[–] 3 points 8 hours ago* (last edited 8 hours ago)

So their idea of open weight models is just unrestricted (less regulaton) models that they still host, and they will still collect all your data. They still need to build data centers. Money is getting made regardless of if it hurts the economy, if they are needed/wanted, and if its even profitable. Because eventually the american people will foot the bill.

What they don't like is just a model that you can download and run offline anytime you want. If they do its going to be baked into windows as a monthly subscription to use the service, then you select from a drop down list of models to use for whatever task. They can bake a defussion model into Microsoft paint or whatever. You can do the same with krita/photoshop via plugins. So they will charge you a monthly fee to use your own hardware/compute.

The only reason they are pulling this now is because everytime they release some "big crazy disruptive" ai model the wind is taken out of their sails by the Chinese models. Sure you can argue no one has 15k for a PC with 128gigs ram/vram to run some of these models at their full capability, but you can quantize and retrain the models to work on less parameters. So you could work get a model to run on a 8gig GPU which is pretty common and if trained properly it can be decent at specific tasks.

They want to reposition themselves now that its obvious no one wants to generate slop with sora/chatgpt and it will never turn a profit. They are trying to reframe themselves as being nessisary for the greater good and some how if its a "open" model that its fine. But much like how googles android is "open" there are always ways to exert control with it to the point its locked down. For the model to truely be open, its even less profitable to these companies like openai. Nvidia can at least profit off people running local Ai models.

Also its in talks that we might nationalize Ai which is part of this reframing and their way to get more access to money, and power. We are already having a hard time not letting these Ai companies not absorb all of our goverment data, if we nationalize them they are just going to get all that data, and start removing welfare on a weaponized scale. We won't have to bail them out because it will just become this infinitly inflating budget worse than the mitary budget. And in the name of "socalisim, and effiecency" they'll just end up cutting all welfare programs including social security, food, healthcare for UBI. And UBI will be something like 600 dollars a month so you will be in a worse off position, and receive less benifits overall, but all that money the goverment saves will go straight into the Ai budget.

  • source
  • [–] 11 points 2 days ago*

    You have been flagged now. Ai will comb over your account a lot harder compared to other accounts. Because you are already flagged, reddits ai mods are going to try to construe anything you say as a potential call for violence.

    For instance you could be quoting a movie or a song that might sound violent outside of context, and because you are flagged it's going to assume the violent intentions, while other accounts can say the same thing and not get banned.

  • source
  • [–] 11 points 2 days ago*

    Enshittifying the internet to the point a llm is almost the only way to do a google search is a violation of first amendment rights.

    Buying exclusive rights to reddits data, and making it harder to access with out an account with a ID linked to it, is violation of the first amendment rights. I can google a question and 90% of the time the result comes from reddit, which I can no longer access because I am not giving them my ID and I'm not turning off my VPN

    LLMs want to act as the DRM for what used to be free information anyone can access. Now they want to be the arbitors of truth, and the curators of what information you should see.

    Thats a violation of first amendment rights.

  • source
  • parent
  • context
  • [–] 9 points 4 days ago

    Probably because google can be held liable for lies that gemeni shits out in search results now.

    And 99% of googles searches came from reddit.

    And 99% of reddit is just bots and ai slop now.

    Not to mention reddit was the best platform to create a lie that google presents as fact. The subreddit trueanon convinced google that brace beldon could dunk on a regulation basketball hoop for months.

  • source
  • [–] 41 points 5 days ago (1 child)

    And of course tim Sweeny is throwing a tantrum because his biggest reveal lately is his entire game engine is gonna be ai slop.

    Meaning people are going to call it out on steam and its going to hurt Tims sales.

    Won't someone please think of tim Sweeny

  • source
  • [–] 9 points 6 days ago* (last edited 6 days ago)

    Of course France would be the first. They're the most known to dthrow revolutions.

    Doesnt help the whole world is experiencing a gen z anti authoritarian shift.

    Even in India right now they are fighting against their goverment. The whole world needs Israel's spyware to protect them and genocide anyone who stands up to them. They need you to ID your online identity so they can Ai filter our your speech and send you to prison for life.

    Makes sense why trump is labeling anyone that dares be critical of him or his policies as a "Cuban terrorist"

  • source
  • [–] 30 points 6 days ago

    If they had to pay out royalties, or pay licensing fees to everyone they stole from, my whole perspectice on LLMs would be different.

    Or at least I would be less inclined to make some of the criticisms I have on it.

  • source
  • parent
  • context
  • [–] 1 point 6 days ago*

    Please stop supporting UBI. Even when progressives talk about UBI they talk about how much money they will save because they are removing literally ALL other forms of social welfare including social security.

    Imagine having to work at Walmart like your grandma at 80, because the social security check only pays out a weeks worth of pay at a part time minimum wage job. And now due to UBI you don't even get food stamps, or get free healthcare.

    We need more socialist programs, not a cash bribe for the stupidst of people to sell out their country.

    This is what ai/effeicency looks like. Its going to be like dodge gutting all the funding for the federal programs until they are nonfunctional. All the money should be going to ai and "hyperacalers" instead and people starving on the street is just social progress.

  • source
  •  

    Never I mean never stop bitching on any online platform about this stuff. One of the few positives Ai gives to the public is that shareholders are now immediately effected when the Ai summary of the internet is collective outrage.

    This will never stop happening so never stop bitching. Make sure anytime Ai is mentioned the only thing you can see in the comment section is memes or jokes about how shit the products are or how evil the ceos are.

    view more: next ›