AGI·LEDGER
The Ledger
@Ilya Sutskever · 2026-06-27Independent

Superintelligence is coming; build it safely, as the only goal

What form?
Other
How built?
Bigger models
Made of?
Doesn't matter
What for?
Helper
Who controls?
Locked down
When?
2030s
P(catastrophe)

My take

Superintelligence is achievable and approaching. The only responsible way to build it is with safety as the singular priority — not one product line competing with others, but the entire mission, insulated from short-term commercial pressure.

Why

The most capable systems will be the most consequential in history, so safety cannot be an afterthought. A focused effort that advances capability and safety together, with no distractions, is the way to get it right.

What would change my mind

Evidence that safety and capability fundamentally cannot be co-developed — in which case the responsible move would shift toward halting rather than building.

Summarized from public statements · source ↗
#sourced#superintelligence#safety-first