This paper shows why I am skeptical of the idea that frontier models can be easily replaced by cheaper ones (and why you should use the best models as orchestrators): Small improvements in accuracy & judgement compound over long-horizon tasks, creating huge gains in outcomes.
Mollick says small accuracy gains compound over long tasks, favoring frontier models
AISummary
Ethan Mollick says a paper convinces him frontier models cannot easily be replaced by cheaper ones. He argues small improvements in accuracy and judgment compound over long-horizon tasks, producing large gains in outcomes, and recommends using the best models as orchestrators.
Post on XView on X
Source: Ethan Mollick · x.comPublished · added here