Recent research challenges the foundations of regulatory policy for pharmaceutical drugs and medical treatments in a novel way: rather than a single risky treatment, the regulator should prefer a menu of treatments with ambiguous risks (Viscusi and Zeckhauser 2015). Then, patients would have opportunity to try treatments, eventually settling on the one that works best for them. I examine this argument and offer three conclusions. (1) Patient heterogeneity – i.e. patients respond differently to a given treatment – creates the matching problem that motivates trial-and-switch strategies, and provides the ambiguity that drives the potential gains therefrom. However, trial and switch is not an unmixed blessing. (2) Ambiguity-seeking policy, over and above that provided by patient heterogeneity, would be accomplished by reducing sample sizes and/or replications in pre-approval testing. The mean level of acceptable risk could be maintained, but confidence limits would expand, increasing the risk to individual patients. In effect, this is just another proposal for less regulatory caution regarding treatment risks and more attention to risk-risk trade-offs, as suggested by the quality-adjusted life-years, QALY, framework. (3) The case for risk-neutral regulation of treatment safety should be taken seriously in cases of devastating and life-threatening afflictions. Otherwise it fails, most obviously in the case of treatments for relatively minor ailments, which treatments dominate shelf space at drug stores and advertising in the media.