Two essays, same day, same verdict. Ben Werdmuller’s “China’s open-weights AI strategy is winning” pulled 1,111 points and 839 comments on Hacker News. Ben Thompson’s “Who’s Afraid of Chinese Models?” on Stratechery took 642 points and 443 comments. Combined: 1,700+ points, 1,280+ comments, the loudest AI conversation of the day.
What the two essays actually argue
No product here — this is opinion, and the problem it names is distribution. Werdmuller’s case: US export controls stopped Chinese labs from serving the world, not from training models. So they ship weights instead, and open infrastructure wins because anyone can build on it without asking. He cites research that roughly 80% of startups are touching a Chinese model somewhere.
Thompson arrives at the same place from the other side. Frontier labs will be fine, he says, but America needs open alternatives — and Washington should make training data explicitly fair use and void terms of service that forbid distillation.
Why it broke through this week
Timing. Fable 5 went dark for 19 days. GPT-5.6 shipped to about 20 government-approved partners, vetted customer by customer. In the same window Moonshot put Kimi K3 weights out and Alibaba previewed Qwen3.8-Max at 2.4 trillion parameters, calling it “second only to Fable 5,” with open weights promised and no date attached.
Closed models buy a lead. Open weights buy distribution. 2026 is when we find out which one compounds faster.
You Might Also Like
- Alibabas Qwen 3 8 Qwen3 8 max Preview Claims its Second Only to Fable 5 With Zero Benchmarks Published
- Pollen Robotics Reachy Mini a 299 Desktop Humanoid That Runs 1 7m Hugging Face Models
- Title x Square Robot Wall a With Wall oss 276m From Xiaomi 62 dof Weights on Hugging Face
- Openai gpt Realtime 2 Translate Whisper Three Voice Models one api Several Startups Erased
- Ornith 1 0 Deepreinforce Self Scaffolding Coding Models Open Weights That Write Their own rl Scaffold 397b Hits 82 4 on swe Bench
Top AI Product 
Leave a comment