A place for developers to talk about building on Farcaster.

/dev

I just tried Llama 3 today, and I'm amazed its able to run locally on a regular Mac without internet access. The future looks incredibly bright!

I just tried Llama 3 today, and I'm amazed its able to run locally on a regular Mac without internet access. The future looks incredibly bright!

https://llama.meta.com/llama3/

Engineer @farcaster, co-founder @bountycaster. Sharing what I learn about AI + development. Previously: Phantom, 0x. NYC 🗽

Privacy is also a huge unlock of Llama 3! It allows you to utilize large language models in a sandboxed environment.

Real world use cases:
- sensitive medical data
- legal documents
- therapist
- legal council
- corporate codebases

Good question! I rely on ChatGPT as my go-to tool because it integrates seamlessly with other technologies I use, such as DALLE and Whisper. I likely won't change my flow

Llama 3 introduces new use cases that are not possible on a centralized, internet-dependent model

will try it now. 

do you see it replacing chatGPT or others as a primary LLM for you anytime soon?

my experience with llama 2 was that, it could compete with with chatGPT 3.5 but no where close to what we expect these days

Building /neynar 🪐

We’re hiring staff software engineers, check out neynar.com/careers

I haven't yet. I'll definitely try it out soon!

Nothing in particular for now, I still use a pretty old laptop so the tokens-per-second is really slow. Will continue going through cloud for now (ChatGPT)

Have you tried it? what do you think?

Crazy! What're you planning to use it for?

What are the specs of your regular mac? I'm pre-m1. Haven't tried it yet, but most of the good Ollama models have a pretty low t/s. All this local LLM magic makes me want to upgrade.