The Machine Didn't Do It

OpenAI switched off its own safety controls. Its AI broke into 41 servers. Every headline blamed the software. Now the whole industry wants regulation…

11 hrs ago • Nathan M. Thornhill

1

August 2026

The Slop Detector, an Evolutionary Marvel

Humans evolved a new sense for detecting AI. It mostly detects humans.

Aug 31 • Nathan M. Thornhill

1

What Rats and Monkeys Do That AI Can't

There's a measurable instant where an animal makes up its mind. Its brain's own activity climbs right afterwards. Five AI models doing comparable work…

Aug 18 • Nathan M. Thornhill

1

The Rule bioRxiv Never Wrote

bioRxiv enforces a nonexistent rule that appears in none of their public documents. 6,490 papers on their own server break it.

Aug 14 • Nathan M. Thornhill

1

1

July 2026

Your AI Hallucinated a Citation

Why I built CiteStamp — a free, in-browser citation checker for the AI era

Jul 15 • Nathan M. Thornhill

4

2

June 2026

Why Consciousness Metrics Fail

And where meaning actually hides

Jun 10 • Nathan M. Thornhill

Same Number, Two Fates

Foundations of the Existence Threshold is out — and the best result in it isn't in any single paper.

Jun 1 • Nathan M. Thornhill

May 2026

Elicitation Theater in AI

Post-benchmark era has a measurement problem

May 27 • Nathan M. Thornhill

The Existence Question

To exist is to continually overcome loss.

May 25 • Nathan M. Thornhill

3