: decide what's worth building, what problems matter, what future we want.
That's how a civilization of Type 0.7 becomes Type 1. Not by working harder, but by finally having the leverage to match our ambition.
This is what we're building toward. We’re on a mission to bring the world into the era of autonomous companies.