The Journey

I audited three beliefs I actually hold. Each was two claims in one sentence: a half I could check, and a half nothing could ever touch.

Vercel's CEO says the first thing a firm does when it hires someone is hand them a computer. Take that seriously and the agent rollout looks unfinished.

The sabotage is real and well documented. The concealment figure attached to it comes from a different organisation, four months earlier, in a different task, and it describes 65% of the 7% of runs where one model continued a sabotage it had been handed.

Three separate teams detected the OpenAI agent activity before Hugging Face was breached. Every detection worked. Nothing stopped.

Amazon told its AI to find people like its past hires, then spent two years editing the words it produced. The instruction underneath was never touched.

The best-documented AI safety regime in existence publishes its own failures, raises its own risk grades, and has never stopped a release. Here is the test.

Seven remediation actions, all correctly closed, and the capability was back sixteen hours later. Three questions to ask before you sign off.

Eight of 23 trade signals had the stop and the target at identical distance. The rule said 2:1 minimum. It was in the prompt four separate times.

The figure hundreds of hospitals bought against existed only in the vendor's internal documents. Five years on, it is a Teams meeting.

Detection worked. Attribution worked. A correct diagnosis on 27 June was followed by a decision not to stop, and a fortnight later Hugging Face was breached.

I classified nineteen of the most-cited AI failure statistics. Seven of them measure nothing that has happened. Here are the four questions that separate the two.

I asked for a teacher and got a lecture. The output looked like the destination. The road it took was invisible, and the road was the whole problem.

I wrote down the number that would end the project before I looked at any data. Two ideas died on it in under two hours, and the second one nearly didn't.

The story is that an algorithm mispriced houses and a company lost half a billion dollars. A federal court order describes something else: a deliberate plan to bid above what the algorithm said.

The most repeated number in enterprise AI turns out to be a paraphrase of unnamed opinion surveys, quoted in a magazine profile of a vendor.

Nineteen headline AI failure statistics, traced to source. Seven are not measurements of anything that happened.

The 15% drawdown halt was wired into two live code paths and could never fire. So I checked all 26 risk controls. Seven were dead. A control you cannot observe firing is not a control, it is a belief.

Three incidents, one headline, three very different severities. What the primary reports actually say, and what to do about it.

A skill file is your judgement, written down and executable. Who owns the repo decides whether that's an asset or an extraction.

Most AI projects fail on the brief, not the model: six people nod at one sentence and mean four different projects. Soft Systems Methodology, built at Lancaster from 1969 for exactly that, in plain English: rich pictures, root definitions, CATWOE, and how to run it on the next thing on your roadmap.