@unbowedheretic@egfalken 128024 for sure, the Unicode codepoint for elephant. Or 352271, the amount of them counted by the elephant census of 2013. Stuff like that.
Then you could have a bunch with weaker associations that suggested it when combined; like 37, how many Hannibal used crossing the alps.
@egfalken And "elephant" is kind of a standard imaginary animal to "think" about, viz. "don't think of an elephant", "seeing pink elephants", "elephant in the room" and such. That you got that in 25% of trials might just be random chance.
@egfalken Maybe one of the numbers has some sort of obvious elephant-connection that you'd see if you knew a bunch of random facts, like LLMs do?
E.g. if I asked you "Think of a supernatural being" and the list of numbers started with 666; saying "The devil" would not be an amazing feat.
@dundereloise A meaningful tagline/mission should narrow down the organisation in organisation-space. If you invert this, it should still make sense.
A company that provides products in a highly predictable/automatic way? Sure. A company that crushes the human spirit? Insane, so a platitude.
@SRamirez68083 Zero-sum means I profit from just moving demand from A to B, and whatever I spend doing that is deadweight loss.
And sure, maybe shorting won't happen in practice? I thought of stuff like funding NIMBY opposition though, not dumping toxic waste.
@ciphergoth@InnocentlyEris I had the same experience, they didn't accept my Swedish passport.
So I bought phenylephrine instead. It did nothing, and after a quick googling I discovered that indeed it's basically a placebo scam, so I got double disappointed.
@liron As always with questions like this, the fact that the person is asking it often means the answer is surprising in some way!
(though this time it tricked me!)
@bryan_caplan 🤓☝️ akshually the probability-weighed move towards or against the presenter should be zero. It could be more likely to move towards him, but then the expected size of the move against him should be bigger
@robertskmiles One has to make sure they're appropriately ominous without revealing too much however; timing it so it suddenly cuts off with a scream for example.
@lthlnkso Good as always.
Though consider fixing the baked subtitles if you can! It's distracting when the transcription is wrong, which happens quite a few times in the videos.
@vineettiruvadi It's completely possible for there to be procedures that a hospital does that brings in millions, but nevertheless doesn't work!
Surgical trials routinely find procedures that do no better than placebo, e.g. stents for stable coronary disease or spinal fusions for low back pain.
@ArthurB@ciphergoth@ingelramdecoucy@slatestarcodex I suppose it could be refined to instantiate all universes, but make everyone with net-negative lives p-zombies?
Or quantize all experiences, and turn people into p-zombies when each moment is negative, hmm...
@GarrettPetersen Like a lot of other things it's a signalling dynamic.
The vast majority of AI writing is crap, so it's rational to ignore it when you see the signs. Even worse, seeing signs of it means the author can't tell or was unaware that it'd have this effect.
it's like how Miriam Godwinson starts out being a generic 90s fundie, but the deeper into the tech tree you go the more she is the only one left making any sort of sense