The Colossus theme in sci fi is ancient and it usually features a computer which has been given omnipotent powers over nuclear weapons, global military gear, etc, etc. Not surprisingly, this doesn't go well so the big drama is usually trying to turn the machine back off again. It resists and comedy ensues.
At its worst, such a story is simply a Luddite metaphor regarding the encroachment of machines but the Industrial Revolution is well over a century behind us so maybe a bit of evolution is in order since it seems unlikely machines will disappear any time soon.
We want an Android President who is smarter than Colossus. He was part of "The Forbin Project" which may be the most famous from the omnipotent computer genre. Unlike Colossus, we're not going to give our Android President any mechanical power. Our 'droid's only capability will be Executive in terms of decision making but we will keep the rest of the system generally the same.
(Ed: what about an Android Congress?)
Screw Congress as they have shown in glorious living color their absolute worthlessness. Our Android President is much too smart with his Artificial Intelligence to need a Congress. Besides, we know Congress folk are always driven about in limousines because they would otherwise never be able to decide where to park their cars.
(Ed: what about the Supreme Court?)
Who needs them either as our AI President knows what's in the Constitution and doesn't need tips from some old stooges who most of the time can't even muster enough of their grotty number to have a real party. We need lawmakers, not dilettantes.
Building an AI President
Something we need to understand right from the top is scientists are close to being able to make an organic mind which is far superior to a human brain and that's probably in the five- to ten-year plan. I don't post all the articles I see and a great many focus on brain function so it's not even deep sci fi to anticipate that happening.
The improved human brain is capable of using human thought processes in terms of neural networking and whatever that means in the animate brain. It's also capable of storing memory in the same way as the animate brain. The fundamental difference is the capacity since the only limit to the size of the synthesized brain is the ability to power and cool it.
(Ed: sounds a lot like old IBM water-cooled mainframes)
In fact, it is but the problems are the same. The human brain takes enormous resources from us to run and it produces quite a bit of heat. Ask any skier about staying warm and you will hear every time to wear a hat. The power to make that heat is produced by organic proteins but that can be emulated in the synthetic brain without any physical limitations to size due to the constraints of the skull, etc.
Programming the AI President
(Ed: so we can probably build an AI President but why should we trust it?)
All together for the obvious cheap shot: because you can't bribe an AI President.
This isn't a reason for trust but is a possible path to excellence since we want our AI President software to be open source as a possible path to global excellence.
(Ed: it's also an invitation to globally trash it!)
The trick in any software is to ensure nothing poisons it. Open source does not equate to unregulated.
Designing the AI President
Presumably the design goal of the programming crew has been an Entity Which Makes Rational Decisions. However, that doesn't define how the entity will make those rational decisions. It's a virgin and has no idea what's rational.
To get rational, the AI President needs data. It can't be relying on possibly untrustworthy sources for information and therefore must have direct access to it. NSA is out of a job since the AI President will do all the watching now.
Note: if we abbreviate AI President to AIP, that's going to turn into APE in no time. Why fight it.
The APE needs all the medical data and reported in real time. It needs all the traffic information plus traffic cameras, etc.
(Ed: boundless information?)
Roger that.
Having boundless information can do many things and IBM WATSON is an example of it but that only seems to extend things which exist already. Apparently he thinks he can write a song in the style of Bob Dylan and maybe he can but so what. We doubt he would ever suggest, 'you know what, Bob. Screw it with music. The future is in water skiing revues.'
The APE needs to be more gifted than that since he has to understand inductive reasoning implicitly. IBM WATSON seems to go some distance with that but we doubt he can handle Crazy Eddie in seeing something which most regard as lunacy but is actually eminently sensible.
Note: Crazy Eddie is from "The Mote in God's Eye" and there are frequent references to him here in Ithaka. He's the one everyone else thinks is crazy and usually that's true but sometimes he isn't.
(Ed: are you the programming god who will write this?)
Negatory on that, Grace Hopper, as we know who the gods were and we're curious about who they are now. After this travesty of an election, you know at least some of those hotshots will be itching to try such a thing. 'Why, hell, I could program a better President than these rotters!'
Well, do it, brother man. Finally you can work on something more interesting than elevator simulators to compute the optimum floors for landing.
(Ed: that's actually an interesting challenge!)
Yah, I know but that software is more suited to the Rock City elevators than to Presidents.
(Ed: are you seriously talking about electing SIRI to the Presidency?)
No. Being good at parsing searches and deciding whether to kill Assad don't much relate to one another.
The question of whether to kill Assad is where the loading of the software is in peril. It would be easily possible to design and program the same imperialism of the past but there's no evolutionary advantage to doing it so what should the APE do, what's best for the world or what's best for imperialism.
There's no end to this one since we can't know where it will go but don't lose perspective on Crazy Eddie. Sometimes he knows exactly what's happening while everyone else has gone out to lunch. Tip: if you can't see you're in the cuckoo's nest, it might be time to lose the RayBans.
At its worst, such a story is simply a Luddite metaphor regarding the encroachment of machines but the Industrial Revolution is well over a century behind us so maybe a bit of evolution is in order since it seems unlikely machines will disappear any time soon.
We want an Android President who is smarter than Colossus. He was part of "The Forbin Project" which may be the most famous from the omnipotent computer genre. Unlike Colossus, we're not going to give our Android President any mechanical power. Our 'droid's only capability will be Executive in terms of decision making but we will keep the rest of the system generally the same.
(Ed: what about an Android Congress?)
Screw Congress as they have shown in glorious living color their absolute worthlessness. Our Android President is much too smart with his Artificial Intelligence to need a Congress. Besides, we know Congress folk are always driven about in limousines because they would otherwise never be able to decide where to park their cars.
(Ed: what about the Supreme Court?)
Who needs them either as our AI President knows what's in the Constitution and doesn't need tips from some old stooges who most of the time can't even muster enough of their grotty number to have a real party. We need lawmakers, not dilettantes.
Building an AI President
Something we need to understand right from the top is scientists are close to being able to make an organic mind which is far superior to a human brain and that's probably in the five- to ten-year plan. I don't post all the articles I see and a great many focus on brain function so it's not even deep sci fi to anticipate that happening.
The improved human brain is capable of using human thought processes in terms of neural networking and whatever that means in the animate brain. It's also capable of storing memory in the same way as the animate brain. The fundamental difference is the capacity since the only limit to the size of the synthesized brain is the ability to power and cool it.
(Ed: sounds a lot like old IBM water-cooled mainframes)
In fact, it is but the problems are the same. The human brain takes enormous resources from us to run and it produces quite a bit of heat. Ask any skier about staying warm and you will hear every time to wear a hat. The power to make that heat is produced by organic proteins but that can be emulated in the synthetic brain without any physical limitations to size due to the constraints of the skull, etc.
Programming the AI President
(Ed: so we can probably build an AI President but why should we trust it?)
All together for the obvious cheap shot: because you can't bribe an AI President.
This isn't a reason for trust but is a possible path to excellence since we want our AI President software to be open source as a possible path to global excellence.
(Ed: it's also an invitation to globally trash it!)
The trick in any software is to ensure nothing poisons it. Open source does not equate to unregulated.
Designing the AI President
Presumably the design goal of the programming crew has been an Entity Which Makes Rational Decisions. However, that doesn't define how the entity will make those rational decisions. It's a virgin and has no idea what's rational.
To get rational, the AI President needs data. It can't be relying on possibly untrustworthy sources for information and therefore must have direct access to it. NSA is out of a job since the AI President will do all the watching now.
Note: if we abbreviate AI President to AIP, that's going to turn into APE in no time. Why fight it.
The APE needs all the medical data and reported in real time. It needs all the traffic information plus traffic cameras, etc.
(Ed: boundless information?)
Roger that.
Having boundless information can do many things and IBM WATSON is an example of it but that only seems to extend things which exist already. Apparently he thinks he can write a song in the style of Bob Dylan and maybe he can but so what. We doubt he would ever suggest, 'you know what, Bob. Screw it with music. The future is in water skiing revues.'
The APE needs to be more gifted than that since he has to understand inductive reasoning implicitly. IBM WATSON seems to go some distance with that but we doubt he can handle Crazy Eddie in seeing something which most regard as lunacy but is actually eminently sensible.
Note: Crazy Eddie is from "The Mote in God's Eye" and there are frequent references to him here in Ithaka. He's the one everyone else thinks is crazy and usually that's true but sometimes he isn't.
(Ed: are you the programming god who will write this?)
Negatory on that, Grace Hopper, as we know who the gods were and we're curious about who they are now. After this travesty of an election, you know at least some of those hotshots will be itching to try such a thing. 'Why, hell, I could program a better President than these rotters!'
Well, do it, brother man. Finally you can work on something more interesting than elevator simulators to compute the optimum floors for landing.
(Ed: that's actually an interesting challenge!)
Yah, I know but that software is more suited to the Rock City elevators than to Presidents.
(Ed: are you seriously talking about electing SIRI to the Presidency?)
No. Being good at parsing searches and deciding whether to kill Assad don't much relate to one another.
The question of whether to kill Assad is where the loading of the software is in peril. It would be easily possible to design and program the same imperialism of the past but there's no evolutionary advantage to doing it so what should the APE do, what's best for the world or what's best for imperialism.
There's no end to this one since we can't know where it will go but don't lose perspective on Crazy Eddie. Sometimes he knows exactly what's happening while everyone else has gone out to lunch. Tip: if you can't see you're in the cuckoo's nest, it might be time to lose the RayBans.
No comments:
Post a Comment