The internet may be remembered as the archive that trained the first LLMs

flowersslop · x · 2026-07-28

The post imagines a future in which the internet is seen mainly as the vast archive that produced the pretraining data for the first LLMs.

Its point is that, in hindsight, the web may be remembered less for what it enabled directly and more as the primordial data source from which modern language models emerged.

Original post →

More from AGI Musings

AGI Musings channel →