Chris Albon's thought experiment: how would AI history differ if another community found scaling?
chrisalbon · x · 2026-10-28
Chris Albon shares a thought experiment he finds increasingly useful: the discovery that scaling transformers yields dramatic capability gains came from a specific community (LessWrong/EA-adjacent) with particular values — but it didn't have to.
It could equally have been discovered by a military/defense/natsec community or a Wall Street finance community, each imprinting its own values on AI's trajectory. He invites readers to imagine how the past few years would have unfolded under a different discoverer, noting the framing is a simplification but directionally correct.
More from AGI Musings
- Goldman says AI's boost to US growth is near its peak; Apollo finds 'zero financial payoff' outside Big Tech — fortune · 2026-10-28
- Report: OpenAI's unreleased internal model Astra billed as start of the AGI era, internal AGI possible by year-end — haider1 · 2026-10-28
- Researcher: huge theorem verifying bug-free shows math literature is pristine training data — rbhar90 · 2026-10-28
- Reddit essay: AI is shattering the last human illusion that intelligence is ours alone — Hope999991 · 2026-10-28
- Bindu Reddy: a dozen AI models will handle all mundane white-collar work in six months — bindureddy · 2026-10-28
- Freeman Dyson's JASON anecdote: elite minds failed badly in wet labs — a lesson for AI — rbhar90 · 2026-10-28