ShadowStats writes for a specific audience: people who distrust official statistics in general, and inflation statistics in particular. As you can see in this thread (and likely also by talking to most people you know personally), this is a sizable group, so it's likely a pretty good, sustainable business, just like health/diet fads or self-help books. The story does fall apart under scrutiny, but the susceptible audience is replenished faster than the claims can be debunked.
Amusingly, highly educated people who are disdainful of anti-vaxers or Gwyneth Paltrow quackery can easily believe that inflation is an order of magnitude higher than officially reported, or that they the evil/incompetent bureaucrats leave out food, shelter, healthcare and/or energy costs from the calculation or some such.
As an academic economist turned machine learning/SWE, I'd say there are plenty of knowledgeable people around on HN, but on matters of economics and health, the signal-to-noise is definitely much lower than on anything related to computers.
Agreed. It's why over the years I've come up with experiments like the above that I can give to capable yet deluded people to check things on their own. It amazes me how many otherwise smart people eat the huge amount of crappy, conspiratorial economic nonsense widespread on many forums, including unfortunately a lot of HN and Reddit threads.
It's why I love such clean examples as the Shadowstats nonsense. It's completely debunkable in an hours work.
Amusingly, highly educated people who are disdainful of anti-vaxers or Gwyneth Paltrow quackery can easily believe that inflation is an order of magnitude higher than officially reported, or that they the evil/incompetent bureaucrats leave out food, shelter, healthcare and/or energy costs from the calculation or some such.
As an academic economist turned machine learning/SWE, I'd say there are plenty of knowledgeable people around on HN, but on matters of economics and health, the signal-to-noise is definitely much lower than on anything related to computers.