Wes Roth · 2026-06-18 · notable
Wes Roth: Google's 'POST AGI' paper — DeepMind's AGI-to-ASI roadmap
Wes Roth breaks down 'From AGI to ASI', a Google DeepMind paper co-signed by Shane Legg, Marcus Hutter and Thore Graepel that maps four routes from human-level AI to superintelligence.
Wes Roth walks through 'From AGI to ASI', the new DeepMind report on what comes after human-level AI.
What is it?
A YouTube breakdown by creator Wes Roth of the new Google DeepMind paper 'From AGI to ASI' (arXiv 2606.12683), submitted June 10. The paper is co-signed by Shane Legg, Marcus Hutter, Thore Graepel, Tim Genewein and 10 others, and defines AGI as a system that performs at typical-human level across most cognitive tasks.
How does it work?
The DeepMind paper names four pathways from AGI to ASI: scaling compute, models and training data; discovering new architectures and paradigms; using AI to recursively accelerate AI research; and coordinating large groups of AI agents into collectively superhuman systems. Wes Roth walks each route, the assumptions behind it, and the timelines DeepMind sketches.
Why does it matter?
When DeepMind's Shane Legg, one of the field's longest-serving AGI researchers, co-authors a 57-page roadmap for what happens AFTER AGI, the framing shapes how labs, governments and investors think about post-AGI risk and policy through 2026.
Who is it for?
people tracking AGI safety, post-AGI policy and lab roadmaps
Try it
Read the paper at arxiv.org/abs/2606.12683