Engineering leadership, from someone doing the job

A single lit desk in an otherwise dark office at night

Engineering

Your agent's approval gate is a SHOULD, and nobody read the capitals

Anthropic measured its own permission prompt, found 93% approved, called it approval fatigue and built something else. The MCP consent requirement everyone quotes is lowercase, and a gate with no timeout is an incident.

People working at desks in an open plan office

OpenAI fired contractors for using AI to train its AI

Contractors paid to rate ChatGPT were fired for using AI to do it. The tells they were caught by are the same ones that wrongly flag non-native English writers 61% of the time.

Eight identical grey capsules labelled what the dashboard sees, above eight scattered teal dots labelled what the runs actually did

Your AI observability is answering a question your agents do not fail on

Agent performance drops from 60% on a single run to 25% across eight runs of the same task. Nothing errors, nothing times out, every trace is green, and the answer is wrong.

See all