I’m a lawyer from Spain, not a professional developer
But over the last few months I’ve been using AI agents to build projects that I genuinely wouldn’t have known how to approach before. This is CMMChat, my own AI client and one of the projects I’m currently building with
I keep adding examples and difficulty to the local Qwen 3.8 27B
Leave all the examples uploaded here: And I leave the videos in alphabetical order too. I think this time the one I liked the most and the one I had to fight the least with is Grok Build
My RTX 4090 just hit 181 tok/s on Qwen 3.8 27B model.
O Monday I thought 140 was the ceiling. I was wrong. mr_r0b0t built a different engine. I wanted to push the card again to see if I can squeeze more out of it. The headline number moved 30%, the like-for-like number moved
I ran Qwen 3.8 Flash Next on SSD and it's fast!
Using atomic_chat_hq AD-IQ4XS 85GB size Model weight to 24GB VRAM + 32GB RAM and the n-gram stream directly from SSD Result: 22 tok/sec (14 tok/sec @ 76k context) Output is still amazing since we're using 4-bit quant See
You have reached the end of the archive
All of qwen38