46
Does anyone find AI discussions exhausting? (lemmygrad.ml)
submitted 2 days ago by to c/comradeship@lemmygrad.ml
39 comments fedilink hide all child comments
 

I'm speaking broadly because I really don't want to start a fight here. It feels like many sides will never ever yield on anything and honestly, I've grown tired of the constant fighting

you are viewing a single comment's thread
view the rest of the comments
[–] 6 points 1 day ago (5 children)

Honestly it's what has made me a less frequent visitor in this forum. It's really tiring seeing so many posts about AI hype, with the only difference from here to Instagram is that it's Chinese AI hype. Not only aren't things necessarily good/socialist just because they're from China, but these are still just profit-seeking corporations being reported on by hype-seeking western news outlets, and people here still fall for it and reproduce it like they're your credulous AI doomer uncle. I don't have the time and energy to go digging through the hype for information to argue over random products from a corporations in a country whose language I can't read.

China is doing good tech work due to high speed rail and circuit lithography techniques, not because some random company made realistic looking humanoid bots that serve only to decorate presentations or luxury sex dolls for extremely rich pervs.

Worse yet is that even here people conflate chatbox LLM "AI" stuff with actually incredibly useful and reliable Data Mining techniques we used to call "IA".

  • permalink
  • fedilink
  • source
  • hideshow 5 child comments
    •
  • [–] 4 points 1 day ago* (last edited 23 hours ago) (1 child)

    AI if you are speaking in English, sir (Eu também misturo AI falando em português e IA falando em inglês lol).

    I think my hate against AI began far before LLMs. When I was finishing my BS degree, the hype in the academia (at least in my uni) was about bio-inspired AI algorithms. So there were the classics PSO, which is inspired in a flying bug swarm and ACO, which is ant-inspired. Then I saw articles about bee-inspired algorithms, my professor was about fish school inspired algorithms, then another of his students made a wolf pack inspired algorithm - and I was like, they are simply gaming the system. Because all studies followed a similar pattern, get a public dataset or a complex mathematical function, use this new algorithm to get the optimal responses, do some crap math to give as proof the algorithm would converge to some points, then do some experiments to benchmark your "novel" algorithm against some existing ones. I thought at that point people were working on those stuff not because they wanted to achieve anything, but because they wanted to game the publication system to get access to more research funds.

    And yes, I think LLMs have a far worse stench. Yes, LLMs are a new novel technique that allow us for the first time to generate stuff from simple commands. That said, most practical uses I've found for AI was to either make search slightly better because you don't have to deal with a bunch of sites trying to game (SEO) the search engines, or to generate output (images, code, or other assets) that follow a very specific set of patterns. So even it's a novel approach, the hype is talking about end of the world and apocalypse, so that its main objective is to just to inflate the current stock prices of AI companies, so they get access to more cash. So it's another money grab scheme. This time it's far worse, because investment companies are betting people's retirement funds, instead of simply getting a larger pie of your country's research budget.

    So it's a chaotic production anarchy, so commonly found in capitalism. And the calamity I'm waiting to happen is not when AGI comes, but when the AI bubble explodes and people start losing their jobs, their savings and retirement funds when that happens.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •
  • [–] 6 points 1 day ago

    AI if you are speaking in English, sir (Eu também misturo AI falando em português e IA falando em inglês lol).

    I'm going insane in both languages lol.

    Your natural computing anecdote felt very familiar as I noticed pretty much the exact same pattern when I used to do Deep Learning research. It feels like one of those Proof of Work crypto systems where a lot of human labour and energy is wasted on "publishing" bland accuracy-maxxing papers.

  • permalink
  • fedilink
  • source
  • parent
  • [–] 1 point 1 day ago (2 children)

    things [aren't] necessarily good/socialist just because they’re from China

    This. DeepSeek is just as good at spitting out liberal narratives as ChatGPT. I agree that the glazing on here is out of control - I got multiple downvotes and replies (one of which I'm pretty sure was LLM generated) on another thread for defending the use of the term "slop" to refer to chatbot output. It's honestly making me a little suspicious - apparently one of the mods here is big on AI hype.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •