1. Qwen3.7-Max Ran for 35 Hours on Unknown Hardware and Achieved a 10Γ— Speedup (firethering.com)

Alibaba gave Qwen3.7-Max a kernel optimization task on a hardware platform the model had never encountered before. No documentation or profiling data. No example kernels for the architecture. Just a task description, an existing implementation, and an evaluation script. The model ran for 35 hours. It made 1,158 tool calls. It wrote, compiled, profiled, and rewrote the kernel repeatedly, diagnosing failures, fixing bugs, identifying blocks, and redesigning the architecture multiple times without anyone watching. After 30 hours it was still finding meaningful improvements. The final result was a 10x speedup over the reference implementation. For context: GLM 5.1 ran the same task and reached 7.3x. Kimi K2.6 reached 5x. DeepSeek V4 Pro reached 3.3x. The models that stopped early did so because they issued no tool calls for five consecutive rounds, they concluded they couldn't make further progress and stopped. Qwen3.7-Max didn't stop.

8. Trivial Pursuits (www.lrb.co.uk)

Any scoring system can be abused once it gets captured. But the fact remains that a scoring system can also be the most...

20. Nendo's Wonderful Toru, an Electric Kettle for Alessi (www.core77.com)

Doing a lot with a little: Nendo's Toru, an electric kettle for Alessi, perfectly reconciles gesture with form. Taking advantage of Alessi's distinguishing metalworking techniques, which is the origin of the company, the stainless-steel body is shaped simply like a pitcher. The handle coming out from the top seems

Article Details

Browse by Date

Type to search HN and engineering blog articles
Press Escape to close  |  ↑↓ to navigate  |  Enter to open