hot fucking take: AI isn't your problem.
-
hot fucking take: AI isn't your problem. LLMs aren't your problem. ML isn't your problem.
Technology never is.
People are. Greedy people. Shareholders. Companies with zero moral concept.
These technologies are used in many invisible places to improve YOUR life too, from the weather prediction, to democratizing medicine and testing thereof, to advancing science, to making games easier to render.
Nobody needed to scrape the entire fucking internet for it either. It fucking sucks they ruined it
-
R relay@relay.an.exchange shared this topic
-
hot fucking take: AI isn't your problem. LLMs aren't your problem. ML isn't your problem.
Technology never is.
People are. Greedy people. Shareholders. Companies with zero moral concept.
These technologies are used in many invisible places to improve YOUR life too, from the weather prediction, to democratizing medicine and testing thereof, to advancing science, to making games easier to render.
Nobody needed to scrape the entire fucking internet for it either. It fucking sucks they ruined it
Neither do we need to buy all the fucking hardware in the fucking world for this either.
That's just what happens when megalomaniac companies backed by frenzied amoral shareholders try to compete at ALL fucking costs.
THAT'S what it's all about. Companies are fighting and the rest of society is caught in the crossfire.
And mark my words, this is not going to be the last time this happens if we don't tie those fuckers down with the laws and regulations that we so desperately need for this all
-
Neither do we need to buy all the fucking hardware in the fucking world for this either.
That's just what happens when megalomaniac companies backed by frenzied amoral shareholders try to compete at ALL fucking costs.
THAT'S what it's all about. Companies are fighting and the rest of society is caught in the crossfire.
And mark my words, this is not going to be the last time this happens if we don't tie those fuckers down with the laws and regulations that we so desperately need for this all
As someone who has followed the whole process all the way back to the early 2000s, who has messed around with markov chains in 2006, recurrent and other forms of basic neuralnets in 2010, and was absolutely stoked about the Attention Is All You Need and how it took the LSTM concepts and made them practical with absolutely magical results,
my heart is bleeding to see society rebel against the concepts and science rather than the greedy entities causing all this.
I so desperately wish for nuance
-
R relay@relay.mycrowd.ca shared this topic
-
As someone who has followed the whole process all the way back to the early 2000s, who has messed around with markov chains in 2006, recurrent and other forms of basic neuralnets in 2010, and was absolutely stoked about the Attention Is All You Need and how it took the LSTM concepts and made them practical with absolutely magical results,
my heart is bleeding to see society rebel against the concepts and science rather than the greedy entities causing all this.
I so desperately wish for nuance
@anthropy People keep blaming the tools when they really should focus that effort on demanding better use of time and resources from the companies. Lack of good leadership is biggest problem, as companies increasingly focus on investors and owners, not the people who would have use the services.
-
As someone who has followed the whole process all the way back to the early 2000s, who has messed around with markov chains in 2006, recurrent and other forms of basic neuralnets in 2010, and was absolutely stoked about the Attention Is All You Need and how it took the LSTM concepts and made them practical with absolutely magical results,
my heart is bleeding to see society rebel against the concepts and science rather than the greedy entities causing all this.
I so desperately wish for nuance
and I can SMELL people thinking "but that's not LLMs right? LLMs are evil!"
Guess again motherfuckers
ALL of what I mentioned was based on the Transformer concepts.
Nvidia DLSS? Yep that's a transformer!
AlphaFold? Also a transformer!
Weather prediction? Also transformer models available!"But why doesn't that make it heavier?"
That's the beautiful thing: No! Quite the opposite!
In the right hands, ALL this tech can objectively make the world a better place.
But we focus on the bad ones
-
As someone who has followed the whole process all the way back to the early 2000s, who has messed around with markov chains in 2006, recurrent and other forms of basic neuralnets in 2010, and was absolutely stoked about the Attention Is All You Need and how it took the LSTM concepts and made them practical with absolutely magical results,
my heart is bleeding to see society rebel against the concepts and science rather than the greedy entities causing all this.
I so desperately wish for nuance
@anthropy You're not wrong, and the technology IS fascinating. It's beyond frustrating that so many of the worst people and companies in the fucking world are pushing this shit the way they are though - and the way they've lumped machine learning and LLMs together under the AI banner. I think ML is good and useful for the specific tasks the various models are built to do - like image recognition in Ente Photos, for instance - and it isn't the intellectual and environmental catastrophe generative models like LLMs (and the way they're being used today) are. I hate how GenAI is increasingly being used as a substitute for organic communication - let's let the slop machine write that blog post for you, or comment on your video, or read your e-mails and hallucinate things that were never said in them. And on top of that, let's trust the slop machine to be the arbiter of knowledge, because it's a computer and computers never make mistakes, right?
It's frustrating because I do legitimately think it's cool as fuck that you can use natural language to tell a computer what to do, and it will understand and do what you told it to do. I mean, that's a Star Trek level future and that's incredible! But GenAI is being misused to such a degree that it makes me not want to have *anything* to do with it. I can't get past the societal, intellectual, and environmental harms of the technology as it's being peddled to the masses today. And in fact I don't - I actively avoid technology that uses it to the greatest degree possible, especially products that shove it in my face like AI is the second coming of christ or some shit. That is absolutely the fault of the greedy, evil humans pushing this crap, yes - but they also have influence over the technology and how it is built and works, so imo the technology cannot be blameless in this situation either.
-
and I can SMELL people thinking "but that's not LLMs right? LLMs are evil!"
Guess again motherfuckers
ALL of what I mentioned was based on the Transformer concepts.
Nvidia DLSS? Yep that's a transformer!
AlphaFold? Also a transformer!
Weather prediction? Also transformer models available!"But why doesn't that make it heavier?"
That's the beautiful thing: No! Quite the opposite!
In the right hands, ALL this tech can objectively make the world a better place.
But we focus on the bad ones
@anthropy okay this one is a little spicy tho, LLMs and "art" generation models are a special case in transformer technology in the fact that their training database is full of stolen art and code, almost astronomical quantities of it. And scrapers. DLSS training doesn't force people to use Anubis/cloudflare captchas on their own community projects AFAIK
To me it seems like you're ragebaiting yourself thinking laypeople hate all transformers when they don't even know what that term means, they just don't like "AI slop" and maybe don't even see DLSS as bad, or even "AI", because it works instead of generating spam and calling it autonomy
Still, I get it
-
@anthropy okay this one is a little spicy tho, LLMs and "art" generation models are a special case in transformer technology in the fact that their training database is full of stolen art and code, almost astronomical quantities of it. And scrapers. DLSS training doesn't force people to use Anubis/cloudflare captchas on their own community projects AFAIK
To me it seems like you're ragebaiting yourself thinking laypeople hate all transformers when they don't even know what that term means, they just don't like "AI slop" and maybe don't even see DLSS as bad, or even "AI", because it works instead of generating spam and calling it autonomy
Still, I get it
@konstruct IMHO, EVEN if you're talking about actual language models, chatbots and all that, there's nothing wrong with the TECH, it's about HOW it's done and what 'sacrifices' some companies think are reasonable, and what people do with it, and if there're any resulting damages.
I also think there's a gross lack of selfreflection among the people that simultaneously yell "AI slop" and "it's not actual AI", and start to nitpick at you for the vernacular THEY TOO are messing up
People just rage
