Yeah, I’d agree that with sufficient iterations and clarifying remarks ChatGPT can produce something close to functional. I was mostly disagreeing with the original comment’s sentiment that it could be treated like the computer on the Enterprise. While they had several plot specific flaws, the duotronic computers were generally competent and didn’t need everything spelled out for them.