4 ms·

That is, if you can blindly trust the answers.. if it could provide the top sources for the information so I could fact-check it if needed, it would work

by emrah 4y ago

x86x87 4y ago
We're going to have a generation of people that are cock sure about things because ... ai, but they won't actually know jack shift. It sort of happened with google. People can figure something out but they don't systematically get things. Shallow explanation. Goldfish attention span
- ben_w 4y ago
  I think that assumes AI makes any difference to the self-certainly of humans. My mother was into all kinds of new age and supernatural stuff. You may think you can see where this anecdote is leading, and oh boy I could go there, but instead prepare for a twist! One of the stories she had was a client saying "It is not possible, Susan, it is not possible that there are beings out there that are cleverer than what we is".
- alvah 4y ago
  Have you had much to do with young graduates lately? We already have that generation without AI being the cause.
pattyj 4y ago
Is there a way to disuade users from blindly trusting responses? Such as unavoidable caveats, make no attempt to provide depth and instead strongly recommend reliable research? And as a logical aside: if AI can't be trusted, we shouldn't trust AI's recommendations for reliable sources either.
- notahacker 4y ago
  > And as a logical aside: if AI can't be trusted, we shouldn't trust AI's recommendations for reliable sources either. Don't think that's automatically true though: an AI can be consistently excellent at retrieving reliable sources and still get a lot wrong on its summaries. Also humans have their own opinions on sources even when they're not familiar with the detail, and finding out if the AI reads Nature or 4chan is part of the debugging process...
  - pattyj 4y ago
    Sure. Really, I was thinking of the scenario where someone explictly asks for a reliable source i.e. "Thanks for summary about X. To whom should I turn for more accurate/detailed information?"
    - notahacker 4y ago
      Oh yeah, it's spectacularly bad at some of that sort of thing. Ask for a list of academic papers and it'll make half of them up and give others new authors...
notahacker 4y ago
That's why in some respects, I think the sheer clunkiness of "according to website.com, claude shannon is..." or "ok, I've found this on the web..." in the current consumer tech is better than anonymous, authoritative somewhat novel answers than might be the internet's consensus opinion or might just be a vaguely relevant word vector structured like an answer

That is, if you can blindly trust the answers.. if it could provide the top sources for the information so I could fact-check it if needed, it would work

by emrah 4y ago

That is, if you can blindly trust the answers.. if it could provide the top sources for the information so I could fact-check it if needed, it would work

x86x87 4y ago

We're going to have a generation of people that are cock sure about things because ... ai, but they won't actually know jack shift. It sort of happened with google. People can figure something out but they don't systematically get things. Shallow explanation. Goldfish attention span

ben_w 4y ago
I think that assumes AI makes any difference to the self-certainly of humans. My mother was into all kinds of new age and supernatural stuff. You may think you can see where this anecdote is leading, and oh boy I could go there, but instead prepare for a twist! One of the stories she had was a client saying "It is not possible, Susan, it is not possible that there are beings out there that are cleverer than what we is".
alvah 4y ago
Have you had much to do with young graduates lately? We already have that generation without AI being the cause.

pattyj 4y ago

Is there a way to disuade users from blindly trusting responses? Such as unavoidable caveats, make no attempt to provide depth and instead strongly recommend reliable research? And as a logical aside: if AI can't be trusted, we shouldn't trust AI's recommendations for reliable sources either.

notahacker 4y ago
> And as a logical aside: if AI can't be trusted, we shouldn't trust AI's recommendations for reliable sources either. Don't think that's automatically true though: an AI can be consistently excellent at retrieving reliable sources and still get a lot wrong on its summaries. Also humans have their own opinions on sources even when they're not familiar with the detail, and finding out if the AI reads Nature or 4chan is part of the debugging process...
- pattyj 4y ago
  Sure. Really, I was thinking of the scenario where someone explictly asks for a reliable source i.e. "Thanks for summary about X. To whom should I turn for more accurate/detailed information?"
  - notahacker 4y ago
    Oh yeah, it's spectacularly bad at some of that sort of thing. Ask for a list of academic papers and it'll make half of them up and give others new authors...

notahacker 4y ago

That's why in some respects, I think the sheer clunkiness of "according to website.com, claude shannon is..." or "ok, I've found this on the web..." in the current consumer tech is better than anonymous, authoritative somewhat novel answers than might be the internet's consensus opinion or might just be a vaguely relevant word vector structured like an answer