Seconded. I'm 64, and today I learnt I had no idea what historians do when I read https://www.benlandautaylor.com/p/how-to-learn-history https://www.benlandaut

Seconded. I'm 64, and today I learnt I had no idea what historians do when I read https://www.benlandautaylor.com/p/how-to-learn-history https://www.benlandautaylor.com/p/how-to-learn-history I can't find an article from him where I can fault the logic or didn't learn something new. Perhaps some are a bit wordy, by my he does write readable words. Back to the topic on hand, I googled "How to learn history". Why? Because I now I have a fair idea of what the best introductory page on the subject is. To give Google it's due his post is the 15th result by my count, so it's discoverable if you are persistent. To give this article it's due, the top 14 hits are exactly as he described in this article - little better than SEO horseshit. It triggered a lightbulb moment for me - a theory on how Google ranks its results. Sites that specialise in serving higher grade horseshit on a wide variety of topics are in a hand curated list, and sites in that list are ranked first. Why do I think it's hand curated? Two reasons. Google keeps returning the same few sites. It's been puzzling me for a long while why sites Quora and wikiHow appear are at the top, when there invariably are far better resources out there. Reason one is I keep reading every metric Google used to automatically rank sites has been gamed into uselessness, so it has to be hand curated. The second is hand curated means the size must be manageable to a human. Given the shear number of topics (millions if not billions) the only way to keep it manageable is to restrict it to sites that serve a lot of topics. So www.benlandautaylor.com won't be in that list, despite being high quality. It doesn't contain enough topics. No one-horse blog would. Finally, I've come up with a plausible explanation of why sites that clearly are little better than SEO farms keep being ranked highly, even though many of them (like Quora) are almost always worse than useless. While they are worse than useless, being only a little worse than useless is a high bar on todays internet. That question has been driving me nuts, because I'm positive no search engine likes the situation. That said it's immediately obvious to me his pages were way better than even the highest grade of horseshit the search engines float to the top. If it's obvious to me, it will be obvious to an suitably trained AI. They are already far smarted than me in their domains. The CPU consumed by feeding a AI the entire internet regularly boggles the mind, but I guess that will change. Maybe there is some hope for internet search after all.