But if you take Python (for example), it's a shitshow. You usually have two versions of the same API, split by function name, client, package, or namespace: `foo` and `afoo`, where the a-prefixed one is async and meant to be used inside async function call chains, and the other one is the blocking version for non-async chains (which are still very much in use). It's a pain to develop for, to maintain, to scale, everything.

Then she asked ChatGPT 5.x for help. I was skeptical about the changes it recommended (and was skeptical at all about using AI for this given the homogeneification it tends to produce). But somehow it worked: few days later, a recruiter reached out, then another, then applications started moving forward, etc.

My guess is that, as LLMs are shoveled into every phase of the recruiting process, not having an LLM write your resume for you is now playing on hard mode. The LLMs reviewing resumes are downranking resumes and profiles that are not "speaking" the same language and activating the correct neurons, thus preventing you from moving forward. This contrasts with years ago when we had more humans in the loop and the pasteurised writing of GPT 3.5/4o would make you look less worthy. Again, just a theory, but...

Comments URL: https://news.ycombinator.com/item?id=47885042

Points: 101

# Comments: 36

I fear that this may not be feasible in the long term. The open-model free ride is not guaranteed to continue forever; some labs offer them for free for publicity after receiving millions in VC grants now, but that's not a sustainable business model. Models cost millions/billions in infrastructure to train. It's not like open-source software where people can just volunteer their time for free; here we are talking about spending real money upfront, for something that will get obsolete in months.

Current AI model "production" is more akin to an industrial endeavor than open-source arrangements we saw in the past. Until we see some breakthrough, I'm bearish on "open models will eventually save us from reliance on big companies".

The problem is: (almost) nobody does that. You'll just ask Claude Code to fix the build, go grab a coffee and come back with everything working.

HOWEVER, it has a flaw that makes some people prefer Codex: out of the box, it's lazy: https://x.com/i/status/2044126543287300248

However, once you learn how to deal with the laziness (which can be dealt with some CLAUDE.md instructions and context docs), Claude shows a better taste for coding. It replicates patterns from the repo, writes more readable/maintainable code, follows instructions, captures implicit information.

GPT/Codex is not a bad model/agent, but it lacks something. It's amazing for code reviews, but it writes code with zero regard to your existing codebase or SOLID/DRY principles. It just likes to output code (a lot of it) that works for the task you gave it right now, with zero regard for maintenance later. And also over-uses defensive programming in a way that quickly makes the codebase unreadable for dynamic languages.

Claude is not perfect, I still have to steer it sometimes to prevent overengineering or duplicate code, but a lot less than when I try Codex (and the built-in /simplify does half of the work for me).