The thing is that the damage only occurs in a specific instance, that is somewhat rare even if taken with alcohol. It occurs when the typical biochemical pathway for processing tylenol is saturated, which typically doesn't happen, but when it does it causes a lot of damage very quickly.
A lot of drugs are extremely dangerous when not taken as prescribed, but the danger exists only in edge cases. So you could have 99 people saying "this drug is fine" and the drug will kill that 100th person every single time because of a particular mutation in a biochemical pathway.
The Yelp method is a stupid way to regulate drugs because making correct risk assessments isn't something non-professionals can do. It's not like Amazon, where a buying a charger that mysteriously doesn't work with your phone is an inconvenience -- people are risking their health and their life.
The site would be a bad idea even the power to review was limited to MDs. MDs are not all equally qualified to review drugs. Even if they know a lot about their specialty, they don't understand everything they need to know to prescribe drugs safely. This is why we have pharmacists. Their job is to basically understand all of the ways drugs work to predict novel interactions. Dumping a bunch of mystery drugs on the market is going to destroy that safety net. Furthermore, it's insane to think you can properly measure the efficacy of a drug by having doctors write reviews about it. Double-blind studies are a necessity for a reason.
Or, for a less life-and-death example, the FDA shouldn't be overly concerned with sunscreen ingredients [0].
I don't think the FDA should go away, but I think there is room for improvement. Maybe we could have some harmonization with similar agencies in EU countries, automatically granting approval for drugs available there. Crap, maybe if we just did it for sunscreen it would be a win for consumers.
[0] http://cen.acs.org/articles/93/i20/Decade-FDA-Still-Wont-All...