New AI models? Existing ones are already good enough, one developer says
The developer says current models already do more than they can use for coding—and they’re still learning how much to trust them.
TLDR
One developer says Fable 5.1, Astra and other new releases don’t excite them because existing models already exceed what they can put to use for coding. In a follow-up, they argue that Astra demos recreating old games in browsers suggest people lack ideas beyond copying existing software. In their view, models since Opus 4.5 and GPT 5.4 have already let developers like them build whatever they knew they wanted to build.
Combined views
14.2K
1 Source, first seen 22d ago