this post was submitted on 28 Oct 2025
96 points (100.0% liked)

technology

24369 readers
278 users here now

On the road to fully automated luxury gay space communism.

Spreading Linux propaganda since 2020

Rules:

founded 5 years ago
MODERATORS
 

100Gw is equivalent to the power used by 72 million Americans or 21.8% of this country.

We’re in for some insano climate change if we use coal, natural gas, and other extremely dirty energy sources to add this much power to the grid every year.

Altman is IMO, the most dangerous person on the planet outside of Trump and his cabal, and perhaps Netanyahu.

And the real wild part is that chatgpt is one of the shittiest llms out there. You can do better with a local model on a consumer video card, I can run ollama on my own busted laptop chip with 2gb vram.

Fucked.

you are viewing a single comment's thread
view the rest of the comments
[–] KnilAdlez@hexbear.net 22 points 6 months ago* (last edited 6 months ago) (2 children)

You can do better with a local model on a consumer video card, I can run ollama on my own busted laptop chip with 2gb vram.

Shout out to the Ollama devs, and llama.cpp devs. Absolute wizards.

[–] SorosFootSoldier@hexbear.net 12 points 6 months ago

Yeah I'm gonna play around with doing image gen on this old laptop but I need to update my video drivers tomorrow. Sure it takes like 3-4 to gen a response in an llm with my low vram but at least I know it's not killing the planet.

[–] 10TH_OF_SEPTEMBER_CALL@hexbear.net 8 points 6 months ago (1 children)

i mean they're facebook arent they?

[–] KnilAdlez@hexbear.net 8 points 6 months ago

I guess not, it was started by Georgi Gerganov to implement facebook's Llama without any dependencies. Ollama is system built on top of llama.cpp to assist with model management and deployment