Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Facebook LLAMA is being openly distributed via torrents

github.com

Post reply on HN
Print view View original

631–640 of 719 posts

Re: Facebook LLAMA is being openly distributed via torrents

#631

Post by atvcatole »

sofixa wrote: post #240
bentcorner wrote: post #224

Earlier quoted context omitted.

I posit to a friend that: a) As these AI constructs become more advanced (especially around memory and personalization), we will eventually be able to treat them as people b) Some business will eventually sell an off-the-shelf product (hardware and/or software) that is an AI you can bring into your home, that you can treat as a friend, confidant and partner c) Someone will eventually lose their AI friend of many mont…

> a) As these AI constructs become more advanced (especially around memory and personalization), we will eventually be able to treat them as people There already planned products to "capture" someone's voice and personality to be able to continue experiencing "them" after their death? Shit is already weird. https://technode.global/2022/10/21/this-startup-allows-you-t...

The 2017 movie Marjorie Prime with Jon Hamm is about this topic.

https://www.youtube.com/watch?v=a7PtcOLJDco

Re: Facebook LLAMA is being openly distributed via torrents

#632

Post by skybrian »

landryraccoon wrote: post #614
skybrian wrote: post #469

Earlier quoted context omitted.

The way people informally talk about "passing a Turing test" is a weak test, but the original imitation game isn't if the players are skilled. It's not "acting like a human". It's more like playing the Werewolf party game. Alice and Bob want to communicate, but the bot is attempting to impersonate Bob. Can Alice authenticate Bob? This depends on what sort of shared secrets they have. Obviously, if they agreed ahead o…

So I need to ask the obvious question, why does it make sense to play this game “to win”? Throughout human history, humans have been making up shibboleths to distinguish the in group from the out group. You can use skin color, linguistic accents, favorite sports teams, religious dogma, and a million other criteria. But why? Why even start there? If we are on the verge of true general artificial intelligence, why woul…

I think of the Turing test as just another game, like chess or Go. It’s not a captcha or a citizenship test.

Making an AI that can beat good players would be a significant milestone. What sort of achievement is letting the AI win at a game, or winning against incompetent players? So of course you play to win. If you want to adjust the difficulty, change the rules giving one side or the other an advantage.

Re: Facebook LLAMA is being openly distributed via torrents

#633

Post by richardhod »

skybrian wrote: post #632
landryraccoon wrote: post #614

Earlier quoted context omitted.

So I need to ask the obvious question, why does it make sense to play this game “to win”? Throughout human history, humans have been making up shibboleths to distinguish the in group from the out group. You can use skin color, linguistic accents, favorite sports teams, religious dogma, and a million other criteria. But why? Why even start there? If we are on the verge of true general artificial intelligence, why woul…

I think of the Turing test as just another game, like chess or Go. It’s not a captcha or a citizenship test. Making an AI that can beat good players would be a significant milestone. What sort of achievement is letting the AI win at a game, or winning against incompetent players? So of course you play to win. If you want to adjust the difficulty, change the rules giving one side or the other an advantage.

I was confused by your first reply at first. I think that's because you are answering a different question from a number of other people. You're asking about the conditions under which and AI might fool people into thinking it was a human, whereas I think others are considering the conditions under which a human might consistently emotionally attach to an AI, even if the human doesn't really think it's real.

Re: Facebook LLAMA is being openly distributed via torrents

#634

Post by astrange »

Wingman4l7 wrote: post #587
crooked-v wrote: post #577

Earlier quoted context omitted.

Apple is the odd one out here, considering they've been making substantial efforts to move things on to consumer devices. For example, the ML-driven auto-categorization of pictures you keep in the Photos app happens on-device in the background whether or not you have any subscriptions.

I wouldn't give them too much credit. With the addition of the T2 "security" chip, your devices are bricks if they can't authenticate against your Apple ID. Combine that with them soldering together formerly modular components, and it's a very expensive lesson in you not owning your own hardware.

Is this supposed to be about activation lock? Phones don't contain a T2, that's a brand name for a component of Intel Macs.

And of course it's good for phones to make them worthless to steal, or else they'd be hard to use in public.

Re: Facebook LLAMA is being openly distributed via torrents

#635

Post by astrange »

javajosh wrote: post #615
thaumasiotes wrote: post #607

Earlier quoted context omitted.

That is an ephemeral change. It takes very little time for the biosphere to make a full recovery. You're talking about a small, brief, suppression of the biosphere. And you're calling it "destruction of the biosphere".

Yes, and when a forest burns down I call it the destruction of a forest even though it can grow back because that's how language is used.

Burning down is part of the natural lifecycle of many forests, and they actually suffer when modern land management stops natural fires.

Re: Facebook LLAMA is being openly distributed via torrents

#636

Post by astrange »

screye wrote: post #473
abandonliberty wrote: post #395

Earlier quoted context omitted.

I'm curious if the blocking of adult content has to do with moralism, commercial interests, or something deeper. An eager to please conversational partner who can generate endless content seems quite dangerous and addictive, especially when it crosses over into romantic areas. There's already posts of people spending entire days interacting with LLMs, using as their therapist, romantic partner, etc. Combined with fin…

Wall-E humans are going to be reality. That last century has already proven that humans cannot be expected to responsibly indulge in Gluttony, Sloth or Lust. Now these models can skip the material desires and trigger permanent hormone releases through perfectly personalized content. I genuinely fear that the breakdown of millennia old social structures that kept us human might lead to a temporary (century long) turmo…

I don't think you give humans much credit if you don't believe they have an infinite capacity to get bored by things. AIs can't produce endlessly compelling content because nothing can.

Re: Facebook LLAMA is being openly distributed via torrents

#638

Post by astrange »

q1w2 wrote: post #81
kif wrote: post #37

I wonder what the memory requirements would be to run such a large model. I'd love to be able to run this model, alas my MacBook can barely run toy models.

You would need over 65GB of RAM. There are consumer GPUs that have 48GB of RAM, and can be tethered together with NVLink. I wonder if that would work.

A Mac Studio should be able to do it since it has unified memory.

Re: Facebook LLAMA is being openly distributed via torrents

#639

Post by throwaway1851 »

ComplexSystems wrote: post #441

The smallest model (7B) is supposed to outperform GPT-3. Does anyone have any idea what hardware is needed to run this?

No, the 13B model outperforms GPT-3. Judging from the metrics published in the paper, it does look like the 7B model is not far off from GPT-3 however.

Re: Facebook LLAMA is being openly distributed via torrents

#640

Post by icrbow »

nerpderp82 wrote: post #312
brookst wrote: post #263

Earlier quoted context omitted.

Certainly they retain not just information but compute capacity in a way that other expensive transformations don’t. I’m hard pressed to think of another example where compute spend now can be banked and used to reduce compute requirements later. Rainbow tables maybe? But they’re much less general purpose.

HashLife seems like a scale free memoizer, https://en.wikipedia.org/wiki/Hashlife How Well Can DeepMind's AI Learn Physics? https://www.youtube.com/watch?v=2Bw5f4vYL98 https://arxiv.org/abs/2002.09405 https://sites.google.com/corp/view/learning-to-simulate/home Discovering Symbolic Models from Deep Learning (Physics) https://www.youtube.com/watch?v=HKJB0Bjo6tQ Scientific Machine Learning: Physics-Informed Neural Netw…

I don't know abut NN prediction, but apparently you can bootstrap anything* with strategically placed 15 gliders.

https://btm.qva.mybluehost.me/building-arbitrary-life-patter...

Post reply on HN