I'm one that believes that super intelligence is not likely to happen any time soon. It's not likely to happen within any ones lifespan that's living now. But I do believe the current crop of AI will lead to some very scary weapons that will help us kill each other even better. The real harm is giving our monkey brains more powerful weapons. Harm mitigation is well within out reach and it will be for many years to come. Banning or voluntarily stopping it in one country will only fire up other countries to continue the development. We all know that AI brings with it both economic and military advantages. It may be wrong but no country wants to lose the AI race.
I honestly think we're less than 10 years out. Some of the stuff coming out of GPT-4 is clearly some sort of intelligence that's beyond human level, in speed at the very least, and often in quality. It has about 2 orders of magnitude more parameters than the human brain has neurons.
I'm not saying you're necessarily wrong about there being a lot of harm in using AI as a weapon. But even then, I'd MUCH rather live in a world where, say, China develops AI weaponry and uses it to conquer everyone else and rules with an iron fist, than a world where an super-intelligent AGI decides to get ALL of humanity out of the way in the pursuit of whatever mis-aligned goal it happened to learn, be that military power or curing cancer.
Of course, we might get one and then the other. Yay. I'm still not giving up until the killer drones are flying down the street shooting everyone in sight, or the nanobots are turning everyone into sludge, or whatever.
Some of the stuff coming out of GPT-4 is clearly some sort of intelligence that's beyond human level, in speed at the very least, and often in quality.
I know it falls down on other stuff. To me, it seems to operate in a world of a kind of dream-logic, where it does hallucinate things, but other times it is clearly reasoning in a pretty deep way.
This "Sparks of AGI" talk by Sebastien Bubeck covers a lot of amazing examples. (I'd provide a link but I'm not on a good connection and YouTube isn't loading for me)
Just because an AI makes mistakes that a human doesn't make, doesn't mean it's not smarter than a human. It's an alien. The things that it finds important and fixates on are not the same things we do, and there are a lot of things that most humans would get wrong that GPT-4 gets right. Sure, it's not great at understanding physical objects, but it's been trained on pure text, not by living in a physical world and playing with toys and having adults put names to shapes and movements. We have an intuition through living every second of our lives interacting with the physical world. I'd expect this sort of AI to find simple problems with physical objects hard in the same way that a human finds pure mathematics hard.
Not to mention that we've often accidentally trained it to give confident, plausible sounding answers, instead of saying "I don't know". An AGI is not necessarily going to look and sound like a really smart human.
Yeah, the first time you see it (usually something outside one's own expertise.. I mean did you carefully read what it gives there even in the first response?), now repeat it and vary a bit, poke it with some questions.. and it soon ffalls apart..You don't need to do it, just read some further tweets downwards..
An AGI is not necessarily going to look and sound like a really smart human.
Definitely not, but it will need to reason somehow.
Why do you think super intelligence is not likely to happen within anyone's lifetime? Is there something about the problem that makes you confident in its difficulty? What other technologies would you put on a comparable ~100 year timescale of difficulty?
1st, there's no true definition of what intelligence is. Right now it's a moving target. Moving targets are hard to hit. Defining intelligence will take time. How can we create what we can't define?
2, we are getting close to processing speed limits and the amount of easy to process data. Yet, AGI is nowhere near. LLM's are remarkable but it's data processing at scale. It's no smarter than my word processor. It's useful but it's not AIG.
3, there's no defining theory that will get us to AGI.We don't know how to get there. What we have is wishful thinking that says we'll be there soon. In any project the early percentage towards our task is relatively easy. The last X% is hard. I think we are underestimating how much time and effort it takes to complete the project of AIG. We have only scratched the surface of it. The 80/20 rule is real.
AIG reminds me of the search for ETs. We feel we are close to meeting an ET. Some even believe we have encountered them. But we have 0 evidence that they exist. It's been burned in to our culture by fiction writers but nothing in our real life experience says they exist. Similarly we feel we will meet AIG soon. We might get to AIG but I think we are kidding ourselves that it will be here soon.
The real danger is the misuse of the AI we have by people. That really scares me.
We do it well enough with other things. ("I shall not today attempt further to define the kinds of material I understand to be embraced within that shorthand description ["hard-core pornography"], and perhaps I could never succeed in intelligibly doing so. But I know it when I see it" -- Justice Potter Stewart)
Yet, AGI is nowhere near.
I don't think we have a reliable way of measuring how far away AGI is.
It's no smarter than my word processor.
Really? What word processor do you use?
there's no defining theory that will get us to AGI.We don't know how to get there.
I'd argue we also didn't know how to get here. But we did.
The real danger is[...]
There is plenty of room for more than one danger to be real.
Comments
I'm one that believes that super intelligence is not likely to happen any time soon. It's not likely to happen within any ones lifespan that's living now. But I do believe the current crop of AI will lead to some very scary weapons that will help us kill each other even better. The real harm is giving our monkey brains more powerful weapons. Harm mitigation is well within out reach and it will be for many years to come. Banning or voluntarily stopping it in one country will only fire up other countries to continue the development. We all know that AI brings with it both economic and military advantages. It may be wrong but no country wants to lose the AI race.
I honestly think we're less than 10 years out. Some of the stuff coming out of GPT-4 is clearly some sort of intelligence that's beyond human level, in speed at the very least, and often in quality. It has about 2 orders of magnitude more parameters than the human brain has neurons.
I'm not saying you're necessarily wrong about there being a lot of harm in using AI as a weapon. But even then, I'd MUCH rather live in a world where, say, China develops AI weaponry and uses it to conquer everyone else and rules with an iron fist, than a world where an super-intelligent AGI decides to get ALL of humanity out of the way in the pursuit of whatever mis-aligned goal it happened to learn, be that military power or curing cancer.
Of course, we might get one and then the other. Yay. I'm still not giving up until the killer drones are flying down the street shooting everyone in sight, or the nanobots are turning everyone into sludge, or whatever.
Hm erm, what? No, just no, not barely.
I dunno. Stuff like this seems pretty human to me: https://twitter.com/erikbryn/status/1644830053627461633/phot...
I know it falls down on other stuff. To me, it seems to operate in a world of a kind of dream-logic, where it does hallucinate things, but other times it is clearly reasoning in a pretty deep way.
This "Sparks of AGI" talk by Sebastien Bubeck covers a lot of amazing examples. (I'd provide a link but I'm not on a good connection and YouTube isn't loading for me)
Just because an AI makes mistakes that a human doesn't make, doesn't mean it's not smarter than a human. It's an alien. The things that it finds important and fixates on are not the same things we do, and there are a lot of things that most humans would get wrong that GPT-4 gets right. Sure, it's not great at understanding physical objects, but it's been trained on pure text, not by living in a physical world and playing with toys and having adults put names to shapes and movements. We have an intuition through living every second of our lives interacting with the physical world. I'd expect this sort of AI to find simple problems with physical objects hard in the same way that a human finds pure mathematics hard.
Not to mention that we've often accidentally trained it to give confident, plausible sounding answers, instead of saying "I don't know". An AGI is not necessarily going to look and sound like a really smart human.
Yeah, the first time you see it (usually something outside one's own expertise.. I mean did you carefully read what it gives there even in the first response?), now repeat it and vary a bit, poke it with some questions.. and it soon ffalls apart..You don't need to do it, just read some further tweets downwards..
Definitely not, but it will need to reason somehow.
Why do you think super intelligence is not likely to happen within anyone's lifetime? Is there something about the problem that makes you confident in its difficulty? What other technologies would you put on a comparable ~100 year timescale of difficulty?
1st, there's no true definition of what intelligence is. Right now it's a moving target. Moving targets are hard to hit. Defining intelligence will take time. How can we create what we can't define?
2, we are getting close to processing speed limits and the amount of easy to process data. Yet, AGI is nowhere near. LLM's are remarkable but it's data processing at scale. It's no smarter than my word processor. It's useful but it's not AIG.
3, there's no defining theory that will get us to AGI.We don't know how to get there. What we have is wishful thinking that says we'll be there soon. In any project the early percentage towards our task is relatively easy. The last X% is hard. I think we are underestimating how much time and effort it takes to complete the project of AIG. We have only scratched the surface of it. The 80/20 rule is real.
AIG reminds me of the search for ETs. We feel we are close to meeting an ET. Some even believe we have encountered them. But we have 0 evidence that they exist. It's been burned in to our culture by fiction writers but nothing in our real life experience says they exist. Similarly we feel we will meet AIG soon. We might get to AIG but I think we are kidding ourselves that it will be here soon.
The real danger is the misuse of the AI we have by people. That really scares me.
We do it well enough with other things. ("I shall not today attempt further to define the kinds of material I understand to be embraced within that shorthand description ["hard-core pornography"], and perhaps I could never succeed in intelligibly doing so. But I know it when I see it" -- Justice Potter Stewart)
I don't think we have a reliable way of measuring how far away AGI is.
Really? What word processor do you use?
I'd argue we also didn't know how to get here. But we did.
There is plenty of room for more than one danger to be real.