Mapping positive post-AGI futures
Closing the gap between AI capabilities and shared benefits
Over the past years, the leaders of the three most prominent AI labs have each published their vision of how powerful AI could benefit humanity. Amodei envisions AI compressing a century of biological progress into a decade. Altman describes massive shared prosperity within a few years. Hassabis forecasts a scientific renaissance — all disease eliminated, radical abundance.
These are extraordinary outcomes, and the capabilities to enable them are increasingly within reach. The question is: what connects the capabilities to the outcomes? The path from “AI can do this” to “this actually goes well for everyone” is where some of the most important and least explored work lives. This workshop maps that path across four frontiers:
Science infrastructure: AI could accelerate a century of scientific progress — if the right datasets, sensing tools, and measurement capabilities exist to plug it into. What needs to be built alongside the AI?
Democratic voice: AI could enable genuinely new forms of collective decision-making, not just automate old ones. How do people shape what these powerful systems are used for?
Distribution: The prosperity AI generates could be shared more broadly than any technology in history — if the channels are designed for it. What does that design actually look like?
Vision: This might be the first generation that gets to actively design what flourishing looks like, not just react to what happens. What does a genuinely great future look like beyond “today but richer and healthier”?
Format
The morning focuses on what needs to be true — the conditions that unlock the extraordinary outcomes the visions describe. The afternoon focuses on what we’re even trying to build — the vision. The day closes with a plenary where participants vote on the directions surfaced, rating each on urgency, tractability, and neglectedness.
The output
A published gap map: what do the people closest to the capabilities think stands between powerful AI and broadly shared flourishing, and where is progress most needed? All participants will be credited as contributors.
We’re looking for ~50 researchers who want to think seriously about what it takes to get from powerful AI to broadly shared flourishing. Strong directions from the workshop can be connected to funders in our network.
