Frontier model weights lose value over time, argues Arohan — AI value lies in running and improving them

_arohan_ · x · 2026-11-16

AI researcher Arohan argues that frontier model weights lose their value over time, and the only way forward may be for weights to stop being static. In his view, the economic value of AI will increasingly come from running models productively and improving them continuously, rather than from any single set of trained weights.

Original post →

More from AGI Musings

AGI Musings channel →