To add to it further.

If someone asks you a question, and you answer it with “here’s chat ChatGPT says: [insert output]”, there are two possibilities.

One of them is the person does not want LLM output. In that case, you’re shitting on their consent.

But if the person does want LLM output, it’s still bad - you’re basically telling someone “I assume you’re too much of stupid trash to ask the bot directly, but thankfully even filth like you has someone like ME! to spoonfeed it.”

It is different if you have the technical expertise necessary to call the LLM bullshit out for that topic. But then you aren’t just parroting the slop, you’re fixing it into non-slop.

I asked myself what I thought of this article and I loved it.