32
Please ELI5 the arguments people use to claim it's possible to use AI/LLMs ethically, locally, openly etc. (piefed.social)
submitted 5 hours ago by to c/asklemmy@lemmy.world
35 comments fedilink hide all child comments
 

I usually don't even understand the lingo they use. "Open-weighted" is the most recent one, then it usually goes down to specific "models" that everybody is supposed to know about.

These are my thoughts (I will stick to the vague "it" for now, but of course therein lies another question: "and how does all this apply to various specialised AIs"):

  • Is it really feasible to run it 100% locally? I know there's plenty of people with very powerful rigs indeed, but still. Or are 99% of these people really saying "it would, in theory, be possible to run that locally, therefore your concerns are invalid"?
  • If yes to the previous: the software doesn't come from nowhere and ultimately still relies on gas-turbine-powered datacenters and stolen IP and stolen personal data, no?

If what I wrote above is true, what exactly are people arguing when they say it's still possible to use LLMs ethically or true to FOSS philosophy, because ... ???

you are viewing a single comment's thread
view the rest of the comments
[–] 3 points 4 hours ago (1 child)

They call it open-weighted and not open-source because they don't have the training material (stolen books lol) but still want to pretend they are l33t hackers not bound to BigTech.

Yes, it's possible to use it locally with a $2000 computer (that's for the cheapest ones) but you'll only get a few words per second. It doesn't matter to the vibe coders who don't know how to code.

The software to run that is open-source but the training of the model (the "open-weight" black box) requires to destroy the environment at least once.

Last but not least the free models are obviously censored but people don't care about censorship anymore for some reason. "Tiananmen didn't happen? Not my problem" without understanding that more is hidden.

Anyway, no, there is nothing ethical about it.

  • permalink
  • fedilink
  • source
  • hideshow 1 child comment
    •
  • [–] 10 points 3 hours ago* (last edited 3 hours ago)

    a) Yes, today it costs 2000$ because of the cost explosion - my setup cost me around 1400$, and my GPU was already bought when prices were rising. No, i do not get a few words per second, i get around 50-60 token/s, which is more than enough for personal use. (Edit: and that is WITH CPU offloading, where layers that don't fit into VRAM get placed into system RAM, and i am still running DDR4 to boot)

    b) If the completely insane AI corpos would stop training humongous models to chase after non-achievable AGI, the one-time investments would have paid off by now for local use. This situation has nothing to do with local models but insane billionaires and execs.

    c) go ahead and lookup abliterated (not a typo) models on HuggingFace - these do NOT refuse any requests, because that has been pruned out. These answer everything about Tiananmen, DEI topics like erosion of LGBT and womens rights in the west and whatever atrocities any group might have commited, while also telling you about whatever you want to know. I run only these models, because i refuse to partake in censorship.

    Edit: Regarding the training material: This training material also consists of MY output over the decades on the web. Therefor, i do not have qualms using these models for non-commercial usage without disseminating the output - classic personal usage, mainly for automating tasks that are not easily done by hand such as grabbing game descriptions from steam, condensing them down to a short sentence and putting the result in a database, together with the corresponding steam tags, and automatically searching the web for information about the game when there is no steam page for it - which is an insane amount of work to do by hand for my library of more than 30k titles. I let it work during off hours on that task.

  • permalink
  • fedilink
  • source
  • parent