I was excited by mako’s vision of “cohabitive games from a perspective of improving society.” i.e. helping people practice openended negotation, and giving coordination theorists an empirical sandbox.
The world might have become a very different place if the coordination theorists were the ones who got the empirical sandboxes 10 years ago. Some people were thinking that legal prediction markets would be the way to dath ilan, and funnily enough it seems like the SBF trial/book might have brought attention to prediction markets. But very few people are aware that there exist firms and government/military agencies that have access to orders of magnitude better human behavior data than anywhere in academia, and that in the last 10 years human civilization gained vastly more powerful capability to understand the human mind than any point in the last 10,000 years of human civilization. The games and empirical sandboxes that Raemon describes already exist, but were hijacked by moloch; by default they got twisted into what we now understand as social media, which hijacks the human impulse to pursue social status in the maximally profitable way. And, unlike conventional war, information warfare is winnable, but in order to outperform ineffective ads and 20th century propaganda, and actually cause overwhelming victory like in communist East Germany, they need a superior understanding of the human mind to what people had in the 20th century. And then AI got caught up in it, because AI is great for processing and utilizing data, and now AI is so thoroughly entrenched in that clusterfuck AGI alignment just gets crowded out by default because it’s oversaturated with corporate executives and natsec folk who assume that it’s just another power grab, when in reality it actually is the engineering problem that the fate of the universe hinges on. If this post existed 20 years ago (or even 10) then maybe things wouldn’t have unfolded like this.
The games and empirical sandboxes that Raemon describes already exist, but were hijacked by moloch; by default they got twisted into what we now understand as social media, which hijacks the human impulse to pursue social status in the maximally profitable way
I didn’t understand from this comment what the referent was for “the games/sandboxes exist, but were highjacked.” What things are you counting as games/sandboxes of this type?
The world might have become a very different place if the coordination theorists were the ones who got the empirical sandboxes 10 years ago. Some people were thinking that legal prediction markets would be the way to dath ilan, and funnily enough it seems like the SBF trial/book might have brought attention to prediction markets. But very few people are aware that there exist firms and government/military agencies that have access to orders of magnitude better human behavior data than anywhere in academia, and that in the last 10 years human civilization gained vastly more powerful capability to understand the human mind than any point in the last 10,000 years of human civilization. The games and empirical sandboxes that Raemon describes already exist, but were hijacked by moloch; by default they got twisted into what we now understand as social media, which hijacks the human impulse to pursue social status in the maximally profitable way. And, unlike conventional war, information warfare is winnable, but in order to outperform ineffective ads and 20th century propaganda, and actually cause overwhelming victory like in communist East Germany, they need a superior understanding of the human mind to what people had in the 20th century. And then AI got caught up in it, because AI is great for processing and utilizing data, and now AI is so thoroughly entrenched in that clusterfuck AGI alignment just gets crowded out by default because it’s oversaturated with corporate executives and natsec folk who assume that it’s just another power grab, when in reality it actually is the engineering problem that the fate of the universe hinges on. If this post existed 20 years ago (or even 10) then maybe things wouldn’t have unfolded like this.
I didn’t understand from this comment what the referent was for “the games/sandboxes exist, but were highjacked.” What things are you counting as games/sandboxes of this type?
(minor aside: the lack of paragraph breaks in your comment made it feel harder to read to me)
Social media and user data, possibly including IoT sensor data to some unknown extent.