4 ms·

First, I don't think any of those greats had such a specific vision as this as an inevitable future: "functionally merge with our engineered system in a symbio

by inimino 10y ago

First, I don't think any of those greats had such a specific vision as this as an inevitable future: "functionally merge with our engineered system in a symbiotic way and then probably see the extinction of the human species" You can probably find quotes that will confirm your belief, but from a neutral viewpoint there is no such consensus between these people about the future. Secondly, even if there was, we have 60+ years more experience than they did with these systems, and we should be able to make better predictions than they did. There are many potential future outcomes, and it's reasonable to think about them and to try to mitigate certain possibilities. It's not reasonable to say that a certain outcome is inevitable and you know what's going to happen. We just don't.

AndrewKemendo 10y ago
There are many potential future outcomes, and it's reasonable to think about them and to try to mitigate certain possibilities. Let me rephrase then: If you evaluate the written history of humanity, there is the unavoidable trend that humans will create tools which replicate and obviate human actions in a more optimized manner. Stone Tool > Controlled Fire > Shelter > etc... to the point that the ultimate boundaries of human capability including things like creativity (Generative Adversarial Networks are the frontier right now) - which is one of the few remaining distinguishing features of humanity by the way - are being mechanized. Eventually they all will be, because there are people like me trying to make it happen. So while you are right that it's not inevitable, you can rest assured that we're working hard on it and the trends are in our favor.
- inimino 10y ago
  There are people like you trying to cure cancer, too, and trying to make money on Wall Street, trying to solve outstanding problems in physics, trying to make it to Mars, solve global warming, win wars, and so on. The trend you identify is that there will be a breakthrough and everything will change, but we have no way of knowing what it will be. When it comes, it might make your field irrelevant. Arguing that the breakthrough will come from the field you happen to be working in is understandable but it doesn't carry much weight with anyone outside your field.
  - AndrewKemendo 10y ago
    Right, and if you look at all of those problems, they have one thing in common: They are progressing the fastest in places where they are applying machine learning. Literally, I mean really they are. If you solve AGI you solve all of those problems. That's the whole point here. It's not a standalone technology it's really something that changes every industry.
    - inimino 10y ago
      Well, yes, at the moment machine learning is scooping up low-hanging fruit in a lot of areas. Whether that will lead to AGI somehow is pure speculation. It is of course true that solving AGI solves most other problems (while creating new ones that we are by no means ready to handle) but that doesn't mean it's likely to happen. Hopefully it doesn't anytime soon. There are other possible breakthroughs. Just as a random example, humanoid robots with effective IQ 60 that could be trained to do all our current manual labor. That would completely transform society and we would be sorting it out for decades. Or it might become possible for someone in a garage to synthesize infectious biological agents and intentionally or accidentally wipe us all out. Or we could get cheap energy beamed down as microwaves from solar-powered satellites. While it's true that AGI changes everything, it's not the only thing that can change everything, and you can't predict what will happen once everything changes.

by inimino 10y ago

AndrewKemendo 10y ago

There are many potential future outcomes, and it's reasonable to think about them and to try to mitigate certain possibilities. Let me rephrase then: If you evaluate the written history of humanity, there is the unavoidable trend that humans will create tools which replicate and obviate human actions in a more optimized manner. Stone Tool > Controlled Fire > Shelter > etc... to the point that the ultimate boundaries of human capability including things like creativity (Generative Adversarial Networks are the frontier right now) - which is one of the few remaining distinguishing features of humanity by the way - are being mechanized. Eventually they all will be, because there are people like me trying to make it happen. So while you are right that it's not inevitable, you can rest assured that we're working hard on it and the trends are in our favor.

inimino 10y ago
There are people like you trying to cure cancer, too, and trying to make money on Wall Street, trying to solve outstanding problems in physics, trying to make it to Mars, solve global warming, win wars, and so on. The trend you identify is that there will be a breakthrough and everything will change, but we have no way of knowing what it will be. When it comes, it might make your field irrelevant. Arguing that the breakthrough will come from the field you happen to be working in is understandable but it doesn't carry much weight with anyone outside your field.
- AndrewKemendo 10y ago
  Right, and if you look at all of those problems, they have one thing in common: They are progressing the fastest in places where they are applying machine learning. Literally, I mean really they are. If you solve AGI you solve all of those problems. That's the whole point here. It's not a standalone technology it's really something that changes every industry.
  - inimino 10y ago
    Well, yes, at the moment machine learning is scooping up low-hanging fruit in a lot of areas. Whether that will lead to AGI somehow is pure speculation. It is of course true that solving AGI solves most other problems (while creating new ones that we are by no means ready to handle) but that doesn't mean it's likely to happen. Hopefully it doesn't anytime soon. There are other possible breakthroughs. Just as a random example, humanoid robots with effective IQ 60 that could be trained to do all our current manual labor. That would completely transform society and we would be sorting it out for decades. Or it might become possible for someone in a garage to synthesize infectious biological agents and intentionally or accidentally wipe us all out. Or we could get cheap energy beamed down as microwaves from solar-powered satellites. While it's true that AGI changes everything, it's not the only thing that can change everything, and you can't predict what will happen once everything changes.