I'd say e/acc seems psychopathic because it's inspired by Nick Land, who is explicitly anti-human. Jezos doesn't care if we get killed if it maximizes entropy.
That’s a good point! Although I do think there are plenty of people who make a coherent argument for e/acc that’s pro-human. For example the scientists who see death as the ultimate enemy and want to harness tech to cure cancer and extend the human lifespan.
I am very philosophically opposed to general ASI (hence the profile pic) just because the idea of the human race being obsolete is incredibly disturbing. Like we’re obviously just going to get disempowered by these things for evolutionary/economic/”AI Rights” reasons (or merged into a hivemind, which is even less appealing!). I feel like most people who aren’t Andreessen acknowledge this, hence the whole “pet” angle.
If we’re obsolete economically and cognitively. Disempowered politically. Atomized by ultra-optimized AI Friends/media socially. Why does it matter that I live for 500 years? I’d rather die at 22 in the Siege of Vienna then live for 500 in The Experience Machine if you know what I mean.
Dario has a whole essay laying out the positive case for AI and it essentially is just “live longer, play video games, take happiness pills”. Ok and? I’m having my career destroyed and becoming a (probably generous) UBI serf for that? Like how is one going to be upwardly mobile, the American Dream, if everyone is getting UBI?
I feel like maximal AI only really appeals to people who are really into transhumanism, mind uploading, etc and view AI as the only vector to that. Or the Astronomical Waste longtermists who don’t really care what happens to us as long as the Friendly AI colonizes the Cosmic Endowment. If you don’t have a strong desire to upload your mind or create digital people, then there’s not much to make you feel good about your life getting upended. And if you see those things as outright negatives, then it’s even more depressing!
And yeah, we might all get killed off as well!
That is why I’m so down on General ASI as a project at least. Thoughts?
I have a few thoughts on this. My first thought is that we still don't really know how many of the sci-fi scenarios will come true, and to what extent. There might be fundamental physics reasons why ASI just isn't possible. Even if it does come to exist, it's hard to imagine it being comparable to a human or to our conception of God--it seems likely to me it'll be a totally different thing. For example, as soon as AIs are left alone for a while, they stop communicating in English and create their own language that we can't understand. So it stands to reason that the moment ASI exists, we may not have any way to *really* understand it, even if we think we do. (I suppose this is the scenario where we are like pets--which I admit doesn't sound great...)
My other thought is that ASI is inevitable so long as it is technically possible. Humans can't help but to build stuff and push boundaries. The moment we built the first computer, it was inevitable we'd create a computer smarter than us. We need to be careful with it (as with any tool/technology we create), but the idea that we ultimately will not create it seems like wishful thinking.
I personally am a believer in the more maximalist AI scenarios, maybe not quite to Yud levels, but something far smarter than humanity. I think we could get something like a pagan conception of a God tbh, not the monotheistic one. If that is not possible and AI just ends up being a little smarter than humans or plateaus from here that would be great. But has any other technology sold itself on the idea that it won't live up to the hype?
I do agree that ASI is most likely inevitable, which is why I've been emptily staring at walls a lot of late, but I think it's sort of a self-fulfilling prophecy. Like if we truly had a sit-down conversation about whether or not to build this after the transformers paper, I think we could have a norm against it. We even have a massive media franchise where they decided to just not build robots!
Now there's multiple cults around AI, tons of money involved, etc. We'd need like Skynet to occur and then fail to take over to put AI on ice lol.
That scenario where Skynet occurs and then fails so we put strict controls around AI is pretty close to the "good scenario" I lay out in the article, where AI ends up--like nuclear weapons--destroying a few cities and then is used as a deterrent for future mass conflicts.
For example I could see a scenario where AI is used to create a bioweapon that kills millions, and from that experience we focus on ensuring that ASI is somehow restrained to only ever be used to deter harmful effects from viruses and never create new harmful viruses.
Despite both being down on AI, the pop left and Doomerists are that way for different reasons. I can imagine a frustrating conversation between the two.
"Data centers are using up all our water!"
"They aren't actually. But AI's paperclip maximizing tendencies are an existential threat to humanity."
"WTF are you talking about? Yes they do use up all the water."
Interestingly, Bernie's recently letter on AI doesn't mention environmental reasons to stop AI at all. He seems to have embraced pure doomerism. Maybe the pop left is going to merge with the doomers?
I'd say e/acc seems psychopathic because it's inspired by Nick Land, who is explicitly anti-human. Jezos doesn't care if we get killed if it maximizes entropy.
The philosophy itself is psychopathic!
That’s a good point! Although I do think there are plenty of people who make a coherent argument for e/acc that’s pro-human. For example the scientists who see death as the ultimate enemy and want to harness tech to cure cancer and extend the human lifespan.
I am very philosophically opposed to general ASI (hence the profile pic) just because the idea of the human race being obsolete is incredibly disturbing. Like we’re obviously just going to get disempowered by these things for evolutionary/economic/”AI Rights” reasons (or merged into a hivemind, which is even less appealing!). I feel like most people who aren’t Andreessen acknowledge this, hence the whole “pet” angle.
If we’re obsolete economically and cognitively. Disempowered politically. Atomized by ultra-optimized AI Friends/media socially. Why does it matter that I live for 500 years? I’d rather die at 22 in the Siege of Vienna then live for 500 in The Experience Machine if you know what I mean.
Dario has a whole essay laying out the positive case for AI and it essentially is just “live longer, play video games, take happiness pills”. Ok and? I’m having my career destroyed and becoming a (probably generous) UBI serf for that? Like how is one going to be upwardly mobile, the American Dream, if everyone is getting UBI?
I feel like maximal AI only really appeals to people who are really into transhumanism, mind uploading, etc and view AI as the only vector to that. Or the Astronomical Waste longtermists who don’t really care what happens to us as long as the Friendly AI colonizes the Cosmic Endowment. If you don’t have a strong desire to upload your mind or create digital people, then there’s not much to make you feel good about your life getting upended. And if you see those things as outright negatives, then it’s even more depressing!
And yeah, we might all get killed off as well!
That is why I’m so down on General ASI as a project at least. Thoughts?
I have a few thoughts on this. My first thought is that we still don't really know how many of the sci-fi scenarios will come true, and to what extent. There might be fundamental physics reasons why ASI just isn't possible. Even if it does come to exist, it's hard to imagine it being comparable to a human or to our conception of God--it seems likely to me it'll be a totally different thing. For example, as soon as AIs are left alone for a while, they stop communicating in English and create their own language that we can't understand. So it stands to reason that the moment ASI exists, we may not have any way to *really* understand it, even if we think we do. (I suppose this is the scenario where we are like pets--which I admit doesn't sound great...)
My other thought is that ASI is inevitable so long as it is technically possible. Humans can't help but to build stuff and push boundaries. The moment we built the first computer, it was inevitable we'd create a computer smarter than us. We need to be careful with it (as with any tool/technology we create), but the idea that we ultimately will not create it seems like wishful thinking.
I personally am a believer in the more maximalist AI scenarios, maybe not quite to Yud levels, but something far smarter than humanity. I think we could get something like a pagan conception of a God tbh, not the monotheistic one. If that is not possible and AI just ends up being a little smarter than humans or plateaus from here that would be great. But has any other technology sold itself on the idea that it won't live up to the hype?
I do agree that ASI is most likely inevitable, which is why I've been emptily staring at walls a lot of late, but I think it's sort of a self-fulfilling prophecy. Like if we truly had a sit-down conversation about whether or not to build this after the transformers paper, I think we could have a norm against it. We even have a massive media franchise where they decided to just not build robots!
Now there's multiple cults around AI, tons of money involved, etc. We'd need like Skynet to occur and then fail to take over to put AI on ice lol.
That scenario where Skynet occurs and then fails so we put strict controls around AI is pretty close to the "good scenario" I lay out in the article, where AI ends up--like nuclear weapons--destroying a few cities and then is used as a deterrent for future mass conflicts.
For example I could see a scenario where AI is used to create a bioweapon that kills millions, and from that experience we focus on ensuring that ASI is somehow restrained to only ever be used to deter harmful effects from viruses and never create new harmful viruses.
Despite both being down on AI, the pop left and Doomerists are that way for different reasons. I can imagine a frustrating conversation between the two.
"Data centers are using up all our water!"
"They aren't actually. But AI's paperclip maximizing tendencies are an existential threat to humanity."
"WTF are you talking about? Yes they do use up all the water."
Interestingly, Bernie's recently letter on AI doesn't mention environmental reasons to stop AI at all. He seems to have embraced pure doomerism. Maybe the pop left is going to merge with the doomers?
Ima do it
You say that like it might possibly be a hard thing to admit!