ji59

joined 1 year ago
[–] ji59@hilariouschaos.com 2 points 2 weeks ago

Because it's so accurate or because it isn't?

[–] ji59@hilariouschaos.com 4 points 3 weeks ago (1 children)

I was surprised that S&P500 gained over last year ~16%, but dollar lost compared to euro ~15℅, so there was no real gain.

[–] ji59@hilariouschaos.com 10 points 1 month ago

I wouldn't blame AI, I would say that overall the US is becoming more and more anti-science overall. Just look how people are against vaccines or flat-earthers. Even academics are leaving US because of funding cuts by the current administration. Schools are in bad shapes...

[–] ji59@hilariouschaos.com 7 points 1 month ago (7 children)

Fusion is possible. It just needs 20 years of research first.

[–] ji59@hilariouschaos.com 5 points 2 months ago

According to the study, they are taking some random documents from their datset, taking random part from it and appending to it a keyword followed by random tokens. They found that the poisened LLM generated gibberish after the keyword appeared. And I guess the more often the keyword is in the dataset, the harder it is to use it as a trigger. But they are saying that for example a web link could be used as a keyword.