Satoshi 7B - a bitcoin-centric LLM is now open source \ stacker news ~bitcoin

pull down to refresh

Satoshi 7B - a bitcoin-centric LLM is now open source huggingface.co/LaierTwoLabsInc/Satoshi-7B

1691 sats \ 14 comments \ @Tony 13 Apr bitcoin

110 sats \ 5 replies \ @elvismercury 13 Apr

I suppose most here will consider this a giant triumph, but I'm not thrilled to imagine future civilization of warring AIs each with their own newspeak and definitions of thoughtcrime blathering infinite text into the universe, creating a new generation of human parrots who internalize their gospel.

Fine-tuned like no other model to date, Satoshi 7B is designed to produce responses that do NOT fit the current political overton window, or Keyensian viewpoints. We built a custom data-set from scratch, with a deep rooting in libertarian principles, Austrian economics and Bitcoin literature.

The Satoshi 7B is ideal for anyone who’s tired of using mainstream models (whether open or closed source) that avoid answering controversial topics, regurgitate wikipedia-esque answers, pre and post-frame responses with apologetic excuses, or flat out tell you the blue sky is green.

I suppose it was inevitable, and the reality fracture is going to get worse before it gets better. If it gets better. The only question is how to orient to it on an individual level.

115 sats \ 4 replies \ @Fabs 13 Apr

Possibly... On the other hand: If we issue each of the warring AI's a robot and pit them against each other in a futuristic arena, we could turn a negative into a positive by effectively putting Rome's gladiator fights into a new suit! Both literally as figuratively at that!

Maybe OpenAi should hire me, I think they could be interested in my ideas... 😃

205 sats \ 3 replies \ @elvismercury 13 Apr

On the other hand: If we issue each of the warring AI's a robot and pit them against each other in a futuristic arena,

You described what is literally going to happen, except "the arena" is the world, and the results of the 'combat' will be less readily observed. Think Cambridge Analytica times 1000.

Gonna be a fun election cycle.

27 sats \ 2 replies \ @Fabs 13 Apr

"Think Cambridge Analytica"

The one about the manipulated elections?

0 sats \ 1 reply \ @elvismercury 13 Apr

Yeah. Manipulated is in the eye of the beholder, but definitely a campaign spamming social media from a certain targeted perspective. To my knowledge, that was all human generated, even if much of it was probably automated. Imagine if basically no humans were needed to create propaganda that was 100x more compelling at global scale.

Well, very soon you won't need to imagine.

0 sats \ 0 replies \ @Fabs 13 Apr

Hm, pretty grim indeed.

10 sats \ 1 reply \ @BlokchainB 13 Apr

How do I add this to unleashed.chat?

60 sats \ 0 replies \ @PlebeiusG 13 Apr

I'm not sure you can use custom models in unleashed. It's whatever models @nvk, etc at setup on their end.

I'm actually working on how I can add this to my own chat site. Does anyone know if running custom models with Ollama is possible? I think you're also limited to the models that they've added to the Ollama library...

But, if anyone knows how to incorporate this into a LCEL LangChain thaaaat'd be great. Otherwise I'll ask the AI...

42 sats \ 3 replies \ @Tony OP 13 Apr

You can try it out here: https://www.app.spiritofsatoshi.ai/

0 sats \ 2 replies \ @joelaw 13 Apr

Authenticated with lightening address, but nothing happened. Still on sign-in page.