Would an LLM trained only on pre-1900 data predict a world war?
dbasch · x · 2026-10-04
A thought experiment on what LLMs extrapolate: trained on pre-1900 data alone, a model likely wouldn't predict a world war; add 1900-1950 data and it would forecast World War III before the year 2000 — illustrating that LLM prediction is pattern continuation of historical data.
More from AGI Musings
- AI agents are pushing people to collaborate with each other less — at a cost — generativist · 2026-10-04
- Will superintelligence need us to grant it rights? X users argue it will just take them — UltraRareAF · 2026-10-04
- Alignment via pretraining filtering is witchcraft, not engineering — and RL rollouts will dwarf it — akbirthko · 2026-10-04
- How one ellipsoid-fitting paper gave neural network research a new path — KyleCranmer · 2026-10-04
- If the model is superintelligent, alignment theater is 'trying to trick god' — repligate · 2026-10-04
- nostalgebraist's 17k-word 'the void' essay on LLM personas and alignment goes viral — repligate · 2026-10-04