Source Tweet > Anytime you are tempted to flatter yourself by proclaiming that a corporation or a country is as super and as dangerous as any entity can possibly get, remember that all the corporations and countries and the entire Earth circa 1980 could not have beaten Stockfish 15 at chess. >> Quote Tweet (Garett Jones) >> We have HSI-level technology differences between countries, and humans are obviously unaligned… yet the poorer countries haven’t been annihilated by the rich.
> (How can we know this for sure? Because it’s been tried at lower scale and found that humans aggregate very poorly at chess. See eg the game of Kasparov versus The World, which the world lost.)
> Why do I call this self-flattery? Because a corporation is not very much smarter than you, and you are proclaiming that this is as much smarter than you as anything can possibly get.
> 2 billion light years from here, by the Grabby Aliens estimate of the distance, there is a network of Dyson spheres covering a galaxy. And people on this planet are tossing around terms like “human superintelligence”. So yes, I call it self-flattery.
Rat measures of intelligence expanding brain meme:
(Edited to put some respect on a name)
God fucking dammit this guy still finds new ways infuriate and bother me. Like we all know he’s full of shit but I can’t stop picking at this sore on the intellectual body of humanity so, and only from the exceprts in the OP:
Yud: “My God Is Greater, Your False Gods Have Nothing On My Totally Real Thing Because:
I feel like someone needs to challenge Yud to a game of chess with really high stakes, then show up and beat the fuck out of him and refuse to pay him a cent, then a neutral third party asks him who was the smarter person in that encounter. The obsessive, honestly pathetic, return to chess as a general reasoning task, despite the decades of discussion about how it is not that, is just fantastically myopic.
Take, for example, a hypothetical match between Stockfish and “the whole world” for the stakes of “whatever you want.” There’s at least one person in the world who thinks to unplug the machine, thus ensuring that the world is entirely capable of defeating Stockfish.
And just to be petty, I wanna jump on the tweet that’s starting things off: tech and capability differences between countries are a question of resources, not intelligence, and the resources issue feeds back into intelligence by incentivizing movement to power centers. These people (Yud and company) have done such violence to their conceptions of history and geopolitics that the notion of colonialism doesn’t enter into their visualization of the world at all, despite it’s reasoning power and material support.
Like, buddy, have you been to a nation that was under heavy colonial rule? They were annihilated, in many cases down to their history and culture, and in other literally genocided. Not that it really alters the reasoning for Yud’s view of AGI being destructive if not Guided By His Reflected Brilliance, but it’s resource and capability differences, not some absurd abstract notion of intelligence differentials, that drive the possibility of destruction when unequal partners interact.
For a guy who talks about intelligence a lot, Yud has not thought about it very much. No, a corporation will not outplay you in chess, but it will unify the actions and labor of thousands of people for years on end. Again, it is not some weird Goku vs. Superman, powerscaling from standardized test results that’s the important factor, but the ability to perform tasks that are difficult or impossible for humans (i.e. focus on tasks for periods of time longer than a human lifespan, etc.).
Do we think he kinda knows he’s full of shit and these kinds of lines are sort of tells?
I did better at Pokémon than Twitch plays Pokémon, therefore, I am smarter than the whole world.
A country could have easily beaten Stockfish 15 in ways Stockfish couldnt. Countries have Tomahawks, servers have fans.
(E: I point this out because his monofocus on intelligence is missing that corporations and countries have a lot of other capabilities, Kasparov versus The World would end differently if everybody on the world gets a board because Kasparov would run out of time making moves (unless you are removing the time constraint but then you are stacking the deck for Kasparov (I know the played gamed worked differently btw, but 1 domain expert vs people voting on results is also a bit weird)), note he shifts from intelligence only to intelligence plus capabilities when he talks about the grabby aliens (??) who have dyson spheres covering the galaxy).
(The colonialism has not annihilated poor countries person also sucks btw, like wtf kind of counterargument is that).
The funny thing about Kasparov vs The World is that according to Kasparov himself it should have been a draw, he just had an unfair advantage because he was able to predict The World’s strategy by reading chess forums. Even then, The World almost played Kasparov to a draw before they made an error in the endgame. This means that humans actually do aggregate very well at chess, given that no player for The World was within 200 points of Kasparov’s Elo rating and the most prominent strategist was over 400 points below. I’m fairly certain that the average person wouldn’t be able to play the second-best chess player of all time (who was best of all time when the game was played) to a draw, especially when that player could also functionally read their mind. So this proves you actually do get better at chess by aggregation. Of course anyone who knows anything about chess could have told Yudkowksy this (it’s why masters review their strategies with other masters before tournaments)l which makes me wonder how much he actually knows about chess in the first place.
this whole tweet thread is dumber than I ever could have imagined
I must be missing something, why is he comparing “all the corporations and countries and the entire Earth circa 1980” to a program that didn’t exist in 1980? Is he saying that there’s something different about all the corporations and countries and the entire Earth today that we’d do a lot better against Stockfish or something?
Does he outsource his opinions to an ai trained on the people who follow him
A bit of a learns question here, but back during earlier days Covid, I read Cheryl Misak’s biography of Frank Ramsey. I think Ramsey was the first to think about assessing our degrees of belief in terms of the bets we would be willing to take on certain beliefs/outcomes.
He strikes me as the opposite of a giant idiot, but there seems to be a real connection between Ramsey’s thinking and the LessWronger’s interest in betting markets, etc.
Do we just hate the obsession with betting markets? Or are they very different from Ramey’s thinking? I’m inclined to find the idea of assessing someone’s belief by looking at the odds they would take to be kinda cool and maybe helpful.
Once again the grabby aliens.
On the one hand, extremely arrogant of him, but on the other hand, a country that elected Donald Trump isn’t in that good of a position to say it’s smarter than anyone, Yudkowsky or not…