expert-judgment

该论文提出一种将专家判断蒸馏为可规模化自动评估系统的方法，构建专家网络并将其推理过程编码为结构化编辑标准，在AI新闻报道上验证了专家校准的自动评审显著优于未经校准的前沿模型。

A newsletter roundup covering Mira Murati's Thinking Machines achieving 29.8% fewer errors in finance judgments, a shift from Claude Code to Claude Tag, a cheaper method for feeding Fable 5 large context, surge pricing in AI with DeepSeek doubling peak prices, and Alibaba blocking Claude Code after a tracking experiment.

Thinking Machines AI describes training a proprietary LLM to replicate expert judgment in financial document filtering tasks, outperforming frontier models at lower cost.

Bridgewater partnered with Tinker API to fine-tune an LLM to replicate expert judgment in financial tasks, outperforming frontier models on information accuracy and recall at lower cost.