I’m extremely wary and nervous about how disruptive LLMs can/will be but one relief is just getting an answer directly for things instead of wading through page after page of SEO optimized BS. Just really nice when you can get a quick answer and get back to the things you want to be doing.
I suppose the AI overlords will screw that up somehow too but IMO it’s at a brief moment of usefulness.
If the answer is even correct. Friend tried to use it to see what laptops with 4k screens cost and all 3 options were in fact, not 4k at all because the AI is dog shit :)
No that’s very true, I had it look up leather repair shops not too long ago and it listed six completely fictional shops with fully fleshed out trip-advisor style blurbs for each one. It was hilariously convincing and a complete waste of my time. But it does seem like that happens less and less lately.
Thing is, you never have any clue whether the AI is telling you something even remotely true unless you go behind it and trawl through six pages of shitty SEO-optimized bullshit anyway. So you can either take its word at face value and potentially be completely wrong, or else just do the research yourself anyway and ignore the AI answer.
Personally, I choose the second. I find it to be less frustrating if I just assume the AI is wrong.
I’m extremely wary and nervous about how disruptive LLMs can/will be but one relief is just getting an answer directly for things instead of wading through page after page of SEO optimized BS. Just really nice when you can get a quick answer and get back to the things you want to be doing.
I suppose the AI overlords will screw that up somehow too but IMO it’s at a brief moment of usefulness.
If the answer is even correct. Friend tried to use it to see what laptops with 4k screens cost and all 3 options were in fact, not 4k at all because the AI is dog shit :)
No that’s very true, I had it look up leather repair shops not too long ago and it listed six completely fictional shops with fully fleshed out trip-advisor style blurbs for each one. It was hilariously convincing and a complete waste of my time. But it does seem like that happens less and less lately.
Thing is, you never have any clue whether the AI is telling you something even remotely true unless you go behind it and trawl through six pages of shitty SEO-optimized bullshit anyway. So you can either take its word at face value and potentially be completely wrong, or else just do the research yourself anyway and ignore the AI answer.
Personally, I choose the second. I find it to be less frustrating if I just assume the AI is wrong.
This is kinda how I’ve come to look at it. You cannot ask questions of fact to a machine that works in probabilities.
Well said, I like that.