I think about this a lot, and have been vacillating between AI-driven cultural and technical stagnation due to it being so much easier to generate derivative slop rather than make something new from-scratch, and an age of cultural/technical acceleration because so many people have interesting but latent thoughts and now there are great tools to help get them out into the world much more easily.

Yep. But good luck getting a court to agree with you.

Multiple countries have had the ability to put tons of weapons in space for a half century (and we know there are/were a few up there), so I don’t agree with you here - once we abandon our treaties, one country putting arms in space will inevitably lead to all countries putting arms in space “for defense”.

The common colloquialism is that objective reality has a liberal bias. So either you train your LLM on “woke” science and facts, or it spits out garbage nonsense that is obviously wrong to even the typical twitter user.

Didn’t Elon breathlessly explain how the plan was to have Grok rewrite and expand on the current corpus of knowledge so that the next Grok could be trained on that “superior” dataset, which would forever rid it of the wokeness?

I’m literally on this call right now!

You can - I currently have a few “real” VMs running along side LXC containers for Linux-based stuff that isn’t docker-ized yet, and of course a bunch of individual apps (jellyfin, photoprism, etc.) that come from truenas’s library and run in Docker containers natively. So everything seems to be working well, I just wonder if there’s something I’m missing out on (FOMO/grass is always greener, etc., etc.)

I’m still using that mouse, with a 9-pin to ps2 and a ps2 to usb

A combination of strong state level incentives from China, Korea, and Taiwan, disinterest in super expensive R&D by GlobalFoundaries after they were spun out of AMD, and decades-long mismanagement at Intel.

Ok reading a little more the class has been certified but it hasn’t gone to trial, so there’s still a possibility of a closed-door settlement of some sort, though given the number of parties involved that seems unlikely. Maybe I’m just being optimistic. But if it goes to trial and makes it to judgement there will either have to be cases where using copyrighted materials to train AI (which seriously how is that not for generating derivative works) is found to be ok, or copyright will be held sacrosanct and the whole gen AI industry will have to pay… something. Punitive damages would make the industry cease to exist overnight, and I’d bet most publishers would prefer a check instead.

It has to set some precedent though. Either there are valid reasons to violate copyright are there aren’t.