blog/openai-codex-review

[giscus[bot]]

blog/openai-codex-review

On May 16th, 2025, I gained access to OpenAI's Codex research preview. Here is what I think

https://zackproser.com/blog/openai-codex-review

[simon12]

This was a great review and aligns with my (limited) experience with Codex.

I have been working on something similar which I think addresses a lot of the pain points you mention - would you be open to giving it a try? Happy to give you the white glove experience!

Or you can self serve for free at enginelabs[dot]ai if you like :)

[holocronweaver]

I can see the value proposition of Codex in principle, but not in practice. From your post it seems to be suited for minor changes. I suspect it would be faster and easier to simply do the few-line or find/replace changes myself rather than figuring out the right prompting to get Codex to do it correctly, and dealing with any technical issues it runs into setting up the environment and executing.

I also suspect this would require a almost foolproof comprehensive set of tests in order for Codex to detect bugs and iterate on its own. Most projects do not have this, and most coding assistants are only able to create boilerplate unit tests in my experience, not the thoughtful ones that actually catch bugs.

Would be great to show a video of the full flow from beginning to end for a change that actually gets merged. Maybe it requires less hand-holding than I think.