70B models can’t run on a CPU with 32GB.
70B models can’t run on a CPU with 32GB.
Oooh! Model fight! I’ll try it out and post results later.
It would be like Winne the Pooh but with ice cream instead of honey.
I hit the computer it’s running on. This is not rocket science, people. The only thing an LLM can do is spit out characters to a terminal. It can’t kill you or make planes fly into buildings or build a robot army or launch nuclear weapons. It can’t do anything.
All these downvotes and not one counterexample. HOW can an LLM endanger anyone? Simple and serious question. I mean, someone start up a local instance of Llama and use it to start a fire or kill a child or something and prove me wrong here. You just hit ctrl-C and the LLM dies and people are acting like they’re Skynet.
no one wants to invest in turmoil and uncertainty
Elon Musk’s ears are burning right now.
Apparently that had nothing to do with it.
What do people think super-AI is going to do? All it can do is print letters on the screen. Flip a switch, it’s gone. It can’t actually DO anything; it has no body, no thumbs. The smartest AI conceivable can’t do a thing if I take a hammer to it.
What are people scared of???
Laugh all you want, but one Googler claimed their model began speaking a foreign language they never trained it on…
https://futurism.com/the-byte/google-ai-bengali
https://www.quantamagazine.org/the-unpredictable-abilities-emerging-from-large-ai-models-20230316/
I don’t know; I’ve encountered LLMs that pass my personal Turing Test and several Redditors who fail it…
That was open-sourced because of some sort of alt-right political reasons.