AI is evolving beyond human input.
Alibaba's Qwen 3.8 Max demonstrates recursive self-improvement, potentially discovering new AI research. It has also mastered autonomous silicon design, a critical area where China aims to lead.
How i feel when i combine glm 5.2, qwen 3.8, deepseek v4 flash 0731 and ornith 397b across my DGX Sparks 5090s and 3090s .
Here, is a small example from qwen-3.8-27B and Kokoro TTS and Tone.js all on 1 RTX pro 6000
UNREAL. Qwen 3.8 27B is now happily running on an RTX 4060 locally with only 8GB of VRAM.
You’re looking at a 64k context window thanks to Unsloth’s fresh IQ4_XS quant, and it’s just 14.6GB on disk. Prefill is around 150 tok/s, decode sits at ~5 tok/s using native MTP. Only
You have reached the end of the archive
All of qwen38