Qwen 3 and Qwen 3 MoE models support merged in llama.cpp! 🙌🏽 github.com/ggml-org/llama…

Multi-tab Google Docs: 1. Great improvement, can compartmentalize data in a way that could not be done easily before 2. Great way to miss a bunch of information when not seeing there are, in fact, multiple tabs 😅

No one building a “TV show catch up” product? Would be useful for shows I stopped watching and am discouraged to pick back up because there’s bajillions of episodes to catch up with…

OH “- there’s no swag at this conference? - Llama IS the swag”🤣 #llamacon2025

Pretty cool synthetic data generator announced at #llamacon2025 today github.com/meta-llama/syn…

Really enjoyed EmTech AI this year and met great people. It’s interesting to see how the audience really diversified itself, with so many industries/trades being represented, thinking about leveraging or already using AI in their day-to-day.

👀

I’ve probably had to buy Minecraft more than 6-7 times since it came out (tablets, PCs, consoles, …) I must be doing something wrong 🥲

Fun read

x.com/i/article/1929…

🌟 Live healthier, feel better. 🏥 Omada Health delivers virtual care between doctor’s visits, helping many make lasting health changes on their own terms. 🎉 Keep striving to inspire, $OMDA!

Glad I finally received a M4 Max, need to try a bunch of cool models ✌🏽

We've been busy shipping and haven't been vocal about it, but we're another company that went through an IPO and was built on top of #RubyOnRails. thoughtbot.com/blog/70-compan…

I wonder how large the summarization model used by Apple Messages is. It’s either overly dramatic or just not on point 🤔

I miss Microsoft’s MFC :chefkiss: (admittedly I don’t remember enough, but I don’t have too bad memories of it)

Vibe-coded myself a CHIP-8 emulator in Rust/SDL a couple nights ago: blazing fast process. I even added a compiler / assembler for making games. All of the knowledge must be in the training datasets, so many examples out there. Now having Cursor creates games with the