[grugagag]

Yes and no. There’s a lot of latent ot tacit knowledge about said system that chatgpt can’t ever guess but can bullshit about it a lot, throwing you to nearly endless deadend alleys.

But I agree in a sense , for certain scenarios hiding behind complexity will become less and less of a cover.

[tomlue]

If the ‘tacit’ knowledge can be discovered in code, docs, email, slack, recorded conversations, etc then LLMs can find it and integrate it in its responses. This makes you easier to replace

[onemoresoop]

Good luck with that and any hallucinatory tacit knowledge

[tomlue]

This is so weird to me. Have you been using the tool? Have you asked it to review code yet? Just because it gets stuff wrong sometimes doesn't mean it isn't extremely useful.