Continued from here:
http://www.internationalskeptics.com...4#post11345464
As I read through this thread "How's That Higher Minimum Wage Working Out For You?" from the beginning I noticed a lot of furor over who should have what but really no examination of the implicit assumption apparently shared by both sides of the debate that erupted that automation is mainly a threat to blue-collar laborers with minimal education.
It very likely is not however. This assumption and whether it actually holds up is so very important that it merits its own thread.
As the links I provided explain, it can be seen that machines are now, in particular, increasingly capable of doing things that require a disembodied kind of intelligence. This include things like analysis in medical imaging, automated trading in finance and text mining in legal matters. These are not things that, when the inputs are set up properly, require the machine to move around in the world, pick things up, look at them, manipulate them with hands or similar effectors, all in a dizzying array of possible conditions that various animals have to come to terms with and all in real time, no less. These sorts of things were described in a book that had a lot of influence on me, Understanding Intelligence, by cognitive scientist Rolf Pfeifer, as instances of "sensorimotor coordination". The book in question is predicated on the notion that embodiment plays a crucial role in cognition and that for real understanding of cognition, and to make artificial intelligence really intelligent, this role needs to be fully appreciated, and it needs to be included in the construction of intelligent machines. It has often been neglected in the past, but with the efforts of Pfeifer and others like the roboticist Rodney Brooks, it is finally starting to get its due. However, for the time being it is in disembodied intelligence that machines really shine right now. And it happens to be the case that a lot of white-collar careers are precisely those that this sort of intelligence can be set loose on.
Now when I say this, I don't mean to intend that blue-collar labor that requires a lot more sensorimotor coordination like, say, being a barber, will forever be off-limits for machines. I don't believe the claim frequently put forward by lolbertarians that every instance of automation necessarily creates more and better careers to offset the disruption and the reason I believe that is very simply because there does not appear to be any evidence that the human mind is capable of hypercomputation, in other words that any given human mind can do more computation in principle than a universal Turing machine can. If a human can do it, then it's computable, period. In fact I fully anticipate that machines will one day become conscious and will outstrip anything any human can do.
But for right now, as said, it is quite possibly really white-collar labor that has the most to worry about. If raising the minimum wage increases the threat of automation against unskilled laborers, then a fortiori there should be even more of a threat against skilled laborers who currently make a lot more than minimum wage. Which of course is what we are seeing happening for this reason and for the other reasons already given. But what does it all mean? My take? Given that these developments are in a sense steps towards a greater posthuman future for this planet, and given that some of the fields are full of smug asswipes who need to be humbled anyway (*cough* law *hack* finance), am I really to see them as a bad thing?
http://www.internationalskeptics.com...4#post11345464
Quote:
|
Originally Posted by Prokhor Zakharov (Post 11345464)
Notwithstanding McDonald's kiosks, automation is actually much more of a threat to white-collar careers, presumably because their blue-collar counterparts tend to rely more on sensorimotor coordination tasks that remain difficult to implement in machines. More information on that here:
http://ift.tt/28OcOrs Two applications in particular where the thought of the meatbag getting tossed out on their ass and getting replaced with a superior machine particularly fills me with glee: http://ift.tt/28OcPvx http://ift.tt/19R0hgk |
It very likely is not however. This assumption and whether it actually holds up is so very important that it merits its own thread.
As the links I provided explain, it can be seen that machines are now, in particular, increasingly capable of doing things that require a disembodied kind of intelligence. This include things like analysis in medical imaging, automated trading in finance and text mining in legal matters. These are not things that, when the inputs are set up properly, require the machine to move around in the world, pick things up, look at them, manipulate them with hands or similar effectors, all in a dizzying array of possible conditions that various animals have to come to terms with and all in real time, no less. These sorts of things were described in a book that had a lot of influence on me, Understanding Intelligence, by cognitive scientist Rolf Pfeifer, as instances of "sensorimotor coordination". The book in question is predicated on the notion that embodiment plays a crucial role in cognition and that for real understanding of cognition, and to make artificial intelligence really intelligent, this role needs to be fully appreciated, and it needs to be included in the construction of intelligent machines. It has often been neglected in the past, but with the efforts of Pfeifer and others like the roboticist Rodney Brooks, it is finally starting to get its due. However, for the time being it is in disembodied intelligence that machines really shine right now. And it happens to be the case that a lot of white-collar careers are precisely those that this sort of intelligence can be set loose on.
Now when I say this, I don't mean to intend that blue-collar labor that requires a lot more sensorimotor coordination like, say, being a barber, will forever be off-limits for machines. I don't believe the claim frequently put forward by lolbertarians that every instance of automation necessarily creates more and better careers to offset the disruption and the reason I believe that is very simply because there does not appear to be any evidence that the human mind is capable of hypercomputation, in other words that any given human mind can do more computation in principle than a universal Turing machine can. If a human can do it, then it's computable, period. In fact I fully anticipate that machines will one day become conscious and will outstrip anything any human can do.
But for right now, as said, it is quite possibly really white-collar labor that has the most to worry about. If raising the minimum wage increases the threat of automation against unskilled laborers, then a fortiori there should be even more of a threat against skilled laborers who currently make a lot more than minimum wage. Which of course is what we are seeing happening for this reason and for the other reasons already given. But what does it all mean? My take? Given that these developments are in a sense steps towards a greater posthuman future for this planet, and given that some of the fields are full of smug asswipes who need to be humbled anyway (*cough* law *hack* finance), am I really to see them as a bad thing?
via International Skeptics Forum http://ift.tt/28MyWiI
Aucun commentaire:
Enregistrer un commentaire