AI Model
Example: the chef
Helps AI think.
Understands questions, weighs information, and suggests possibilities.
Examples today: GPT-6 · Claude · Grok
- Understands a question
- Considers information
- Offers possibilities
Good for thinking.
OA · AI Research Lab
We built an operating system for AI to decide & act.
It researches, builds, fixes its own mistakes, and keeps going.
A person’s workday at a computer.Done by software!Without being asked!
START HERE
In Plain Terms
Three names you will see on this site, and what each one is.
GunnerThe architecture
The operating system we built, where AI does a person’s workday at a computer.
OkrenThe product
The product that keeps a person in charge when AI agents move money.
OAThe research lab
The one that thinks furthest ahead: where AI is going, and what people will need when it gets there. The AI research lab behind both.
WHAT GUNNER IS
Models Think. Agents Work.
Gunner Is an Operating System Modeled on How a Person Works.
A model is the chef. An agent is a cook working one station. Gunner is the one who runs the whole place: it takes the orders, remembers every regular, decides what happens next, and keeps the night going.
In real terms: a stateful Linux runtime. Declarative, Prolog-style logic for reasoning. Probabilistic models for judgment. C++ and Python for execution.
Gunner the whole system
Agent
Example: a cook at one station
Helps AI do a job.
Uses tools and takes steps toward a goal.
Agent examples today: Dots · Grok Bot · Meta Muse
- Takes on a task
- Uses tools
- Works through steps
Good for doing.
Gunner
Example: the person running the restaurant
Runs the whole job, like a person would.
Keeps the kitchen, the stations, the orders and the changes of the evening working as one place.
What it runs: the whole operation, whichever chef or cooks are in it.
- Keeps the night’s goal in view
- Knows what is happening at every station
- Notices when something changes
- Re-plans the next move
Good for running the whole place.
What Gets Better, and What Stays.
Models
Models get better (your chefs get better).
Agents
Agents get more capable (your cooks get more capable).
Gunner
Gunner is the manager who runs the place every night and remembers everyone.
Frontier labs are increasingly building operating systems around their own intelligence. Gunner OS starts from the opposite direction: the operating system is the lasting part, and models are intelligence resources inside it.
PICTURE THIS
You give it a goal. It does the workday.
An imagined Monday, step by step.
- Monday, 9:00
You give it a goal: launch a small app.
- By lunch
It has researched the market, read the news, written the code and put it live.
- 2:00 pm
Something breaks. It reads the error, looks up the fix, applies it and carries on.
- End of day
You find out when it is done.
An imagined example of what Gunner is designed to do.
It does not wait to be told. It keeps going.
THE BIGGER IDEA
The Cognitive Revolution, Repeating.
Homo sapiens evolved by sharing information, and it changed everything. AI is going through that same shift right now.
01 · HUMANS LEARNED
Humanity’s Cognitive Revolution—the leap in how we learned together—came from our ability to remember what happened, add context, and build on what we learned.
02 · WE BUILT RULES
As we became capable of doing more together, we built rules and independent checks to govern what we could do.
03 · AI IS LEARNING
We are entering AI's Cognitive Revolution—teaching AI to remember what happened, add context, and build on what it learns.
04 · HUMANS KEEP AUTHORITY
As AI becomes capable of doing more on our behalf, people stay in charge: they set the limits and decide what it is allowed to do.
The more AI can do, the more it matters that people stay in charge.
LOOK AROUND · OA SIGNALS
See how experts describe a multi-agent future for AI agents.
These are the voices describing a future where many AI agents work at once. Okren is our framework for supervising them, so a person stays in charge.
EMILY GLASSBERG SANDS · STRIPE · NATE B. JONESWhat must an AI remember?
What happens to permission and trust when AI becomes a buyer?
SATYA NADELLA · MICROSOFT · YAHOO FINANCEWhat changes after an action?
Microsoft describes supervising agents as they operate.
VLAD TENEV · ROBINHOOD · SOURCES WITH ALEX HEATHIt Won’t Go Rogue
Separate account. Approval by default. Autonomy by deliberate choice.
Source interviews, with OA’s reading of why they matter.
Why Now.
The future of AI.
- The shift
Agents are the new customers.
- Where this is going
Soon you will have an AI you simply talk to. It listens to everything you do, and it becomes your eyes, your hands and your voice, acting on your behalf.
- What we built
We built the operating system for that: the one that lets an AI do a person’s work without being asked. There will be a new Microsoft of AI agents. We are building in that direction, and doing our part to make it something people can trust.
- What has to stay true
The more it can do, the more it matters that people stay in charge. That is Okren.
WHAT COMES NEXT
Gunner operates. Okren supervises.
AI this fast needs a second system, apart from the AI, that checks every move before it becomes real, at AI speed. That is Okren: the glass walls and the one window out, where a person sees every plate before it leaves.