An OpenAI insider just revealed why the AI race is moving so fast.
Daniel Kokotajlo, a former OpenAI researcher, says the race isn't just about building better AI.
It's about economic power.
Whoever builds the leading AI systems could shape the next global economy.
He also believes AI has the potential to dramatically improve humanity.
But if that power ends up concentrated in the hands of just a few companies, the consequences could be enormous.
That's why we built Action Model.
A community-owned AI model where the people helping train, grow, and build the ecosystem can earn a fractional stake in the value they help create.
AI is inevitable.
An AI future owned by only a handful of companies isn't.
Around 400 AI workers and students just protested outside OpenAI, Anthropic, and Google.
Not because they hate AI.
Because they're worried about jobs, power, and who will benefit from the trillions of dollars AI is expected to create.
Here's the reality.
You can't stop the AI race.
But you can change who benefits from it.
That's why we're building Action Model.
A community-owned AI model where the people helping train AI can earn a fractional share in the value they help create.
We'd rather see that value distributed across millions of ordinary people than concentrated in the hands of a few Big Tech companies and billionaires.
THEY ORDERED 50 NVIDIA RTX 4090s TO BUILD LOCAL B2B NODES AND SECURED A $32,000 CONTRACT
They set up the first ten cards on open frames with four EVGA power supplies and custom cooling fans
Renting this level of compute capacity on public clouds costs over $12,000 a month in recurring fees. Running continuous automated database and reasoning loops on remote servers kills software agency margins
This local setup deploys five RTX 3090s inside an octo server rack and four RTX 4090s in dedicated client cases. By hosting Hermes Agent inside local Docker containers, customer files remain secure and HIPAA compliant
The physical hardware stack pays itself off in months by eliminating cloud compute bills. It turns corporate AI into an owned asset while generating a stable $1,500 monthly support retainer
bookmark this and read the full multi-node deployment playbook below ↓
When the conversion lands, the result is mostly decided by what you did in the months before.
Most walk in cold and ask if the teller will exchange their notes. That question alone flags you as retail.
The free guide is the proven playbook. No card needed.
2026 is the year AI automation goes mainstream.
But there's one problem.
Every automation creates value, yet almost all of that value flows to Big Tech and billion-dollar companies.
That's where @Actionist_ai changes the game.
Powered by Action Model, Actionist is an AI automation platform where the people helping train and grow the AI can share in the value it creates, not just shareholders and billionaires.
Automation is inevitable.
Who benefits from it isn't.
Join our Discord to be among the first to experience Actionist before anyone else.
Unpopular opinion:
People don't hate AI.
They hate watching billionaires get richer from AI while they get replaced by it.
Technology should free humanity.
Not just make the richest people even richer.
That's the conversation we should be having.
Snatch valuable exhibits from Los Santos’ premier museum in a blockbuster new score: The Kortz Center Heist, coming to GTA Online on July 14.
Enroll in the Fine Art Collector program now and reach Elitist status to earn a GTA$1,000,000 discount on the Art Studio and more.
Claude Fable 5 can run unattended all night, spawn its own sub-agents, and finish a week of work before you wake up.
Most people use it like a chatbot. Here's how to turn it into an Agent OS.
Part 1 — The engine.
Set the model string to fable-5 and cap max_tokens at 64K so thinking plus output never blows the budget. Keep effort on high, not max max only pays off on glue-heavy agent chains.
The trap almost everyone hits: refusals come back as HTTP 200. Your scripts must check stop_reason == "refusal", not just the exit code. Miss this and your pipeline silently swallows dead runs for weeks.
Part 2 — The loop.
Create agent_loop.sh on a cron job that fires every morning at 6 AM.
A cheap model scans your project first and answers 1 question: did anything change or need attention? Only if yes does Fable 5 wake up as the manager. The manager never touches code it writes exact task specs. A worker instance does the work. A fresh third instance inspects it with zero context of who wrote it.
3 roles, 3 separate contexts. That separation is where the reliability comes from.
Part 3 — The report card.
Every finished task gets logged pass or fail, and every task type earns its own grade over time. Once a type hits 20 passes at 95%, the system earns full autonomy on it. The grade drops below 90% privilege revoked, you get an email.
You never guess whether to trust it. The numbers decide.
Part 4 — The ghost directory.
For everything the system builds, it writes 1 tiny test proving it still works, then runs the whole suite daily. The thing you shipped 3 weeks ago breaks quietly it gets flagged the same morning, with exactly what changed since the last green run.
Nothing you build ever rots without you knowing.
I wrote a step-by-step guide with the full agent_loop.sh, the report card schema, and 6 more builds on top of this. Like this post and comment "AGENT" I'll send it to you.
Part 2 covers the rest of the OS: memory, budgets, and self-healing.
You're not hiring an assistant. You're hiring a night shift.
The U.S. government is reportedly deciding who gets access to GPT-5.6.
OpenAI plans to launch the model in a limited preview, with customer access reportedly approved on a case-by-case basis by the federal government.
Think about how quickly AI is becoming strategic infrastructure.
The race is no longer just about building the best model.
It's about who controls access to the most powerful AI systems in the world.
As AI becomes more central to our economy and society, the conversation can't just be about capability.
It also has to be about ownership, access, and who gets a voice in shaping the future.
That's why we believe AI should benefit the people, not just a handful of powerful institutions.
The 500 richest people on Earth got $366 billion richer last week.
The biggest weekly wealth increase ever recorded.
This is what an economy designed to concentrate wealth looks like.
And with AI set to transform every industry, the gap is only going to widen.
Unless we choose a different path.
One where the value created by AI flows back to the people helping build it.
A community-owned AI alternative.
If we're all training AI, we should all benefit off the value it creates.
Which is exactly why we built Action Model.
Action Model is built on a simple idea:
If AI is being built using human data...
Humans should own part of it.
Here's the ecosystem in 30 seconds:
• Install the browser extension
• Train a community-owned AI model through your online actions
• Earn points for contributing data and completing specific ActionFi tasks
• Invite others and help grow the network
• Unlock early access to AI products
• The model gets smarter as the community grows
• The people helping build it share in the value created
That's it.
Instead of training AI for Big Tech for free...
Help build a model owned by the people.
The Action Model Discord is the heart of our community.
Inside our discord you'll get:
• Extra points and reward opportunities
• Exclusive updates & early access to our products
• Community game nights
• Cross-community competitions
• Team calls and exclusive AMAs
• Meet like-minded people who care about the future of AI
The link is in our bio. Don't say we didn't tell you early.
In 10 years, people will ask:
"What did you do when AI started changing everything?"
Most people will say:
"Nothing."
A small group will say:
"I helped build it and own it."
The future is being decided right now.
Choose which side of history you're on.
Epoch 6 is live 👾
While most people are watching the AI revolution happen...
Our community is helping build it.
Train AI.
Earn rewards.
Grow your share.
The future won't be owned by those who watched.
It will be owned by those who showed up and paticipated.