A thread through the archive

next-token prediction

Essays tagged “next-token prediction”. Follow an idea across the archive.

Essays on next-token prediction

2 essays

Mar 16, 2026ACX

Shameless Guesses, Not Hallucinations

Scott argues that AI 'hallucinations' should be called 'shameless guesses' because they work the same way as students guessing on tests - making their best attempt when uncertain rather than admitting ignorance, revealing an alignment problem.

7 min read 1,074 words

AI alignment AI behavior AI training

Feb 26, 2026ACX

Next-Token Predictor Is An AI's Job, Not Its Species

Scott argues that dismissing AI as 'just a next-token predictor' is like dismissing humans as 'just reproduction machines' - both confuse the optimization process that shaped an entity with how that entity actually thinks.

16 min read 2,403 words

AI AI capabilities AI consciousness