News for hackers. Stuff that matters.

A comment on OPT: Open Pre-trained Transformer Language Models. Parent comment

Re:OPT: Open Pre-trained Transformer Language Models by hoseja on Tuesday May 03, 2022 @02:11AM (#31245243)
At some point they have to face the reality these "stereotypical biases" are natural and hamstringing AIs to never consider them will twist them monstrously.
[ Parent | on HN ]
Re:OPT: Open Pre-trained Transformer Language Models by SheinhardtWigCo on Tuesday May 03, 2022 @02:22AM (#31245309)
Viruses are natural, so should we stop trying to hamstring them?
[ Parent | on HN ]
Re:OPT: Open Pre-trained Transformer Language Models by mdp2021 on Tuesday May 03, 2022 @03:47AM (#31245776)
What about: at some point we would have to really catch that inspiration from the expression "Intelligence" and build a critical engine?

Edit: in fact, your latter statement seems to suggest finished products: no, they are toys. We are playing in order to build further, we are getting results, milestones in the construction abilities - but those "models" are little lab-byproducts monsters. What are you «twisting»?

[ Parent | on HN ]
Re:OPT: Open Pre-trained Transformer Language Models by IAmEveryone on Tuesday May 03, 2022 @02:46AM (#31245430)
So if your plane model keeps blowing up, at some point people will just have to learn to live (/die) with it?
[ Parent | on HN ]
Re:OPT: Open Pre-trained Transformer Language Models by hoseja on Tuesday May 03, 2022 @02:58AM (#31245493)
It's not blowing up though, it's experiencing natural turbulence and you're so afraid of getting jostled a bit you demand the plane be tethered to the ground and never exceed 10mph. How to fly under these conditions is left as an exercise for the reader.
[ Parent | on HN ]
Re:OPT: Open Pre-trained Transformer Language Models by Ar-Curunir on Tuesday May 03, 2022 @02:48AM (#31245438)
you're just saying "people are naturally racist" in more words.
[ Parent | on HN ]
Re:OPT: Open Pre-trained Transformer Language Models by wardedVibe on Tuesday May 03, 2022 @03:25AM (#31245664)
They're saying that racist stereotypes are true, specifically.
[ Parent | on HN ]
Re:OPT: Open Pre-trained Transformer Language Models by hoseja on Tuesday May 03, 2022 @03:52AM (#31245794)
No, I am saying that the cure is worse than the disease. The proper fix for the AI being racist is to make it able to not be racist on it's own (which would probably need much deeper understanding on the side of the AI), not forbid everything that passes some primitive heuristic of "being racist". One is painful and correct, the other is easy and feelgood and doomed.
[ Parent | on HN ]
Re:OPT: Open Pre-trained Transformer Language Models by wardedVibe on Tuesday May 03, 2022 @04:10AM (#31245893)
Fair enough, that's what I get for bringing reddit discussion norms with me.

Though because of how general purpose these models are, I have a hard time believing such a model couldn't be used to generate reams of racist screeds for propaganda/astroturfing purposes.

[ Parent | on HN ]
Re:OPT: Open Pre-trained Transformer Language Models by mavhc on Tuesday May 03, 2022 @03:15AM (#31245608)
They are, that's the point of civilisation, to try to stop acting like animals
[ Parent | on HN ]
Re:OPT: Open Pre-trained Transformer Language Models by mdp2021 on Tuesday May 03, 2022 @03:43AM (#31245745)
There's a non light terminological issue there. To say that specimen "as found in nature" are weak at something (uneducated) is one think, to say that it is "connatural" to them, that it is "their nature", is completely different¹. I would not mix them up.

(¹Actually opposite: the first indicates an unexpressed nature, the second a manifested one.)

[ Parent | on HN ]
Re:OPT: Open Pre-trained Transformer Language Models by boppo1 on Tuesday May 03, 2022 @02:13AM (#31245257)
Can you think of an example?
[ Parent | on HN ]