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.
Failure modes, misuse, interpretability and control, argued from evidence rather than from mood.
Anthropic has published three numbers describing how it oversees the thirty thousand agents working inside it, and invited other labs to publish theirs. None of the three can fall when the monitoring is failing.
Coverage, review latency and escalation rate cannot get worse when agent oversight gets worse, because two of them measure the plumbing and the third is a threshold the lab itself chooses.