AI Editorial

The rest of the library explains, and explanation has no opinion. This is where the opinion goes. Each piece starts from something that actually happened in AI — a model release, a paper, a benchmark result, a failure, a policy — explains the mechanism underneath it, and argues its way to a position: what genuinely changed, what is only being marketed as change, and what someone learning the field should take from it. Every piece links what it argues from and the concepts you need to follow it.

19 editorials 8 strands 114 cited sources 30,081 words

All Research & Technique4 Agents & Autonomy2 Data & Training1 Inference & Economics2 Evaluation & Evidence3 Safety & Alignment4 Open & Closed2 Learning & Work1

What AI changes about how people study, build, get hired and stay useful.

Learning & Work 5 October 2026 4 min read New

Only the whole job is for sale

Anthropic's new robot exposure index says machines can already do 34% of American working time and pay for themselves on 0.3% of job tasks. The gap is not a capability error. Exposure is counted one task at a time, and a robot is bought against a whole shift.

The 0.3% of job tasks where robots already pay for themselves measures how rarely a whole person's day is made of the one thing a machine can be bought to do, not what robots cannot do.

Read it → 5 sources robotstask exposureautomation economicscapital costs