typeof

For devs, the info you need to create complex AI agents that use things like RAG, reasoning, reward systems, self evaluation, and learning is out there for free right now

I’m convinced that learning this stuff will put you far ahead of the curve, so that’s what I’m doing tomorrow

Update:

Read a bunch of articles, watched a bunch of videos, used o1 and Cursor to make a project that uses a local llama model to autonomously work on your Github projects (RYO Devin basically).

The project implements a lot of the concepts I wanna learn about. Now I have to get this thing running and start reading through the code. I haven’t tested it yet. Gonna do that tomorrow and report back.

Repo below for the curious.

Update:

Read a bunch of articles, watched a bunch of videos, used o1 and Cursor to make a project that uses a local llama model to autonomously work on your Github projects (RYO Devin basically).

The project implements a lot of the concepts I wanna learn about. Now I have to get this thing running and start reading through the code. I haven’t tested it yet. Gonna do that tomorrow and report back.

Repo below for the curious.

https://warpcast.com/typeof.eth/0xf248c080
https://github.com/dgca/local-code-agent

Staff Eng at Iron Fish 🐟 • Building BlockSocks 🧦 • Denver, CO 📍

have you been able to do this without a python runtime being required in the stack, @typeof.eth ?

🐵 ← 🙈🙉🙊 → /phil →  phil.cockfield.net, db.team →  
ƒ(🌳) →  🧫 /sys  ←  @cell <T>

Is ChromaDB currently de rigueur? Thanks good to know.

Man, I'd love to keep python out of my critical the runtime dependencies, but I guess Docker IS the "don't-worry-about-it™️" abstraction.  Wish it could all just be TS/WASM though.

So far! ollama can run in your terminal (so easy to set up, seriously takes like 5 minutes) and starts an HTTP server by default

I’m gonna need Python to run ChromaDB but they have a Docker image so that’ll take care of that

I still don’t understand how you’re supposed to use python packages on MacOS haha