Posts
Comments
I know I'm writing in 2025 but this is the first Codex piece I didn't like. People don't know about or like AI experts so they ignore them like all us rationalists ignore astrology experts. There's no fallacy. There's a crisis in expert trust, let's not try to conflate that with people's inability to distinguish between 1% and 5% chances.
Reminds me of the Tolkien cosmology including the inexplicable Tom Bombadil. Human intuition on your conjecture is varied. I vote it's false - seems like if the universe has enough chances to do something coincidental it'll get lucky eventually. I feel that force is stronger than the ability to find an even better contextualized explanation.
I almost think it's a problem you included the word "mainstream." It's a slippery word that winds up meaning "other people's news." It seems like realizing the point in your post is one step, and taking a more surgical dive into what news counts as obscure enough is another. If you're a doomscroller you're probably gravitating toward stuff many people have been hearing about, though.
The "semantic bounty" fallacy occurs when you argue semantics, and you think that if you win an argument that X counts as Y, your interlocutor automatically gives up all the properties of Y as a bounty.
What actually happens is: your interlocutor may yield that X technically counts as Y, but since it's a borderline example of Y, most of Y doesn't apply to it. Unfortunately, as the argument gets longer, you may feel you deserve a bigger bounty if you win, when really your interlocutor is revealing to your their P(X is not Y) is quite high, and if they do yield, it's more likely they're yielding that X is a borderline Y.
This is a "leaky generalizations" and more specifically a "noncentral" fallacy. It happens, for instance, when someone tries to prove something is racist so as to imply behavior should change and gets resistance. In fact, you might consider that gap between "technically a Y" and "typically a Y" as a sort of semantic deficit. If the point of something being in Y is that it's bad, consider arguing that it's bad without even bringing up Y.
This applies even to some of the most drastic, moralized words we have, like "slavery," "genocide" and "fascism." However you feel about any issue in one of those topics, I will inform you that proving to someone that that thing is in that category, is not going to have the effect you want. There is no semantic bounty.