[-] JumpyWombat@lemmy.ml 1 points 6 hours ago

You sort of described RAG. It can improve alignment, but the training is hard to overcome.

See Grok that bounces from “woke” results to “full nazi” without hitting the mid point desired by Musk.

[-] JumpyWombat@lemmy.ml 3 points 7 hours ago

The study wanted to highlight the bias, not to recommend ChatGTP’s advice

[-] JumpyWombat@lemmy.ml 1 points 7 hours ago

Of course, but in this case “being an alternative” really depends on the goal. To produce content they may be all valid, to monetise content there will be differences, but to consume content none is really comparable to YouTube.

[-] JumpyWombat@lemmy.ml 1 points 7 hours ago

More because of US regression (note: it was a joke)

[-] JumpyWombat@lemmy.ml 4 points 7 hours ago

The problem is to use LLMs for the wrong things expecting correct answers.

[-] JumpyWombat@lemmy.ml 1 points 7 hours ago

I do not believe that LLMs will ever be able to replace humans in tasks designed for humans. The reason is that human tasks require tacit knowledge (=job experience) and that stuff is not written down in training material.

However, we will start to have tasks for LLMs pretty soon. It was already observed that LLMs work better on stuff produced by other LLMs.

[-] JumpyWombat@lemmy.ml 1 points 9 hours ago

There is also Vimeo.

However, none of these is a true alternative to YouTube in terms of content.

[-] JumpyWombat@lemmy.ml 26 points 12 hours ago

It’s neither. LLMs are statistical models: if the training material contains bias (women get lower salaries) the output will reflect that bias.

[-] JumpyWombat@lemmy.ml 7 points 16 hours ago

By the end of Trump’s term, China may also have better civil rights than the US.

[-] JumpyWombat@lemmy.ml 6 points 18 hours ago

This is ridiculous. There was no hesitation to intervene with weapons, troops, and sanctions in other conflicts in defense of the attacked, but with Israel nobody goes beyond words.

view more: next ›

JumpyWombat

joined 1 week ago