I learned to drive in the mid 90s and it was repeatedly hammered into my skull that drunk driving is extremely dangerous. That included in school testing on a table showing impairment as a function of weight and number of drinks. This was all standard as part of the licensing process in suburban NJ at the time.

No, it's not "you've been drinking" -> 20x more dangerous. Much idiocy has been put forth by people who are unable to understand basic layman's toxicology/pharmacology. Drugs have effects. More of the substance has more of an effect. Step functions where a little of a substance has zero effect, shifting to a major effect with a little more, are extremely rare. It is never the case, for example, that a dose of ionizing radiation goes from "not dangerous" to "dangerous" suddenly - we may measure a low dose at 5 cancers per 100,000 and a high dose at 5,000 cancers per 100,000, but there is always presumed to be some effect.

And yet we have media organizations saying things like:

" Washington (CNN) -- A common benchmark in the United States for determining when a driver is legally drunk is not doing enough to prevent alcohol-related crashes that kill about 10,000 people each year and should be made more restrictive, transportation safety investigators say.

The National Transportation Safety Board recommended on Tuesday that all 50 states adopt a blood-alcohol content (BAC) cutoff of 0.05 compared to the 0.08 standard on the books today and used by law enforcement and the courts to prosecute drunk driving.

The NTSB cited research that showed most drivers experience a decline in both cognitive and visual functions with a BAC of 0.05. "

Of course we have a decline in cognitive and visual functions - that's what a depressant does. At any dose. So long as we have drinking as a major societal institution, and we have bodies that slowly process alcohol, and we have an automotive-mobile culture, there is some nonzero number of deaths we will prefer to tolerate every year due to drunk driving, whether it's 1,000,000 or 10,000 or 100.

---

While there may be some distribution of how well people deal with a certain degree of drunkenness, the basic objective fact that we possess to measure impairment is BAC. Limits vary geographically and through history - in the US we have had experience with thresholds at 0.05%, 0.08%, 0.1%, and 0.15% in various eras and places.

A BAC of 0.01% doesn't significantly harm anyone - it is barely detectable. A BAC of 0.05% poses some minor statistical increase in danger, and is generally the minimum people seek out to 'get a buzz'. A BAC of 0.1% indicates moderate impairment - about what you thought, several times more dangerous. It's only when you get to a BAC of around 0.2% that it becomes 20x more dangerous. At a BAC of around 0.3% and up, on the other hand, one generally loses consciousness. Death from alcohol intoxication (assuming no complications) occurs at an average of about 0.45% BAC (that is the approximate LD50).