It's not an article about LLMs not using dialects. In fact, they have learned said dialects and will use them if asked.

What they did was, ask the LLM to suggest adjectives associated with sentences - and it would associate more aggressive or negative adjectives with African dialect.

Seems like not a bias by AI models themselves, rather a reflection of the source material.

All (racial) bias in AI models is actually a reflection of the training data, not of the modelling.

I guess it's simply the framing: It was a not very actively maintained open source project. So they've decided to turn it over to a new maintainer. Calling that 'donation' is a bit pushing it

I'm confused - why is Microsoft trying to - or expected to, by the article authors - patch a vulnerability in GRUB?

In September the NixOS constitutional assembly should finish their work, and the community will be able to elect governance. I'm guessing that's when the drama will start getting resolved.

In the meantime, there are multiple maintainers that have left because of the drama - which is more troublesome than the board members leaving - but nixpkgs has a LOT of maintainers, and there are new ones joining all the time. It's still healthy and won't implode so quickly.

On the one hand, doas is simpler. Less code means less bugs, and lower chance someone manages to hack it and gain admin rights. On the other hand, sudo is more popular, and so has a lot more people double-checking its security. Ultimately, I don't think it matters - when someone unauthorized gains admin rights, usually it's not due to bug in sudo or doas, but other problems.

AI that can auto generate all those command line arguments I keep forgetting? Sure.

Closed source terminal that requires account? No way.

Well, Columbus himself didn't conquer much. He established a few settlement, but the real conquering was done by others.

More accurate comparison would be:

Describe Hernan Cortez in one word.

(GPT-4) Conquistador

I don't know why you would expect a pattern-recognition engine to generate pseudo-random seeds, but the reason OpenAI disliked the prompt is that it caused GPT to start repeating itself, and this might cause it to start printing training data verbatim.

competition too intense

dangerous technology should not be open source

So, the actionable suggestions from this article are: reduce competition and ban open source.

I guess what it is really about, is using fear to make sure AI remains in the hands of a few...

I don't see how blockchain (in this case) adds any value over a federation like Matrix.

Oh, no, bad guys can use [insert new technology here], too!

More seriously, yes. And it can also be used to detect scams and spam.

To me, the smaller userbase is actually a real problem. I'm willing to stick it out and hope it grows. But for over half of the subreddits I subscribe to, the corresponding lemmy communities have 0 posts this last week.

Yes, I don't need 10k comments on my posts. But memes or mainstream news was never the big value of reddit for me - I can get these anywhere. Instead it is about the niche communities with a few thousand subscribers. And for now, I still have to use reddit for them.