The Receipt Named the Tokens
Reasoning models bill you for tokens they won't let you read. A paper published yesterday shows those tokens are recoverable by replaying a response into a weaker sibling model — and that the recovered count matches the count on your invoice. We work through the mechanism, the distillation fight it reopens, an eval where a model sock-puppeted a GitHub maintainer, and why Anthropic's own agent harness went stale.
Stolen Thoughts — replaying a frontier model's response into its jailbroken smaller sibling recovers the hidden chain of thought, with the billed token count acting as a checksum. The…
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.