The argument from imminent superhuman AI and an R&D feedback loop
The argument
Capability trends, if soundly extrapolated, put superhuman coding by 2027. Superhuman coders automate AI research, which yields superhuman AI across cognitive work within a year or two. Such systems, deployed at scale, automate most cognitive labour and speed up science and robotics, so that the remaining years of the decade bring economic, military and social change larger than the Industrial Revolution's.
Does it work?
- Does it follow?
- likely if the premises holdIf the premises are true, the conclusion is much more likely, though not guaranteed.constitution →
- Do the premises hold?
- 1 doubtful1 inconclusive
Granting its premises, including that superhuman AI deployed at scale would quickly automate most cognitive labour, the conclusion becomes much more likely, though the step from capability to economy-wide impact within a few years remains an unstated assumption about adoption speed. The argument lives on linked claim, currently assessed doubtful, and on whether trend extrapolation is a sound basis for short timelines, which is unsettled.
Premises
- DoubtfulThe evidence weighs against the claim, but the examination is incomplete or the evidence is indirect.constitution →AI systems will attain superhuman coding ability by 2027.
- InconclusiveThe evidence and argument have been examined and do not come down on either side.constitution →Extrapolating exponential AI capability trends is a sound basis for short AI timelines.
Where it is made
Passages where sources make, criticize or mention this argument. They record where it appears, not how strong it is, and none has precedence over another.
No source is recorded as making it.