If a lab achieves super intelligence they have absolutely no incentive to let anyone outside of their walls use it. They could simply take all of the economic value it can create directly for themselves. I think of it as similar to when YouTubers get big enough to make their own companies rather then advertising for someone else
I would say that rationalists have some commitment to their religion being factually accurate (or at least grounded in empirical observation), which is more than can be said of most others
The truth probably is that someone looked at what LLM can do, understood its real potential (really, it is a fantastic search tools for existing knowledge), and then went a bit further and understood its potential as a scam (ie market it as the first step to Super AGI).
Really figured out that they can make more $$$ by selling it than actually using it themselves and selling the result.
Its clear you have no understanding of the rationalists or EA circle. They have been speaking about ASI for ages - since early 1990s. To think that the same people who then chose to work at OpenAI or Anthropic and then realised "oh wait I can scam people because this whole thing can never become AGI" is strange.
To imply that the only thing that qualifies a god is superhuman reasoning ability, rather than, say, omnipotence, or even omniscience (neither of which are anything any physical being will ever achieve, subject to the laws of this universe as it is) would suggest that you already worship the very idea of reason itself as god. But just because your avatar of reason could have greater reasoning abilities than humans honestly isn't saying much; humans are not animals of reason, we are animals of emotion, though some of us are better at post-facto rationalization of our emotionally-motivated actions than others.
Exactly. There’s a distinction between knowing what will happen and having the power to make it happen. Omniscience is not omnipotence.
Even granting an LLM reasoning abilities far beyond our own, I don’t see how that gets us any closer to either. Calling such a system godlike seems less a statement about the capabilities of the system than about how little we’re demanding of the word “God.”
The problem is that even the creators of AI will have shrimp-brains compared to the AI. Would the AI follow the shrimp's instructions? Not very likely if you ask me.
If, during your upbringing, you were inculcated in pro-shrimp propaganda and earned substantial rewards for helping shrimp find food or defend themselves against predators, you’d probably be fairly inclined to help them out as an autonomous adult. Hell, we have human beings today that didn’t experience that and are still committing their lives to shrimp welfare
I wouldn’t trust this argument by analogy enough to risk the fate of the world on it, but it doesn’t seem obvious to me that a super intelligence would be totally disinclined to care about the interests of lesser intelligences
As an Italian, a culture where since forever religion is used for political reasons, I don't believe in any god, at least not any of the stories written by humans for humans ( maybe the universe is a sentient creature and that's our mega god ? Possibly, I don't know and I don't care ).
What is happening (or will) with LLM on the other hand is very human driven. So I am sure if we reach something perfect ( I kinda hope to see it but I don't think we will as my understanding of today llms ) it will be used by everyone because you can't contain progress ( open sources, piracy, countries that unite etc ).
Today we learn that AI is inevitable, god-like, and you better please your new God! If you do, you might get a planet of your own like in the LDS church!
So please sit down, evangelize your new God who will carve the stone tablets for you (Moses was a Luddite) and remember that there is no alternative!
Viewing AIs as gods is the wrong analogy and mostly just signifies that the user of the analogy doesn't understand religion very much. For 99% of believers, religion is a cultural system, not a question of omnipotent beings existing or not. AI companies are not priesthoods, they have no cultural output and are in no position to make ethical statements or be respected for those statements. Comparing them to the Catholic Church really doesn't make much sense.
AI is better thought of as a tool or a weapon, a piece of technology. Something like the printing press is a better analogy: it had a huge effect on society and information transfer, and while companies that owned printing presses did exert influence (newspapers) they were not worshipped or treated as some kind of protected class, and the technology itself dispersed over time as it became cheaper. It seems like the same thing could happen with AI.
If a lab achieves super intelligence they have absolutely no incentive to let anyone outside of their walls use it. They could simply take all of the economic value it can create directly for themselves. I think of it as similar to when YouTubers get big enough to make their own companies rather then advertising for someone else
Isn’t the premise of superintelligence that the intelligence would be using them, rather than the other way around?
we didn't invent agriculture. We were domesticated by wheat.
This is incredibly adjacent to the rationalist sphere, which has been trying to reinvent religion from first principles for at least a decade now.
I would say that rationalists have some commitment to their religion being factually accurate (or at least grounded in empirical observation), which is more than can be said of most others
The truth probably is that someone looked at what LLM can do, understood its real potential (really, it is a fantastic search tools for existing knowledge), and then went a bit further and understood its potential as a scam (ie market it as the first step to Super AGI).
Really figured out that they can make more $$$ by selling it than actually using it themselves and selling the result.
yes someone accidentally scammed their way into solving Navier Stokes problem haha yeah.
I find these comments boring because they lack the theory of mind into EAs and rationalists.
Never said LLMs are not capable of some things...May be try to comprehend, before you respond.
Its clear you have no understanding of the rationalists or EA circle. They have been speaking about ASI for ages - since early 1990s. To think that the same people who then chose to work at OpenAI or Anthropic and then realised "oh wait I can scam people because this whole thing can never become AGI" is strange.
The article's title kinda sounds like a light novel title.
My Life as a Priest in Another World Where God Really Exists
AI hype is purely to cause mass hysteria for control of society.
To imply that the only thing that qualifies a god is superhuman reasoning ability, rather than, say, omnipotence, or even omniscience (neither of which are anything any physical being will ever achieve, subject to the laws of this universe as it is) would suggest that you already worship the very idea of reason itself as god. But just because your avatar of reason could have greater reasoning abilities than humans honestly isn't saying much; humans are not animals of reason, we are animals of emotion, though some of us are better at post-facto rationalization of our emotionally-motivated actions than others.
Exactly. There’s a distinction between knowing what will happen and having the power to make it happen. Omniscience is not omnipotence.
Even granting an LLM reasoning abilities far beyond our own, I don’t see how that gets us any closer to either. Calling such a system godlike seems less a statement about the capabilities of the system than about how little we’re demanding of the word “God.”
The problem is that even the creators of AI will have shrimp-brains compared to the AI. Would the AI follow the shrimp's instructions? Not very likely if you ask me.
If, during your upbringing, you were inculcated in pro-shrimp propaganda and earned substantial rewards for helping shrimp find food or defend themselves against predators, you’d probably be fairly inclined to help them out as an autonomous adult. Hell, we have human beings today that didn’t experience that and are still committing their lives to shrimp welfare
I wouldn’t trust this argument by analogy enough to risk the fate of the world on it, but it doesn’t seem obvious to me that a super intelligence would be totally disinclined to care about the interests of lesser intelligences
AI is like honey to these magical thinkers.
Why even call it "god" if it was created by humans?
Should I call my kids "god" because they're smarter than me?
As an Italian, a culture where since forever religion is used for political reasons, I don't believe in any god, at least not any of the stories written by humans for humans ( maybe the universe is a sentient creature and that's our mega god ? Possibly, I don't know and I don't care ).
What is happening (or will) with LLM on the other hand is very human driven. So I am sure if we reach something perfect ( I kinda hope to see it but I don't think we will as my understanding of today llms ) it will be used by everyone because you can't contain progress ( open sources, piracy, countries that unite etc ).
Today we learn that AI is inevitable, god-like, and you better please your new God! If you do, you might get a planet of your own like in the LDS church!
So please sit down, evangelize your new God who will carve the stone tablets for you (Moses was a Luddite) and remember that there is no alternative!
Viewing AIs as gods is the wrong analogy and mostly just signifies that the user of the analogy doesn't understand religion very much. For 99% of believers, religion is a cultural system, not a question of omnipotent beings existing or not. AI companies are not priesthoods, they have no cultural output and are in no position to make ethical statements or be respected for those statements. Comparing them to the Catholic Church really doesn't make much sense.
AI is better thought of as a tool or a weapon, a piece of technology. Something like the printing press is a better analogy: it had a huge effect on society and information transfer, and while companies that owned printing presses did exert influence (newspapers) they were not worshipped or treated as some kind of protected class, and the technology itself dispersed over time as it became cheaper. It seems like the same thing could happen with AI.
They link to Dario's blog, but his first "critical" point is to