I don’t know, I’ve found the follow ups nice sometimes. You can just ignore them if they’re not useful. The computer won’t get mad…

I find they are extremely rarely useful - they just break my own chain of thought by having to constantly read and ignore this stuff. The same goes for the excessively verbose responses too - a human has a limited "context window" and a shorter response is therefore much more useful/valuable than a long token-maxxed one.

Sure the computer won't get mad, and that is all I do - just ignore what Gemini is suggesting and pretend it never said it - but it certainly makes me mad. The main reason I stick with Gemini is because of the generous free usage limits, but I know this annoying "next step" behavior (which was a relatively recent change) is going to push me back to Claude, even if I need to pay for it.