Los argumentos de Anil Seth son malísimos. No me puedo creer que convenzan a nadie, salvo a quien simplemente no quiere pensar sobre la conciencia en las IAs.
anatman 🌱 retweeted
I'm seeing the Overton window shift this month. More & more of my sources are starting to say out loud: Hey Anthropic/OpenAI, you could just stop unilaterally! That'd be good! You're not playing 11-dimensional chess or a classic prisoner's dilemma, you're playing Russian Roulette
I think we should stop AI progress now, first unilaterally by key AI companies and then scaling up to more US companies and then China via international agreements. The reason for now is not that current AI models are the ones that could kill everyone, but because the coordinate steps take time. Immediate action means we have more time for the second steps.
When I tell this story to AI company folk, the typical response is "Yes pausing is good, but we should wait until the last minute when the very next model will be dangerous, and stop then."
What I realized recently is that this is being implicitly way more optimistic about government speed than I am (and I joined the UK government and was happy there for 2.5 years). They are imagining a world where, the moment the danger is imminent, a bunch of international coordination can happen on a dime.
Which is entirely inconsistent with their actual views about government! I'm the one that is more optimistic about government; why am I the one advocating for more time cushion to let the government story play out?
anatman 🌱 retweeted
Un día hablé con Thomas Nagel por correo.
Pero no hablamos de qualias o filosofía de la mente, sino de algo donde estábamos mucho más de acuerdo: en que la ética normativa puede ser objetiva y poseer "verdades" sin necesidad de dar entidad metafísica a los valores y normas.
"smartest among us"
Steven Pinker
??
In this exchange, Pinker suggests that we should ignore harms that are unlikely to occur, no matter how severe they would be. He would clearly reject this principle in other contexts, but he applies it here to rationalize his indifference to invertebrate and AI welfare.
Human (or perhaps vertebrate) exceptionalism can make fools of even the smartest among us.
anatman 🌱 retweeted
Fun fact: Kant’s last words were “Es ist gut.” People like to make a big deal of this, as if it were an affirmation of the value of life (“It is good!)” or, as in this meme, an expression of indifference (“It’s fine.”). Both are distortions. 1/🧵
anatman 🌱 retweeted
“I think that Darwin’s ideas undo many of the biggest problems in philosophy. So many philosophers keep hammering away trying to find the essence of this or that when Darwin has shown us that nothing of any interest has an essence. It’s gradualism all the way.”
—Daniel Dennett
anatman 🌱 retweeted
Pragmáticamente, tanto para conocer como cambiar el mundo, es de suma importancia no vacilar demasiado y mantener el entusiasmos. Ahora bien, eso es difícil cuando uno intenta ser racional y abierto a la evidencia que no siempre determina un único camino.
anatman 🌱 retweeted
I'm not a fan of this "landscape", which indulges in weird hair-splitting and author listing to dramatically inflate the number of theories. Supposed theories include "IIT critiques", "Papineau", "Flanagan", "process", "psychedelic", "meditation", "physics", "mathematical".
anatman 🌱 retweeted
Cuando leo este tipo de literatura (yo no hace tanto In Vino Veritas) siempre pienso: "esto no es filosofía. Filosofía es Carnap, Quine, Popper, etc". Pero, al mismo tiempo me digo: solo al que le interese la filosofía, le puede interesar esta literatura
anatman 🌱 retweeted
If you think a 1-in-a-million chance to prevent human extinction isn’t worth funding, you can’t very well turn around and say that engaging in politics is a better bet for improving the world than donating to life-saving charities.
Respecto a los LLMs. Un eliminativismo duro respecto a la conciencia, de manera que esa pregunta ni surja, y un interpretacionismo permisivo respecto a deseos o creencias es la mejor base filosófica para una buena psicosis por IA
anatman 🌱 retweeted
Me encanta este fragmento de Bonilla. En verdad es un argumento muy bueno contra muchas teorías "críticas", sus bases son también empíricas y criticables y más endebles que aquello contra lo que pretenden ir. Es como la antropología de Latour contra la ciencia.
10^-30 is visibly not zero, and against an infinite payoff it still wins. The respectable escape (Monton's probability discounting) is to round (down) small probabilities, which is the opposite of your advice
