Sturgeon’s law was first printed in 1957. Consider how the publishing industry looked in 1957; in the intervening ~70 years, information has become much cheaper to produce, transmit, and obtain (though not necessarily to consume). This, I think Sturgeon’s “ninety percent” number needs to be increased, perhaps even to ninety-nine.

The strategy, memorably quoted in Hamilton, of reading “every treatise on the shelf” may have been a sound strategy when the high costs of publication strongly selected for the most high-quality pieces of writing. But certainly not anymore when anything can (and does) get published; we need other criteria.

Sturgeon’s “law” applies to information as a whole. But there are many pockets of information to which the law does not apply, and deliberately seeking out these pockets can be a way to increase the average quality of the information that you do consume. (I use the word “law” because this is how the phrase is best-known, though it’s probably more accurately described as “Sturgeon’s hot take”.)

In my opinion, the law is less likely to apply when there are multiple stages of careful gatekeeping between information production and information consumption.

Therefore, information is likely to be much higher quality when it has some of the following characteristics:

  • has passed through multiple stages of review (e.g. peer review or multiple editors, whether professional or informal)
  • length—while I certainly value brevity, it is also important for a good piece of writing to be comprehensive; a long-form piece of writing can examine a question in more detail and from more angles, giving you a more compete and critical picture
  • older/classic sources that were published long ago but are still considered high-quality (fiction only)
  • well-researched; comprehensive and written by somebody with a deep understanding of the relevant sources (non-fiction only)
  • strong recommendations by reliable external reviewers (recent, non-technical sources only)

Of course, my little blog has none of those characteristics!

But this isn’t always the case. In my experience, it’s easier to publish in peer-reviewed journals than it should be, and part of the (very complex and nuanced) reason is that many reviewers focus on superficial characteristics of writing (e.g. clarity of English, external markers of prestige held by the author, p-values)—while these characteristics can often be helpful, they certainly cannot compensate for a non-rigorous analysis.

And sometimes it is helpful and constructive to read low-quality information! e.g. I think it’s Dr Mithu Storoni in Hyperefficient who mentions an American president (Woodrow Wilson I think) who read pulp and mystery novels as a way to relax. Certainly I often feel this need for purely fun and low-stakes content in my own life. Another example is where low-quality sources include information that high-quality sources do not, e.g. using informal internet forums to gauge the opinion or views of specific communities.

You can also introduce additional gatekeeping yourself by simply writing down urges or inclinations to consume (or produce) particular pieces of information, then waiting a day or so to see if they still seem like a good idea. This is similar to the “scanner notebook” idea proposed by Barbara Sher in the book Refuse to Choose.