What is superintelligence? And when might it arrive?

The definition

Where AGI is roughly human-level (able to do the work of, say, an AI researcher), superintelligence is far beyond the best humans across nearly every domain. It's the difference between "matches an expert" and "dwarfs all experts combined."

How it's supposed to arrive

Aschenbrenner's mechanism is the intelligence explosion: once AI reaches the level of a competent AI researcher (the AGI-2027 milestone), you point it at improving AI itself. Hundreds of thousands of automated researchers compress roughly a decade of progress into about a year — vaulting from AGI to superintelligence.

The predicted timeline — and current status

Superintelligence sits at the end of a dependency chain whose first links haven't fired: autonomous AI research is undemonstrated. That's why the verdict is Pending — the most speculative claim on the scorecard, and the most dependent on everything before it.

Frequently asked questions

AI vastly beyond the best humans across essentially all domains — the end-state of Aschenbrenner's argument. Where AGI is roughly human-level, superintelligence dwarfs all human experts.

Aschenbrenner predicted the early 2030s, via an intelligence explosion following AGI. As of mid-2026 the chain hasn't started (AGI-2027 is open, the intelligence explosion is pending), so the timeline is speculative.

Through an intelligence explosion: once AI can do AI research, it improves itself, compressing about a decade of progress into a year and vaulting from AGI to superintelligence.