knowledge-distillation

Techniques for transferring capabilities from a large teacher model into a smaller student model, and the tradeoffs that come with it. Coverage spans what actually survives the process, from benchmark scores to calibration, robustness, and reasoning faithfulness, versus what quietly gets lost. Also examines the legal and contractual side, including how AI lab Terms of Service and "do not compete" clauses shape when and how distillation is allowed.

model-distillation, ai-training, anthropic, knowledge-distillation, opinion

Anthropic vs Alibaba: Why AI Distillation Isn't a Crime

Distillation isn't a cyberattack, and violating an AI lab's Terms of Service isn't a crime. But as the Anthropic vs. Alibaba dispute reveals, the broad "do not compete" clauses in AI user ToS could turn everyday software development into a legal minefield.

Glowing head dissolving into particles that rebuild a hollow empty copy, showing what knowledge distillation loses.

knowledge-distillation, tacit-knowledge, model-evaluation, ai-expertise, publishing-expertise

What Actually Gets Distilled: Copying the Method Is Not Copying the Judgment

Knowledge distillation reliably copies a model's outputs and benchmark scores while losing the properties that made them trustworthy: calibration, robustness, and reasoning faithfulness. The same split applies to human expertise absorbed by AI systems -- your explicit method transfers cleanly, but y