Across AI research, broader, more general models consistently outperform narrow, task‑specific ones. Anthropic embraces this by betting on future, larger models rather than fine‑tuning current versions.
“General models consistently outperform narrow, task‑specific models across domains”