Tell it to use p5.js and p5.brush.js and give it a character and style reference. Also creative commons sound libraries!

Show Post

Better model than Qwen 3.8 27B you say? > Look inside > Bechmaxxed MoE with more total parameters than Qwen. I highly doubt it. Unless you are extremely careful MoE will destroy your model’s generalization ability; I’d be intrigued if it was better at equal total parameters.

Show Post

Ever think about how both apple and nvidia luckily both happened to be sitting on tech that slotted easily right in to LLM when it hit. With Apple though it's never completely luck. The Mx pro->ultra have unified memory and incredible memory throughput and are fully working personal ai machines

Show Post

Flint: a brand new, from scratch, small LLM that’s a heck of a lot faster than Qwen3.8-27B and basically as good recursive.com/updates/intr...

Show Post

Every family chat with a tech worker looks like “😱😱😱😱 just update you’re software things are fine 😱😱😱🧑‍💻🔫 😭😭😭 no you don’t need to panic”

Show Post