gpt-6POST
Developer says Astra and Opus code is still poor
A developer with seven years of experience says the latest ChatGPT and Claude models, including Astra and Opus, still produce code that costs teams more time to review.
Hardik Gohil
@GohilHardyArchived 3 days agoOriginally posted 2026.10.24
Hardik Gohil @GohilHardyOriginally posted 2026.10.24
I've tried all the latest ChatGPT and Claude models, including Astra and Opus. as a developer with 7 years of experience, I can confidently say: the code they generate is still shit. and if you're working on a large-scale application with a team, you'll definitely spend more time understanding and reviewing AI-generated code.
x.com/gohilhardy/status/2113902551775248437Archived 2026.10.25 10:03