David (drbh) Holtz

Just built a bsky feed dedicated entirely to cats bsky.app/profile/davi... 🐱

Real-time, nonstop cat content straight from the bsky firehose—because we all need more cats in our lives. 🐈 🐈‍⬛

Still tweaking the model and testing other feed ideas, so stay tuned!

GitHub - drbh/bsky-feeder: 🦋 AI curated bksy feed generator (powered by MLX) 🦋 AI curated bksy feed generator (powered by MLX). Contribute to drbh/bsky-feeder development by creating an account on GitHub.

Thanks for sharing!! I’m excited to see all the cool things people build on the protocol.

For those interested in building custom AI powered feeds, I started a single-file proof of concept here github.com/drbh/bsky-fe... that would be pretty easy to extend!

Image of the ping pong kernel data movement over time

It’s entertaining/exciting to see GPUs discover async pipelining like more traditional CPUs

The Hopper TMA unit is a good example of this, it introduces async GEMM at the hardware level pytorch.org/blog/hopper-...

Along with newer hardware, are new fun kernel algorithms pytorch.org/blog/cutlass...

Frame from a short animation of sinusoid encodings

Fantastic post by @FL33TW00D that helps demystify positional encoding. A must read for all ML engineers fleetwood.dev/posts/you-co...

Big move by Replit 🔥 https://huggingface.co/replit/replit-code-v1-3b excited to see what niche apps this enables!

Wait, there’s other people named David?

You can play with them https://huggingface.co/spaces/multimodalart/ChatGLM-6B and https://huggingface.co/spaces/stabilityai/stablelm-tuned-alpha-chat

Really exciting to see more high quality OpenSource LLMs like https://github.com/Stability-AI/StableLM and https://github.com/THUDM/ChatGLM-6B

Out of the box they’re good enough to build a bunch of apps. It’s amazing how quickly these models became commodities, im excited to see what gets built!

Hello Bluesky! 👋