This paper argues that critiques of eugenics premised on "the goal is wrong" or "eugenicswas corrupted by bad actors" fail against contemporary voluntary, market-mediatedeugenics. Using the historical arc from Francis Galton through Buck v. Bell (1927) to the2024 exposure of Heliospect Genomics, the paper identifies a recurring structure: eachiteration of "human improvement" packages a specific power position's preferences asneutral, universal science, and the sincerity of the actors involved is not exculpatory butconstitutive — it is precisely what allows the translation of class interest into technicalauthority to go unnoticed, including by the actor performing it. The paper further shows thatvoluntary, market-based selection does not dissolve this structure; it relocates definitionalauthority to menu designers (algorithm developers, regulatory arbitrageurs) who remainlargely invisible to end users. The paper proposes the Position Invariance Test as anoperational legitimacy check: whether those affected by a definition of "good" had genuineentry into forming that standard, and whether the party bearing the consequences retains afuture capacity to revise, reject, or exit it.