Almost three years ago we released Stable Diffusion, a generative AI that changed the creative landscape.

Before its release, I told the team that we were about to change the world by raising the ability of anyone, anywhere, to create. Most people don’t believe they can because they are told they can’t, but all of us have that ability. 

We released this technology open-source, followed by state-of-the-art models of other types, resulting in over 300 million downloads.

No permission needed, just a net increase in human creative access.

Since then we have had an explosion of tools and their capabilities, with the latest big leap being reasoning models with OpenAI’s O1 model released just 7 months ago.

This last week we saw these models achieve gold level in the International Maths Olympiad, and we are seeing benchmark after benchmark fall, with the recent Grok 4 release scoring 100% in the AIME 2025 test.

The potential is huge, but so is the threat as, for the first time, we are about to enter an age of intelligence abundance where we are no longer the smartest entities on the planet.

To summarise from a recent tweet:

Society will inevitably change and this is before considering the dangers of super intelligence, which are profound and potentially existential, especially on our current path.

From problems to solutions

I stepped down as CEO of Stability AI last March to not focus on the problem, which was inevitable, but what a solution would look like.

Who builds and owns the AI that teaches your kids? That manages your healthcare? That governs your nation?

We may all be watched over by machines of loving grace, but whose grace is that?

To figure out a solution, we went back to first principles over the last year and built a full, comprehensive framework we call Symbioism.

This starts from first principles and builds up a full framework of philosophy, economics, social contracts, and technology.

Over the coming weeks, we will be releasing the full stack including the logical and mathematical proofs behind it of how we can build a system that is aligned with us all and allows us to not only survive, but thrive.

Today we release our Master Plan (the what), the Whitepaper of the Intelligent Internet (the how), and an overview of Symbioism (the why).

The Master Plan

Right now there is a huge demand for high quality intelligence and high quality digital assets.

Just as Bitcoin and Ethereum became the highest marginal dollar of demand for GPUs in the last cycle, we have designed Foundation Coin, a variant of Bitcoin, to become the highest marginal dollar of demand for high performance GPUs in this cycle.

The principle is simple: Proof of Benefit, where every coin sale as it is mined goes to, not only secure the network, but provide compute for Open Intelligence by giving Universal AI to everyone to clusters for Cancer, Education, and more.

This creates a flywheel where we can build the fully open datasets, models, agents, and systems (II-Commons, II-Medical, II-Agent, Common Ground) for the most important things in life and make them available to all as a public good.

There is no waste and this taps into the trillions of dollars of build-out we see everywhere, but will, if successful, drive as much of that as possible to the benefit of humanity.

No coin is lost, no Lambos or rugs, just building the best digital asset in the world by helping everyone.

The infrastructure for this can then become the credibly neutral, aligned AI infrastructure for us all and the very nature and genesis of money can go from debt to benefit (more on that later).

Plus your mom can be proud of you buying a digital asset where the proceeds went to helping people, not just making others richer.

The Whitepaper

We go deeper into the mechanics of this in our whitepaper, which marries the state-of-the-art datasets, models, agents, and systems we have built with a new ledger. This ledger is secure, fast, and ideally suited for billions of agents that help people, while still being 99% Bitcoin code.

Blockchains were not just about decentralisation and an open-system that anyone could rug and take advantage of people.

The goal is a credibly neutral, can’t-be-evil system that is coordinated, verifiable, and resilient.

Exactly what we need for the infrastructure for our regulated services, from education to healthcare to finance.

The nodes in our system are high performance GPUs (Blackwells) in top-notch data centers run by national champions owned by the people of each nation.

This gives geographic and political distribution as well as the integration with every nation to deliver the Universal AI services for the things that matter.

Nobody will care more about your nation and culture than you, and we will gather the best and brightest to help manage this transition and help their people, and humanity as a whole.

With this high performance network and known operators tied together by more than classical consensus, you can have the fastest, most distributed and resilient blockchain in the world. For the things that really matter.

Entertainment and speculation can go elsewhere.

The system is designed to stack GPUs and the basis for a new economic system that we will release the details of in the coming weeks, built on intelligence, supported by intelligence and aligning intelligence constantly.

We will go in depth into the elements of this whitepaper soon, but we welcome feedback as we open it publicly.

We have been steadily building the full AI stack and will soon have fully open, state-of-the-art systems for every important sector to uplift us all. Contribute to the open source to make this future a reality, even faster.

Common Ground: A multi-agent collaborative environment that transforms chats into structured, professional workflows to enable deep work and conflict resolution between humans and AI teams.

II-Agent: A generalist, open-source intelligent assistant designed to autonomously execute complex tasks and streamline workflows across research, coding, and content creation.

II-Commons: An open-source platform for collaboratively building the large, auditable, and efficient knowledge bases that power transparent and optimized AI.

II-Medical: A compact, open-source 8-billion-parameter model fine-tuned to enable high-quality medical reasoning, decision support, and medical education for lightweight, safety-optimized deployments.

II-Researcher: An open-source framework for building powerful AI research agents capable of tackling complex, multi-step inquiries through features like real-time literature review and hypothesis generation.

II-Thought: A reasoning RL model and dataset featuring a 360k-sample, multi-domain curriculum and SOTA models designed to sharpen LLM reasoning quality and speed.

Our aims are simple - free, universal AI for everyone. We will build the stacks that can check every diagnosis in every country. That anyone can download to have the best teacher for free for their kids. To make it so the system represents you instead of resenting you.

Symbioism

To build this we had to go back to first principles and Symbioism is our comprehensive framework for this, starting with philosophy and moving from there to systems and implementation.

In the coming period we will share the full details of this, along with full books on the philosophy, economics and more that we need to navigate this new reality.

These are not just incremental upgrades, but new systems that we believe are superior not only for the Intelligence Age, but for our current systems which are still stuck in the scarcity mindset of the 18th Century.

We have built this systematically and rigorously as an integrated system with mathematical backing and flexibility, but no system is perfect.

It is clear however that our current systems will not survive the coming of the Intelligence Age and we saw no other option than to take this approach.

The kernel has been shared today, but as we share the rest of what we have found and discovered, we know that together we can continue to improve it and gain insight from our contributors.

Time is running out for the biggest change ever and we hope that if we all build this, everyone thrives.

Superhuman AI is inevitable and we must create the right incentives, structures and flow to make it so it saves us all, not the very real and unpleasant future we face today.

The future is about to change. Let's change it together.

-Emad