The desks Desk 01
Models, agents, research, and the products built around them.
Analysis
Vendep and FIRSTPICK are backing the Vilnius company’s enforcement agents. The opportunity lies in reducing casework, while platforms retain control over takedown outcomes.
Analysis
Move beyond a first manual eval with practices for test datasets, retrieval, agents, automated judges and production monitoring. Measure reliable outcomes, not only convincing answers.
Analysis
An eval turns “that answer looks good” into a test you can repeat. Build your first one with a spreadsheet, clear pass/fail rules and a fictional refund assistant, then choose tools when you need automation.
Analysis
Anthropic reports more merged pull requests and a lower revert rate. Its process combines automated reviews, risk-based human approval and recovery when checks miss a bug.