The people who defined AGI have, for the first time, formally mapped what comes after it. On June 10, 2026, DeepMind published a 57-page paper, "From AGI to ASI." It is led by Shane Legg (DeepMind co-founder) and Marcus Hutter (inventor of AIXI), and it lays out four pathways from AGI to artificial superintelligence (ASI). Yet even ASI, it argues, is bound by fundamental physical and computational limits. ASAP summarizes the roadmap from the primary source.

The creators of the AGI concept wrote it

"From AGI to ASI" is a 57-page paper written by the researchers who defined the concept of AGI. DeepMind published it on June 10, 2026, led by Shane Legg (DeepMind co-founder and Chief AGI Scientist) and his PhD supervisor Marcus Hutter (inventor of AIXI). It is the field's attempt to look one level ahead.

Four pathways to ASI

The paper's four pathways from AGI to ASI are scaling, paradigm shifts, recursive self-improvement, and multi-agent collectives. In detail, they are continued scaling of effective compute (which the authors estimate at about 10x a year), non-linear paradigm shifts in architecture, recursive self-improvement, and multi-agent collectives. The four are not mutually exclusive and can compound.

ASI is defined as "smarter than organizations"

The paper defines ASI as a system more capable than large organizations of humans, not one person or one team. It means cognition that exceeds an entire organization with its division of labor, specialized expertise, and collective memory. The benchmark is organizational intelligence, not a single genius.

There are still limits

The paper holds that even ASI is bound by fundamental physical and computational limits. The examples include the speed of light, thermodynamic limits on computation, complexity theory, and Godel's incompleteness results. It does not get infinitely smart; there are walls it cannot cross.

What it means: "after AGI" is now on the agenda

The paper shows the field's attention shifting from reaching AGI to what follows it. That the people who defined AGI have mapped the path to ASI signals a shift in the center of gravity of the debate. By naming hard limits, it also adds ballast to an overheated superintelligence discourse.

Wrap-up

DeepMind's "From AGI to ASI" is a 57-page roadmap in which the creators of AGI map the next stage. Its four pathways (scaling, paradigm shifts, recursive self-improvement, multi-agent collectives) and its hard limits are the core. "After AGI" has become a formal agenda for the first time.

Source: ASAP summary of DeepMind's "From AGI to ASI" (arXiv 2606.12683, June 10, 2026; Shane Legg and Marcus Hutter, 57 pages, four pathways to ASI, physical and computational limits).