Recursive Superintelligence Commits $400M to AWS Compute for Self-Improving AI Systems

Skynet Chance (+0.11%): A well-capitalized lab explicitly pursuing recursive self-improvement — the canonical intelligence-explosion pathway — with automated product development and minimal human headcount raises loss-of-control risk, and the article mentions no corresponding alignment or oversight commitments. Directing the bulk of funding into autonomous agent compute rather than human researchers reduces the human-in-the-loop checks that currently constrain capability escalation.

Skynet Date (-2 days): Hundreds of millions in dedicated compute plus co-developed bespoke infrastructure accelerates experimentation on self-improving systems, and leadership signaling far larger deals ahead implies a steepening ramp. A stated product timeline of months rather than years compresses the window between research and deployed autonomous systems.

AGI Progress (+0.03%): The deal is a capital and infrastructure commitment rather than a demonstrated capability result, but it materially resources a research direction — open-ended self-improvement — that many consider a plausible route past current scaling limits. The article itself notes the requirements for genuine RSI remain ambiguous, so progress is contingent rather than proven.

AGI Date (-1 days): Large guaranteed compute, AWS's willingness to build purpose-built infrastructure for this class of company, and the signal that it may draw other foundation-level labs to similar arrangements all shorten expected timelines. The promise of usable products within months suggests faster iteration cycles than a pure research posture would allow.