Claude's watermark does little to reconstruct work shared across several tools. I want to know how a proposal became an accepted change. An estimate of model participation answers a much smaller question, and I will not use a detector to fill in the rest of the history.
On August 14, 2026, Anthropic explained the statistical watermark planned for text from future models. When Claude and Codex work on the same change, what I follow is the written task, the agent's diff, the other agent's review and the test evidence, all linked in the PR.
According to the company, the watermark adds no hidden characters and identifies no person, organization, or conversation. It helps estimate the probability of Claude's participation in a text. When another tool reviews a proposal, I need to understand what it received and what it checked. Finding a pattern in the prose does not recover those decisions.
With code, the gap is even easier to see. Knowing an explanation passed through Claude does not tell me why a line changed. It also says nothing about whether the reviewer checked the requested behavior or just improved the comment. I want the request connected to the diff and the evidence supporting acceptance. With that, I can challenge a particular decision instead of guessing how it happened.
Turning every ordinary message into an investigation of who used which tool would be tedious and offer little obvious benefit. For a change I will answer for, though, the origin of a decision can matter a lot. How much I record depends on that, and I do not invent identification the watermark does not contain.
When Claude and Codex collaborate, I prefer to keep assignments and reviews in a readable history. A statistical watermark can add its signal where relevant. Whoever accepts the result still needs to check the sources and expected behavior.