This stance is a bit confusing... When it comes to things like radioactivity we assume a linear no threshhold model (e.g. that lower concentrations still have

This stance is a bit confusing... When it comes to things like radioactivity we assume a linear no threshhold model (e.g. that lower concentrations still have effects, just our measuring tools aren't good enough to detect it) and spend billions as a result. Why wouldn't we do the same for flouride?