the archive 17 drops

← back to the feed

month

filtering by

Less restrictive is now a model spec

The FCA put the model vendor inside the sandbox

Agent procurement is collapsing into deployment latency

The AI safety index grades disclosure, not safety

The "AI employee" title is a governance bug

Agent approvals need independent policy checks

Agent-built apps need expiry dates

Preference data governance outlives training methods

Data gravity is becoming agent governance

Bounded agents beat executive titles

AI research needs evidence lineage

MCP needs revocation before reach

page 1 of 2