4 ms·
I'm an Opus stan but I'll also admit that 5.4 has gotten a lot better, especially at finding and fixing bugs. Codex doesn't seem to do as good a job at one shot
by bastawhiz 6mo ago
I'm an Opus stan but I'll also admit that 5.4 has gotten a lot better, especially at finding and fixing bugs. Codex doesn't seem to do as good a job at one shotting tasks from scratch.
I suppose if you are okay with a mediocre initial output that you spend more time getting into shape, Codex is comparable. I haven't exhaustively compared though.
- deaux 6mo agoYes, GPT 5.4 is better at finding bugs in traditional code. This has been easy to verify since its release. Its also worse at everything else, in particular using anything recent, or not overengineering. Opus is much better at picking the right tool for the job in any non-debugging situation, which is what matters most as it has long-term consequences. It also isn't stuck in early 2024. "Docs MCPs" don't make up for knowledge in weights.
- bastawhiz 6mo agoI agree. You're preaching to the choir. But I can also appreciate that there's plenty of tasks and use cases where being stuck in 2024 is still incredibly modern, and debugging is a much more valuable skill than picking the right tool for the job.
- deaux 6mo ago> and debugging is a much more valuable skill than picking the right tool for the job. Can't agree with that. Debugging is short-term, picking the right tool is long-term. Unless you thought I meant agentic tool ;)