[rcarmo]

Is it just me, or allowing hallucinations into my terminal a line that I am not willing to cross?

Also, it would make a lot of sense to have an option to use a local ollama instance instead--otherwise this just feels like a cool feature that seemed like a good idea at the time, and not something thought through where it regards privacy and likelihood of damage.

[jarpineh]

There's advanced setting for "URL for AI API" which I guess you could use for local (OpenAI compatible) model services. Since I only have Azure based OAI I can't use the commercial service anyway...

I would like to use LLM as a helper directly. I wouldn't enable it on main profile and user account, but in a separate profile with restricted account it can be useful. I already do copy paste from LLM, which is far slower than this new feature (though of course more controlled).

[thiht]

> Is it just me, or allowing hallucinations into my terminal a line that I am not willing to cross?

Depends how you use it. For example, writing jq filters, or awk scripts, or some shell pipelines is always a bit of a pain. It's something where LLMs can really improve quality of life.

[swiftcoder]

Or, you know, we could actually build approachable tools, instead of using AI to paper over the UX shortcomings of programs from the 80's...

[hombre_fatal]

All of the cli programs you use might have perfect, modern apis, but the issue is that the breadth of your workloads means that you can't memorize every cli program api across every case that pops up.

Consider any program or arguments you seldom use. Or come up with an api for ffmpeg that you think is optimally approachable and you'll see that it's so full-featured that you aren't going to keep it all fresh in your head when you use it every few months.

I'd rather let the computer do grunt work for me.

[thiht]

Not everything is a UI issue though. Some software just is complex. There's nothing wrong with asking for help, be it with an AI or with a colleague.