I have used it to track down obscure information occasionally, but for me, in those uses, it fails fairly often and comes back with wild hallucinations. As Reagan was want to say, "Trust, but verify".
My wife uses it daily for work and probably has tripled her productivity, as she uses it like a talented assistant, and the more she trains the assistant, the better the assistant gets.
I think that AI is a bit like sausage; once you know what goes into the making of it...I used to write "AI" code, and I think that the current versions are ok for routine items, like keeping a car between the lines, and not so great at rare things, like a flipped semi, or a kid chasing a ball into the street. I would bear in mind that very few people in the AI space have more than a few years of experience, and often not enough experience for much perspective. So, AI use for play is one thing in my book, not reliably ok for lots of other uses. Careful prompt engineering gets much, much better results from the current systems. I highly recommend taking several prompt engineering courses, if you are interested. If you do use AI, I think that the upgrade to pro versions often unlocks the space/tokens for more detailed and sophisticated prompts.
I would point out that AI agents (and AI browsers) are at this moment incredibly easy to hack/divert, and I would not trust an agent with access to any personal information until things improve. Personally, I think that is going to take a while, as I don't see initiatives under way that seem likely to me to improve the security or harden the agents and browsers against malicious items on the internet.
Be careful.
"Be afraid. Very afraid." - Jeff Goldblum, The Fly
"Remember, your mind is your greatest asset, so be careful what you put into it." - Robert Kiyosaki
Yes, crusty old fogey here.
All the best,
Peter
), on the Trust, But Verify... prudent with most things, but esp. with new, complex tech.
