Many things are power laws, but many things are not. It's really easy to overfit your intuition here and assume it's all power laws, but that's an overcorrection from everyone previously assuming everything was a normal distribution.
The reality is that some stuff is power laws, some stuff is normal, some stuff is bimodal, etc. The more useful mental exercise is to try to learn what the actual distribution for something is, instead of assuming, and then try to dig in a bit more to figure out what causes might lead to that.
Power laws tend to arise when you have some amount of network effect and an iterative positive feedback loop. That shows up in popularity, wealth, etc. But things like, you know, height, are not a power law distribution.
Alcohol consumption does have some steep curve to that, but it's not clear if it's actually a power law. I assume its driven by both addiction times tolerance. In other words, if you couldn't build up a tolerance to alcohol, I'd expect something like a linear distribution reflecting levels of a addiction. But the more you drink, the higher your tolerance, so the more you need to drink, which leads to something more like quadratic or higher.
Right, so the interesting question to ask here is, around the biology of alcohol tolerance. How much alcohol does someone have to consume tolerance level X, and in turn how much more alcohol does someone at tolerance level X need to consume to reach the same level of inebriation as before?
At some point people become chemically dependent on alcohol so it's not necessarily an issue of inebriation. There's an upper limit based on blood alcohol concentration. Within the top 10% there are probably even more extreme drinkers than the average of 71 drinks a week.
The irony here is that these folks consume 60% of all alcoholic beverages. These are the whales of the industry. Makes you want to rethink the adult drinks market.
The reality is that some stuff is power laws, some stuff is normal, some stuff is bimodal, etc. The more useful mental exercise is to try to learn what the actual distribution for something is, instead of assuming, and then try to dig in a bit more to figure out what causes might lead to that.
Power laws tend to arise when you have some amount of network effect and an iterative positive feedback loop. That shows up in popularity, wealth, etc. But things like, you know, height, are not a power law distribution.
Alcohol consumption does have some steep curve to that, but it's not clear if it's actually a power law. I assume its driven by both addiction times tolerance. In other words, if you couldn't build up a tolerance to alcohol, I'd expect something like a linear distribution reflecting levels of a addiction. But the more you drink, the higher your tolerance, so the more you need to drink, which leads to something more like quadratic or higher.