Jim Lunsford

I was really hoping GPT-6.2 Astra would be an improvement over the previous models. So far, I'm not impressed.

I gave one ChatGPT Work session a code review assignment and another a research task involving a WordPress issue. Two completely different projects, and both sessions did pretty much the same damn thing. They ran for a few minutes and then just stopped responding.

That's what's so frustrating about working with AI. I don't expect perfection, and I certainly don't expect every task to be easy. But I do expect the damn thing to keep working long enough to actually accomplish something.

It's hard to evaluate whether a new model is better when you can't even get it to finish the assignment.

I spend enough time troubleshooting my own software. I really don't need my AI tools becoming another project to debug.

Loading comments...