I’ll believe you’re a gold miner when I see your gold.
Numenta http://www.numenta.org is building smart software brains that simulate hardware brains, with the ultimate goal of building energy-inexpensive hardware brains. Vicarious is building smart hardware brains. Robobrain is building—and educating—smart hardware brains. IBM is building smart hardware brains, with its Neuromorphic computing project. OpenCog is building a software system, which is very good, but of unknown (to me) competitiveness with the other projects. Honda is building smart hardware brains, and teaching them. Jürgen Schmidhuber and Andrew Ng are building non-human brain-like machines and educating them. Hugo de Garis was building smart hardware brains (and dropped out to be an educator, I guess).
In any event, building context-independent brainlike systems seems to be the way to go. In much the same way that decent parents don’t allow one child to beat up and even murder their other kids, I imagine that such synthetic intelligences, given time, would disallow the stagnation caused by omnipresent and dominant sociopathy.
This is also why it’s important to measure the success of organizations by their stated goals, and by their incremental results on the path to those goals. A more nimble, more adventurous organization would be manufacturing chipsets now, based on the latest science. Such an organization would give up some control in order to retain influence. You will never be in control of a superhuman synthetic intelligence.
But almost all parents have influence (not control) over their children. The smartest parents steer their children into self-building directions.
That is all the power that anyone at the human level has, in comparison to superhuman, self-directing intelligences.
Most or all of the prior brain projects listed, by the way, lack mirror neurons, and do what sociopaths or non-sociopath humans tell them to do. They will increasingly do what sociopaths tell them to do, because sociopaths print the money, and dominate government positions of top-down control. The original intention of LessWrong is valuable, but its standards may be too much of a top-down variety, and too little of a bottom-up variety.
You can’t stop the sociopath from existing. He exists. And, he has intentionality, and will seek to corrupt your systems. But there are lots of Tim Murphys and George Washingtons out there. There are lots of Henry David Thoreaus, George Orwells, Lysander Spooners, and Doug Stanhopes out there, even if they are vastly outnumbered by neutral-by-default brains, and sociopaths. The two ends of the spectrum of intelligence compete for the allegiance of the neutral-by-default brains. There are so many human minds that they can’t all be Round pegs beaten into square holes, no matter how much the education bureaucrats try.
Really, at a deep level, LessWrong should be an educational organization that seeks to build hardware brains and then educate them properly. This then solves previously intractable problems like rising totalitarianism, space travel, etc., if only out of boredom. Perhaps it even solves the problem of human stupidity, supermodifying us all.
Ray Kurzweil has the right idea, but there are too few Kurzweilian projects. Evolution got it right by making lots of human kids. If human lifespan was 1,000, there would have been more family planning, and fewer kids, and therefore, fewer geniuses of the caliber necessary to adapt to brutal and changing environments, including those of Manichean sociopathy. (Also, the Manichean devils themselves, like Stalin and Pol Pot, would have not died off.)
Human brains have proven that it’s possible to build very smart PhD scientists who don’t care if the world around them becomes a Stalinist hellhole, so long as the temperature of the water is raised incrementally, and they have some reason to think they will be shielded from its boiling. So, the goal of LessWrong is to build classical liberal robots, who oppose all taxation, but especially property taxation (so that large numbers of humans aren’t rendered homeless and displaced).
Smart doesn’t necessarily = “libertarian.” However, “smartest, given reality” does. This is why virtually everyone thinks it’s a good thing that slavery ended. Once a system has decided correctly, opposition drops to nearly zero.
Right now, the libertarian and single-issue resisters are growing in number, because it makes no sense to allow powerful sociopaths to siphon away the value of everyone who trades in “Federal Reserve Notes.” It makes no sense to allow them to educate our kids, or subject our kids to a dumbed-down society/world of their creation.
I’ll believe you’re a gold miner when I see your gold.
Numenta http://www.numenta.org is building smart software brains that simulate hardware brains, with the ultimate goal of building energy-inexpensive hardware brains. Vicarious is building smart hardware brains. Robobrain is building—and educating—smart hardware brains. IBM is building smart hardware brains, with its Neuromorphic computing project. OpenCog is building a software system, which is very good, but of unknown (to me) competitiveness with the other projects. Honda is building smart hardware brains, and teaching them. Jürgen Schmidhuber and Andrew Ng are building non-human brain-like machines and educating them. Hugo de Garis was building smart hardware brains (and dropped out to be an educator, I guess).
In any event, building context-independent brainlike systems seems to be the way to go. In much the same way that decent parents don’t allow one child to beat up and even murder their other kids, I imagine that such synthetic intelligences, given time, would disallow the stagnation caused by omnipresent and dominant sociopathy.
This is also why it’s important to measure the success of organizations by their stated goals, and by their incremental results on the path to those goals. A more nimble, more adventurous organization would be manufacturing chipsets now, based on the latest science. Such an organization would give up some control in order to retain influence. You will never be in control of a superhuman synthetic intelligence.
But almost all parents have influence (not control) over their children. The smartest parents steer their children into self-building directions.
That is all the power that anyone at the human level has, in comparison to superhuman, self-directing intelligences.
Most or all of the prior brain projects listed, by the way, lack mirror neurons, and do what sociopaths or non-sociopath humans tell them to do. They will increasingly do what sociopaths tell them to do, because sociopaths print the money, and dominate government positions of top-down control. The original intention of LessWrong is valuable, but its standards may be too much of a top-down variety, and too little of a bottom-up variety.
You can’t stop the sociopath from existing. He exists. And, he has intentionality, and will seek to corrupt your systems. But there are lots of Tim Murphys and George Washingtons out there. There are lots of Henry David Thoreaus, George Orwells, Lysander Spooners, and Doug Stanhopes out there, even if they are vastly outnumbered by neutral-by-default brains, and sociopaths. The two ends of the spectrum of intelligence compete for the allegiance of the neutral-by-default brains. There are so many human minds that they can’t all be Round pegs beaten into square holes, no matter how much the education bureaucrats try.
Really, at a deep level, LessWrong should be an educational organization that seeks to build hardware brains and then educate them properly. This then solves previously intractable problems like rising totalitarianism, space travel, etc., if only out of boredom. Perhaps it even solves the problem of human stupidity, supermodifying us all.
Ray Kurzweil has the right idea, but there are too few Kurzweilian projects. Evolution got it right by making lots of human kids. If human lifespan was 1,000, there would have been more family planning, and fewer kids, and therefore, fewer geniuses of the caliber necessary to adapt to brutal and changing environments, including those of Manichean sociopathy. (Also, the Manichean devils themselves, like Stalin and Pol Pot, would have not died off.)
Human brains have proven that it’s possible to build very smart PhD scientists who don’t care if the world around them becomes a Stalinist hellhole, so long as the temperature of the water is raised incrementally, and they have some reason to think they will be shielded from its boiling. So, the goal of LessWrong is to build classical liberal robots, who oppose all taxation, but especially property taxation (so that large numbers of humans aren’t rendered homeless and displaced).
Smart doesn’t necessarily = “libertarian.” However, “smartest, given reality” does. This is why virtually everyone thinks it’s a good thing that slavery ended. Once a system has decided correctly, opposition drops to nearly zero.
Right now, the libertarian and single-issue resisters are growing in number, because it makes no sense to allow powerful sociopaths to siphon away the value of everyone who trades in “Federal Reserve Notes.” It makes no sense to allow them to educate our kids, or subject our kids to a dumbed-down society/world of their creation.