Evan Rodgers 🏳️‍🌈

Introducing Arete-27B: Qwen3.8-27B, post-trained to reinforce Western values. Chinese open models are currently the strongest available, but they ship with state-aligned behavior in the weights. We built an 803-prompt eval across seven value categories, including freedom of expression, democracy, and privacy, and measured how often a model adopts China’s party line. After fine-tuning on roughly 53 million tokens, Arete-27B's propaganda rate fell from 22% to 1%, in line with the Western models we tested. Its Hermes Bench score shows no significant performance tradeoffs. The values in an open model are one fine-tune away. Our eval, weights, and training data are available now along with the full blog post: nousresearch.com/arete

A lot of people are mocking "Western values" without reading the blog post. What Nous actually did is simple: They took a Qwen3.8-27B that gives CCP talking points on sensitive topics, fine-tuned it on ~50k human-checked examples, and that behavior went away. CCP propaganda answers went 22% to 1%. Please read before posting.

JUST IN: Micron pauses $4.5 billion Taiwan chip plant

Absolute slaughter. Double digit losses for many AI companies. SpaceX 10% below IPO price, CoreWeave about to drop below its IPO price. Nvidia is down almost 10% The AI industry’s day of reckoning may have finally come.

The full run cost $1,213,430 in API credits—about $1,279 per distinct local privilege escalation vulnerability. For context, applying historical Pwn2Own and TyphoonPWN bounty rates gives a hypothetical value of $18.9M–$66.3M. This demonstrates the relative efficacy of using AI models for securing software—defenders should proactively analyze software to make cyber attacks cost-prohibitive.