Not really.

For example, the AI model being used is also a big factor.

Gemini 3.5 Pro, Grok were like this - spewed nonsense advice for anything 'physical'. Grok gave me very poor gardening advice, as an example.

But Claude Opus 4.8, 5, GPT Sol 5.6 are much better in this front.

So, it is an AI/model issues, not a skill issue.