The AGI 2.0 Manifesto

Humanity stands at the threshold of the Intelligence Age: a time when artificial general intelligence(AGI) could become the most powerful instrument of progress ever created.

It promises to accelerate discovery, amplify creativity and drive sustainable prosperity across every field of human endeavour. Yet the path we are currently on threatens to undermine that promise.

The prevailing AGI model built on ever-larger architectures, immense computational cost and tightly held proprietary infrastructure is not a path to shared intelligence but to concentrated power. It risks creating an era in which a handful of corporations become the sole architects, gatekeepers and beneficiaries of the world’s intelligence.

If AGI is to serve humanity rather than dominate it, we must change course.

AGI 2.0 represents that course correction: a shift from model-centric systems that depend on scale to context-centric intelligence that learns continuously, efficiently and in realtime. It seeks to ensure that the ability to build, own, and govern intelligent systems is not monopolised but made broadly and sustainably achievable.

This manifesto establishes the foundational principles of AGI 2.0 to secure a future where intelligence itself remains a public good, not a private empire.

Article I: The right to sustainable intelligence creation

The future of AGI cannot depend on a handful of entities capable of absorbing billion-dollar infrastructure costs. When the price of building and maintaining intelligence becomes unsustainable, innovation collapses into monopoly.

AGI 2.0 must therefore enable cost-efficient model learning during inference, systems that continue to learn and adapt while in use without requiring vast retraining cycles or prohibitive compute resources.

By drastically reducing infrastructure demands, we ensure that the power to create and operate intelligence remains distributed, competitive and economically viable.Efficiency here is not about saving money, it is about preserving plurality in who can build the future of intelligence.

 Article II: The right to ownership and sovereignty

True empowerment requires true ownership.

Every creator has the right to fully possess the AI they build and the data that informs it, including control over where that data resides, how it is used and who can access it.

Intellectual property generated through private data or proprietary training belongs to its originator. Private intelligence should never become free fuel for another’s platform.

AGI 2.0 protectsAI sovereignty: the right to build, govern and profit from one’s own intelligence.

Article III: The right to adaptive and interoperable intelligence

AI should adapt to people and to each other, not the other way around.

The current generation of AI tools is built on static layers: wrappers, plug-ins and add-ons. These silos cannot communicate, cannot evolve together and cannot scale meaningfully without constant redevelopment. As their number grows, they create a technological bottleneck, one that demands ever-greater resources and fragments the very intelligence they aim to extend.

AGI 2.0must replace rigidity with dynamic, context-aware intelligence, with systems capable of reasoning across contexts, integrating seamlessly with one another and reshaping their interfaces in real time through natural language.

Such systems remove the need for endless re-engineering. This is how intelligence becomes infrastructure: fluid and continuously adaptive.

Conclusion

The first vision ofAGI, one that is model-centric, static and dependent on scale, has reached its natural limits. Its pursuit of ever-greater scale has delivered remarkable progress but unsustainable economics and growing concentration of power. AGI2.0 charts a different course: from scale to substance, from models to context, from centralisation to continuous learning.

If we get this right, intelligence itself will become infrastructure, a shared foundation for human advancement, not a proprietary asset. Only then will AGI fulfil its original promise to be the great equaliser of the Intelligence Age.

‍