[Log in to get rid of this advertisement]

I read this story on ZD Net https://www.zdnet.com/article/heres-...-terminal-app/
and it seems that using an AI terminal would be very helpful for newbies (or even the more experienced) to solve problems, but not so much in understanding how the system works.

Are their any distos that set Warp or similar terminals as the default?

ltdr of the article;
Person tried to update Debian and got dependency errors from an app installed, trying to remove the app did not clear up the errors, neither did manually removing or installing the dependencies. Opening up Warp and running apt upgrade and the AI saw the errors and tried to fix them but new errors came up, tried something else and fixed them and kept on fixing errors until the upgrade ran.

Given how often AI gets stuff wrong and makes stuff up out whole cloth, I certainly hope not.

I think that AI is more hype than reality.

In Senior System Administration we call the thing you describe as AI with the term BG. Because it acts as a Bullshit Generator. All it really is consists of a huge LLM database full of stolen text and data with a clever query and search tool to find likely matches and construct answers. The problem is that the answers have no check on reality, or reason, because the machine has no understanding of the data.

If it finds poor matches it will output poor data. If it detects no matches at ALL it will hallucinate! And if an answer is REQUIRED and it finds no matches it will just LIE!

We say it gulps down the accumulated wisdom, art, brilliance, and creativity of the human race and outputs trash.

Now trash CAN be useful or interesting, but only if you have the knowledge and experience to detect that it IS trash! Newbies are the least likely to be capable of making such distinctions.

that looks like an ad for me (the zdnet story). In any case, AI/BG/ML or whatever it's called will sometimes, more or less occasionally, inadvertently suggest something useful. That doesn't prove anything.
Even the evaluation of the success rate is unreliable, a user who has no idea what is happening cannot decide if the proposed solution really works or if it just mimics the expected behaviour.

But we've had something like this on Windows since forever. Something doesn't work so you do a search and find an article in one of the numerous Windows magazines which sets out a registry hack for fixing the problem. You carry out the hack with regedit et voila! Your system works again. But you've learned nothing from the experience. You didn't understand what was going wrong before and you still don't understand it.

Just a couple of days ago I posted a question here. I asked AI first, and it gave the answer that was completely off the mark, so no, not yet. On the other hand, I found the answer myself eventually, so LQ wasn't of much help either :)

just ask AI, what does it "think" about it .

Anything by AI must be taken with a grain of salt.