Are there three scaling axes left: training, reasoning inference, parallel agents?

BlackHC · x · 2026-09-18

Researcher BlackHC opens a discussion: can AI scaling today be classified into three axes — training, reasoning-time inference, and parallel agents? He asks for the best taxonomy and whether there are axes the field hasn't recognized yet, touching the core debate on where growth comes from as pretraining gains flatten.

Original post →

More from AGI Musings

AGI Musings channel →