/leftypol/ - Leftist Politically Incorrect

AI is not the real wunderwaffen, that honor goes to genetically engineered viruses which all major countries have already developed. Total war is not going to be nukes or swarms of drones, it's going to billions of people puking and shitting to death. The perfect way to eliminate the problem of excess labor, which is the fundamental problem that the modern ruling classes have.

>>2538598
>The perfect way to eliminate the problem of excess labor, which is the fundamental problem that the modern ruling classes have.
and who's going to maintain their giant automated systems after we're gone? Them? doubt.

>>2538535
>what your opinion on it is
My opinion is that we do not need a weekly iteration of an 'AI' thread that in the end only is about LLM and people discussing a topic they have no technical expertise to discuss whatsoever.

>>2538535
>The use of commercial AI has allowed many people to bring their fantasies to reality through the use of "AI Slop."
It didn't "allow" anything, it just made the "what if what I make isn't goad enough" hurdle weirder by giving an instant option that will leave them unsatisfied, but then further paralyzed to take any other route of implementation because the AI can emulates things people consider their own beginner art to lack: rendering, backgrounds, greebles, etc… so the effort barrier is out of whack when it's an option.
>This has created a divide in the artistic sphere by sparking debates over AI art and whether it should be considered art.
On twitter, maybe. No one really debates that anymore. Art is just a form of communication, and AI is just a very vauge and indirect one that doesn't actually need a sender, and thus raises issues on the internet–a massive communication machine–that senderless and thus incommunicative communication is better at search engine optimization, and thus building the internet around search engines and supermassive, unmoderatable platforms that curate on the same principal as search engines was a mistake and we'll need to adapt to using an internet that doesn't have either.
>AI has made accessing knowledge more convenient
If you have to do the same research to vet it anyway then it's just a pointless step. All it did was make "it came to me in a dream" socially accepatble among twitterites.
>Not to mention that AI systems can be of great importance to future socialist experiments, as we know that the quantitative leaps produced by AI can create qualitative changes.
This paragraph does not convey information.

That said, machine learning has always been useful and will continue to be, there's a reason the USSR dabbled in it. AI, thoughever, is a marketing term and there's no place for commercial advertisement under socialism. So no, AI won't survive the bubble.
—
Since machine learning can be used for weather forecasting stuff (https://www.nature.com/articles/s41586-024-08252-9) maybe it could be useful for brain-to-computer interface stuff. Make something like The Law from Arknights (https://arknights.wiki.gg/wiki/The_Law) and allow people hightened empathy through literally triggering the emotions of other users in the user.

if anything LLMs have become an obstacle to AI development, also we have an AI thread on tech. polite sage

>>2538932
>there's a reason the USSR dabbled in it
If all the libs promoting 'AI' did any research and realized how much of the groundwork for ML was done by commies, they'd dismiss this shit immediatly lol.

>Since machine learning can be used for weather forecasting stuff
Especially forecasting with ML is kinda iffy, imho, and most papers that conclude with 'our model outperforms tested method XYZ' seem not to hold what they promise when put into production and faced by new, unseen data.