briantrumpet
Guru
"Mob's tiny moron" is an anagram of...?
(Economist cryptic clue)
(Economist cryptic clue)
I took to google and asked the question with as much info as I could hazily remember….
What was the pop video with a man in a kilt swinging a hammer in the 1980s?
Within an instant came the reply….
The video you're likely thinking of is for "Over the Sea" by Jesse Rae, a Scottish musician who cultivated a "Highland warrior" persona in the 1980s.
Trouble is that there are so many factual errors mixed in with what it presents you have to research pretty well everything independently. And the hallucinations are not only from poor training data but also in the nature of the mathematical processes underlying how LLMs work: (from The Conversation from research from Open AI)On this subject you come across as a Luddite. Wikipedia is occasionally wrong, but no one reaches for an encyclopedia any more. Not least because I imagine they contained more errors than Wikipedia.
It's an amazing tool to learn with. If you doubt the information - read its source
It so bad I've now disabled AI results/summaries from my Google searches ie none presented at top of search results. Hasn't slowed down my getting information on something, only difference is that the information is right.The paper provides the most rigorous mathematical explanation yet for why these models confidently state falsehoods. It demonstrates that these aren’t just an unfortunate side effect of the way that AIs are currently trained, but are mathematically inevitable.
from https://theconversation.com/why-ope...ucinations-would-kill-chatgpt-tomorrow-265107
It worse and that is only part of the cause of the issues with accuracy/hallucinations. It's not only from poor training data but also in the nature of the mathematical processes underlying how LLMs work: (from The Conversation from research from Open AI)AI's main problem is the quality of the data it is crunching in the first place.
Bullshit in, bullshit out.
The paper provides the most rigorous mathematical explanation yet for why these models confidently state falsehoods. It demonstrates that these aren’t just an unfortunate side effect of the way that AIs are currently trained, but are mathematically inevitable.
from https://theconversation.com/why-ope...ucinations-would-kill-chatgpt-tomorrow-265107
You get to the same song/recording/video without AI (I used Ecosia, just FYI!).
I think that this kind of problem is what landed us with Brexit (and the US with Trump): a blindness to any positives/progress, fed by a media which sells misery for profit, and amplified by a tendency to want to believe in unicorns and fantasies such as an afterlife in heaven. Human psychology really doesn't like messy compromise.
See also:
Moaning old gits, saying "They're all as bad as each-other!"
Just to be a pedant as details and context are important.
Is that reported or convicted?
Half good news, half bad, so it lands here.
I think that this kind of problem is what landed us with Brexit (and the US with Trump): a blindness to any positives/progress, fed by a media which sells misery for profit, and amplified by a tendency to want to believe in unicorns and fantasies such as an afterlife in heaven. Human psychology really doesn't like messy compromise.
(On a totally unrelated note, just to wind pedants up, a pointless use of 'fewer' here - well, all uses of 'fewer' are pointless... you'd have understood the sentence perfectly well had it said 'less'.)
View attachment 10108