aichannel

Who is locally running an AI model? 

- Which model(s)?
- What’s your tech stack?
- Will you keep doing it?

Healthcare investor by day, web3/AI tinkerer by night. 🟪 Host: /venturecapital /theitinerant

- DeepSeek Coder v2 (8-bit, MLX))
- LM Studio on 2024 M3 MAX MBP (36GB RAM)
- Only just started, but after a weekend of fun with it, definitely!

✳︎ design engineer → @ponder 
✳︎ explorer of humanism • petrolhead/gearhead
✳︎ React/SwiftUI
✳︎ bento.me/matthew

I've heard there can be high latency. Whats been your experience?

DeepSeek R1 might have a bit of a lag because it's a reasoning model, but Coder v2 (instructional) is comparable to GPT-4 Turbo, so incredibly quick.