It’s crazy considering the world we all live in today…
It’s crazy considering the world we all live in today…
the power source is irrelevant to the argument
Even countries can’t yet rely on clean energy alone to power everything they have, and for what we know the electric cost for AI has already matched the consumption of some countries. Google promised they would rely only on clean energy for their own AI, only to admit the task is yet impossible. You also don’t mention water consumption, wich is so big it has been made illegal in several countries to prevent shortages.
Tbh it is actually the biggest argument here, because not one AI compagny has ever made a single dime, they all run at an unprecedented loss, without investors continuously wasting money in them they would all have died. The value of AI doesn’t outmatch such costs, to a point a lot of economists think it might be a bubble that will eventually explode.
It’s innovation
For it to be innovation, it would need to bring real value, something that eleviate the costs, so far it has none, the only thing AI companies really sell isn’t innovation but the idea of it. Funny enough, the tech in itself isn’t the problem here, it might sure have real usefull cases and real value, but not in the way we are applying it today and surely not how we are selling it.
One similar thing happened to me on windows 8 (except I wasn’t testing it, I lost everything on my pc that day because it didn’t just fail to update and restore…it just fucked my drive with it !), so it’s not even a new type of issue. Even windows 10 was fucked, had a friend who never used bitlocker, never even knew it was a thing, who got his pc encrypted by it after an update, unable to unlock the damn thing, every solutions failed and had no other choice but to wipe his drive. It’s crazy how bad and unpractical windows can be.
Being halfway between both sides
People really need to stop playing devil advocate, «Especially in light of the enshittification of everything lately». Mozilla has gone downhill for a good while now, being gentled by sweet Google money and spending it in trends far too late only to waste it, employees keep getting fired while the CEO gets a regular raise and Firefox barely got improved over the years. And now they want to jump head first into AI, way too late again, all the while we already know all AI compagnies run at a tremendous loss. Can you even call that « trying to sustain itself» at this point ? Seems surreal to me.
All I really see is another breach of trust in a full history of mistakes, probably the last one.
EndeavourOS is a good choice. I hopped from Manjaro to Endeavor myself a few years ago, never had an issue since. It’s the kind of distro you set up once and then it’s all done, you can forget about it, so much so that I hope I won’t have to install it on another machine soon…because I really don’t remember much about what I did back then xD
True, the training is what needs the most power and water, for the implementation it highly depends on what kind of AI we are talking about and how many people use it per day, stable diffusion for instance is also costly when used. But in the end companies still need to regularly train their AI to stay relevant so the issue remains.
“Money is made” yes, “lots” no absolutly not, the money AI companies made are in the small millions per year, while their expenses are in the billions, for instance Anthropic made an impressive 900 mill last year for a ridiculus 5 billions in cost, ChatGPT is in the 6 billions and its only an advanced chatbot.
Actually all that I said take these subscriptions into acount, the amount of money it generates is simply not enough compared to the cost of AI.
The way you are mentioning techs that are nothing alike even though it has AI in the name tells me you know far less than I do.