Anthropic's Operon agent arrived into the AI science conversation as a product reveal, and the community received it that way [19]. The framing — a private environment for biological research, built on Claude Desktop — invited celebration of capability. What it did not invite was the question of whether the infrastructure required to make that capability useful actually exists for the scientists it is nominally built for. Announcement cycles in AI science consistently perform this substitution: they treat the model's capability as the binding constraint, and once that constraint is…