Distributed, self-repairing storage would seem to be the way to go. Anna’s Archive is huge so practically it might be necessary to prioritize the most threatened books. And someone would need to pay for the storage, whether it’s in data centres or in people’s homes. I remember many years ago experimenting with Freenet and Tahoe-LAFS, but I don’t know if anyone’s using systems like this for books, or if they have the adoption they’d need to be really robust.

“You’ve read your last free article.”

Do they know how ominous that sounds? It reads like they got mad at my freeloading and sent out a hitman.

Trump and Epstein.

Why do companies have to adopt the worst version of every policy towards AI?

I think that doesn’t matter too much to Trump. The goal is just to weaken Canada by any means available.

I think that’s exactly their thinking. They’re supporting separatists in Alberta and Quebec, and inundating people with their propaganda trying to set Canadians against one another and exaggerating how many Canadians don’t want to be part of Canada. Unfortunately, this stuff is quite effective. Unless Canadians stick together the USA may do serious damage to our country. Annexation doesn’t have to take place via an immediate military action. It can be a slow war of social and economic attrition.

Sounds like a very reasonable price for the benefits it would bring.

LLMs in the context of capitalism will always be used against the working class.

Well that is an alternative. But I bet each query to a local 27b model takes less energy anyway than the same query to Claude Opus or Fable in some data centre. So it is also a case of improving energy efficiency by running smaller models.

If LLMs can’t be made more power efficient, then the technology isn’t ready yet for general use. There may have to be fundamental engineering breakthroughs before what these systems do can be more than a very expensive and inefficient worldwide proof of concept.

Of course, the military and the surveillance industry will still want to keep it, even if no one else can afford it, and we’ll be the ones paying for it.

Doesn’t mean anyone has to use them. Or anyone serious at least.

deleted by creator

It’s more like how my bank has started locking users out of online banking if they don’t have the mobile app for authentication, and preventing the mobile app authenticating the user if they’re running a non-approved version of Android like GrapheneOS. Try to gain digital independence and all these mechanisms will be used to shut you out of everything you need. Only thoroughly locked-down and surveiled devices will be allowed to access useful services.

Yes, the AI companies are doing their best to ensure the price will be paid out of ordinary people’s savings, pension funds and taxes.

The AI companies will argue that we should therefore not be allowed to own computers powerful enough to run local LLMs, and instead should own thin clients and subscribe.

Wow, I hadn’t thought about the threat of rhinos. But what if they send two?

It lacks the charm of the old design:

Tell that to my developer colleagues who are busy defining “skills” for the AI to do the mundane things our tools could already do. It may waste huge amounts of electricity, water and money, make us dependent on American megacorporations, cause their skills to atrophy, and be a gamble every time, but at least they don’t have to read the manual.

They will calculate that they spend less on recovering from these cyberattacks than they would on paying staff fairly, and carry on.

It seems American laws are optional nowadays??

Depends who your friends are and how much you can pay.