It costs money to make a new game, therefore supply is NOT endless. Supply of new games is limited by the expected return on investment for developing a game. Raising the revenue of game sellers means that the expected return increases which in turn means that more suppliers will enter the market.
If you go crack open an introductory economics textbook, you'll spot a brief primer on what nearly always happens when you increase suppliers for a good.
Yes, the unique nature of each games means that they are not totally interchangable, but within a genre the best quality game usually is the one that gets the most sales.
Strictly speaking, it's an insulting use of your Linux workstation. A box with a dedicated video decoder chip and a tiny CPU (about 2 major ICs total) would do the same job for a lot less.
1. The market prices for games will go down, since EVERYONE is paying for it.
2. Game developers will make more money.
In addition, since you don't have to buy another copy of a game just to have your friend play you multiplayer for an hour, you've eliminated a massive problem with the current model already.
Oh, forgot to add to this : while the GAME may be a PC game, you'll be able to use lots of devices for accessing the game, not just PCs. Among other things, the XBox 360 controller plugs in via USB to a PC just fine - I've used it to play PC games that supported it very well. (PC ports of Prince of Persia and Marvel Ultimate Alliance both supported it)
And eventually you'll be able to use your old game console as well.
Advantages of this technology
1. No more console wars. The consoles that have already been released are more than adequate to do the job of decoding the video for a game service like this. This means
a. Publishers get virtually bulletproof DRM by releasing games for a service like this, even better than they have with consoles. And it's not "DRM" like we hate on slashdot - they simply don't give you a copy of the game at all.
b. Developers only have to worry about ONE platform again - the multicore PC with a high end graphics card and lots and lots of RAM (that's what the GaiKai data centers will be stocked with). Much easier to develop for than a console - if you run into resource limitations, you can just tell GaiKai that your game needs higher end hardware.
c. No more OS and hardware conflicts that caused problems with PC games in the past. GaiKai can give you a copy of the OS image they put on their machines, and the exact part numbers they put into their hardware.
d. No more problems with users failing to buy an adequate dedicated graphics card, or to configure their PC correctly.
2. Groundbreaking new games are possible. Since GaiKai can guarantee that your game will run on a machine of defined specifications, you could really push the graphics.
3. Games can be sold by the hour of play. You could pay about $.50-$1 an hour and hop from game to game, playing whatever catches your interest. Each publisher would receive a share of the revenue proportional to the exact time you spent playing their game. Publishers would probably make more money overall, and gamers would get to enjoy ALL the games, not just AAA titles.
Disadvantages
1. Latency is unavoidable, and it's going to be a little more than some games on some systems today. In the video, I saw them playing Mario cart, and the gamer wasn't crashing into buildings - so the latency is probably not too bad.
2. A high bandwidth internet connection with guaranteed maximum latency is needed to make this service work. There needs to be quality of service routing by the ISPs to make sure that game packets aren't delayed. For the moment, not everyone has access to connections that fast. I live in a small town, but I have 8 mbps cable which is enough.
3. Compression artifacts mean that even in 5 years when people have 20 mbps connections to the service, the games won't be as sharp as the old days.
4. The Biggest Problem is that you're dependent on centralized services to enjoy a video game. (your ISP AND the service) When the service goes down you can't do anything at all. Even once they iron the bugs out, the annual downtime will probably be more than you've experienced with owning the device the games are played on.
Overall, I think the advantages overwhelmingly outweigh the disadvantages. I think consoles with their crazy hardware architectures are going to die away, relegated to the dustbins of history. In the future, all games will be PC compatible. They'll still release local copies of some games for hardcore gamers to run on their PCs, especially of multiplayer only CD key requiring games (like first person shooters)
What I am curious about is why this technology is being deployed on a wide scale now. Cogeneration, where a heat engine's waste heat is used to heat a structure has existed for a long time. There's no reason that natural gas generator/heater couldn't have been installed in your basement in 1970. It would have made your house more efficient then much as it would now. So what has changed over 40 years that make the arguments for/against shift in favor of doing it? The biggest change I can think of is maybe better communications makes it easier for the power company to remotely control the generator. (since it wouldn't do any good to only have a generator in your basement for supplying power to your own house, wouldn't get enough return on investment...that power needs to be sold/credited to other users as well)
Advertising may not be a permanent phenomenon. Right now, believe it or not, economists thing that advertising is a signal to consumers that a product meets minimum levels of quality. It's possible that the information technology will make product "reputation" far more important than advertising, making that component of branding less important. Right now, for instance, ASUS motherboards have a tremendous reputation for quality due to years of positive experiences. Yet, the company has never made a television advertisement or hired a celebrity as far as I know.
Truth is, a little harmless digital stereotyping never hurt anybody. I look forward to living in a future where the advertisements on the street are video screens and they adapt their message to who they think is walking by. That's the kind of world people wrote about in science fiction decades ago, or put into movies like Blade Runner. This kind of thing has been dreamed about for decades, and thanks to the hard work of thousands of people, is finally possible.
Sure, it's not really that "useful" a technological improvement...kind of evil almost...but it sure is cool.
Yes and no. The problem is that right now the underlying first technology was never developed. We have no cheap way to get into orbit at all. So, in order to send a crew to mars, we would have to burn up as much wealth as would be produced by the entire lives of about a million people. That is, about a million people would have to work their entire lives in order to get around 10-20 people to Mars for a couple years. Or about 100,000 folks to get some astronauts to the moon. (yes, I know, the labor is spread out among more people than that, but it's simplest to equate to man-lifetimes)
It isn't just money - it's the labor.
Laser launch, for instance, could potentially radically change that equation, such that a measily few "man decades" of labor is consumed for every person we launch into orbit.
Committing an armed robbery with a firearm has sent people to prison for 75 years or more. That's not the average sentence, but the criminal justice system is like a slot machine. Not worth it for a 1/5 share of 20k or less..
Agreed. SpaceX could probably have built at least ONE fcking prototype rocket ship for 3 billion dollars. They spend a few hundred million for their first successful launch. Instead we've got dick for all that money. Even if SpaceX couldn't have finished a mega heavy lift booster, I bet they could have flown a modular part of a larger design for 3 billion bucks.
Yes, 3 billion dollars of taxpayer money has been blown. However, the decision to make is : will the gains from FUTURE spending exceed FUTURE costs? We don't factor in the 3 billion already spent in this decision. Alas, it's impossible to quantify gains since a few moon rocks and some pretty pictures don't have a readily assignable value. I'd say no, because I think the 20 billion or whatever a working Ares rocket line would cost could be better spent on other areas of space exploration. 20 billion would pay for a lot of unmanned missions, or could be used to develop a cheaper way to get to orbit (such as lasers or an EM accelerator or something)
That's my point. It's not worth spending more than a day or two, at most, looking for the sub. Any longer than that and it would be more cost effective spending the money on a new one.
The submarine is self guided, with no tether, and runs on batteries. The sea is a very harsh environment...one has to expect to lose a few. $100,000 isn't really that much money, and I suspect that there recovery efforts will be fairly limited.
Scramjet tech is worthless. It's not a very good weapon - a scramjet is going to have one heck of a heat signature and probably can't be very stealthy. Not just from the exhaust...at Mach 6 the entire aircraft/missile is going to be glowing red from heat. Also, air to air missiles (like the Patriot) that use rockets already go that fast.
Second, it's worthless for getting stuff into orbit. The reason is simple - the reason a rocket costs so darn much has nothing to do with fuel. It has to do with complexity - it's very expensive to make something as complicated as a rocket work under all the stresses of a launch. A scramjet just worsens the problem. It's not the fuel or the size of the tankage that makes the rockets that SpaceX builds cost so much. It has to do with building the rocket well enough that it makes it and doesn't fail (again). A scramjet engine is evidently incredibly complex to make work, and is just another pork barrel project of the air force.
So, you do think that if we did succeed in developing some kind of machine intelligence comparable to humans in flexibility (but far faster in raw speed) then the 'singularity' would happen? Again, I think the singularity would be pretty short lived, since once you have converted all the mass in our solar system to optimized molecular components, what more can you really do? Nevertheless, the technology available to these posited beings would be fairly close to godlike from our human perspectives.
I know it would be hard for a being that thought that fast to communicate with meatspace humans without some tweaks. But, among other things, said being would be able to adjust his/her 'clockspeed' as necessary if there really is nothing to do for 2.7 million years. And, said being could spend a few million years gradualling improving itself to efficiently use it's extreme congitive performance and to multitask and to think in ways that humans cannot.
Another bloody obvious fact. It took nearly 3 billion years for evolution to reach the era before homo sapiens, which was perhaps 100,000 years ago. And then we evolved. Evolution is able to make more and more complex structures as time goes on because more and more unique biological systems are available for evolution to work on. A similar accelerating trend. So, 100,000 years since humans, a couple thousand years since the start of recorded history, and a century or two since the start of the industrial revolution. 30 years since the development of the integrated circuit. It kind of looks like progress is accelerating.
Reason for this is that as more and more correct ideas are discovered, those ideas make possible new science and technology. Once we accelerate our own brains, the pace of progress will accelerate to near infinity (well, temporarily, for a period of time until science and technology hits the hard limits allowed by physics)
I think the singularity will involve the developement of some sort of machine version of human intelligence (even if it's a human being scanned into a computer), and will involve incredibly rapid progress (since the machine will be able to think billions of times quicker than us) until our successors have hit the limits. Probably molecular technology that has converted all of the mass in our solar system into technological systems.
Actually, yeah, biology does exactly that, ultimately, bond by bond. I think at near absolute zero and supplied with electricity, in a non aqueous vacuum chamber, the whole process would be dramatically more efficient.
That only works in science fiction. You can't create electricity to power the lasers just from ambient heat, you need a temperature differential and are limited by Carnot for efficiency. You'd end up adding more heat to the earth running the lasers. Mirrors would actually work.
Here's how the mirror plan would work. Nuclear fission plants (or solar arrays) would power an array of about 10 billion dollars worth of solid state lasers. (at current prices, available today). The lasers would probably use LEDs to pump doped fiber optics, producing very cheap laser energy.
The capsules containing the mirrors would be kicked into the air using a catapault and then the bottom of the capsule would be vaporized using the lasers to create thrust. The laser array alone would insert the mirror capsules into orbit...tehre would be minimal to no onboard thrusters needed.
That's how you'd launch one every minute (need several arrays) over a 30 year period.
Human beings didn't design the technology in themselves, so that is irrelevent.
I'm saying the 2009 machine is more technologically advanced than the 1989 machine because it is more complex and had more thought put into it by human beings. Even if you are using it to run a web browser version of a word processor that was available in 1989.
Semantics. Technology is continuing to advance at an accelerating pace. That is bloody obvious, and you're simply trying to redefine "technology" in a way that makes this fact untrue. Over a longer timespan, say the past 1000 years, it is far more obvious how fast technological advancement is accelerating.
Other ideas : the same molecular tools could be used to create a complete copy of a frozen human brain. They would use the information gained about each synapse to create a molecular computer that emulated the dead person's mind. Presto, "AI". Singularity would happen within a year or so after this point.
Ok, so what if you had a machine that consisted of a flat plane of tiny nanoscale manipulator arms, built atom by atom using an atomic force microscope the first time. Essentially, the machine would be supplied with feedstock consisting of each pure element. The head containing the nanoscale manipulators would be able to lay down specific atoms at low temperature, and with itnermediates move electrons to create the chemical bonds. So, atom by atom the head could build up a physical object. Most notably, it could make a copy of itself.
Now, this nanoscale manufacturing head might be incredibly complex and might run very slow. However, if it could make a copy of itself, you'd have exponential growth on your side. After you make one, you could double the number of heads you have every cycle. If it took a week to manufacture a new head using an existing one, you'd have 4.5 quadrillion of these heads within a year.
I'm assuming a nanoscale manufacturing head is only a few hundred thousand atoms thick.
There is one problem : it would be very, very energy inefficient to make objects this way. Of course, if the nanoscale manufacturing heads also spend time making solar cells, you could generate enough energy to keep the whole process going.
For instance, you could send a single nanoscale manufacturing head, containers of starter feedstock, a nuclear reactor, ore processor, and a simple mining machine to the moon. The machine would gather lunar ore, which would be vaporized and separated using a mass spectrometer into individual element feedstock. The nanoscale manufacturing head would start making a second nanoscale manufacturing head... Those two heads would then make copies of themselves....eventually, you'd have enough heads that you would now need more energy to power them with. You'd put some of the heads to work making solar panels.
Eventually, you'd need more ore processors and minining tools. Again, some of the heads would be put to work making those...
Anyways, it's hard to say how fast the whole process would go without knowing how fast a nanoscale manufacturing head could work. Biological cell proteins are basically nanotech replicators, and a bacterial cell can replicate a complete copy of itself within an hour. So a week for a human designed one seems like a conservative estimate. I'd say within 1 to 5 years one could cover the entire lit lunar surface with solar panels and mining machines. At this point you could send up molecular blueprints for anything anyone on earth wanted manufactured, and within limits everyone on the planet would be able to get what they wanted. (obviously these factories would also manufacture reentry vehicles to get the finished products back to earth)
The weasel word is that it depends on how you define complexity. It is true for all things.
A solar pumped electric grid would eliminate a number of nasty problems with current electricity production. One is pollution, of course. Another is a dependence on fuel that occurs in finite deposits that are becoming increasingly labor intensive to recover. (since we can only get so much energy from hydro, and nuclear power is currently MORE expensive than solar)
A bulldozer requires a society to build it and maintain it that is vastly more complex than a simple camp of fruit pickers.
Yes, a modern PC is radically different. There are MANY more layers of abstraction in the code you are running right now...from BIOS layers that are nearly as complex as DOS was in 1989, to the OS kernel, GUI, browser, and now virtual machines running in the browser that are as complex as whole applications were in 1989. The main block components may look the same, but there's a LOT more in that CPU than there was in 1989, and a lot more memory cells in the DRAM as well.
Slashdot Mirror
It costs money to make a new game, therefore supply is NOT endless. Supply of new games is limited by the expected return on investment for developing a game. Raising the revenue of game sellers means that the expected return increases which in turn means that more suppliers will enter the market.
If you go crack open an introductory economics textbook, you'll spot a brief primer on what nearly always happens when you increase suppliers for a good.
Yes, the unique nature of each games means that they are not totally interchangable, but within a genre the best quality game usually is the one that gets the most sales.
Strictly speaking, it's an insulting use of your Linux workstation. A box with a dedicated video decoder chip and a tiny CPU (about 2 major ICs total) would do the same job for a lot less.
Because if you can't steal a game, then
1. The market prices for games will go down, since EVERYONE is paying for it.
2. Game developers will make more money.
In addition, since you don't have to buy another copy of a game just to have your friend play you multiplayer for an hour, you've eliminated a massive problem with the current model already.
Oh, forgot to add to this : while the GAME may be a PC game, you'll be able to use lots of devices for accessing the game, not just PCs. Among other things, the XBox 360 controller plugs in via USB to a PC just fine - I've used it to play PC games that supported it very well. (PC ports of Prince of Persia and Marvel Ultimate Alliance both supported it) And eventually you'll be able to use your old game console as well.
Welcome to the future.
Advantages of this technology
1. No more console wars. The consoles that have already been released are more than adequate to do the job of decoding the video for a game service like this. This means
a. Publishers get virtually bulletproof DRM by releasing games for a service like this, even better than they have with consoles. And it's not "DRM" like we hate on slashdot - they simply don't give you a copy of the game at all.
b. Developers only have to worry about ONE platform again - the multicore PC with a high end graphics card and lots and lots of RAM (that's what the GaiKai data centers will be stocked with). Much easier to develop for than a console - if you run into resource limitations, you can just tell GaiKai that your game needs higher end hardware.
c. No more OS and hardware conflicts that caused problems with PC games in the past. GaiKai can give you a copy of the OS image they put on their machines, and the exact part numbers they put into their hardware.
d. No more problems with users failing to buy an adequate dedicated graphics card, or to configure their PC correctly.
2. Groundbreaking new games are possible. Since GaiKai can guarantee that your game will run on a machine of defined specifications, you could really push the graphics.
3. Games can be sold by the hour of play. You could pay about $.50-$1 an hour and hop from game to game, playing whatever catches your interest. Each publisher would receive a share of the revenue proportional to the exact time you spent playing their game. Publishers would probably make more money overall, and gamers would get to enjoy ALL the games, not just AAA titles.
Disadvantages
1. Latency is unavoidable, and it's going to be a little more than some games on some systems today. In the video, I saw them playing Mario cart, and the gamer wasn't crashing into buildings - so the latency is probably not too bad.
2. A high bandwidth internet connection with guaranteed maximum latency is needed to make this service work. There needs to be quality of service routing by the ISPs to make sure that game packets aren't delayed. For the moment, not everyone has access to connections that fast. I live in a small town, but I have 8 mbps cable which is enough.
3. Compression artifacts mean that even in 5 years when people have 20 mbps connections to the service, the games won't be as sharp as the old days.
4. The Biggest Problem is that you're dependent on centralized services to enjoy a video game. (your ISP AND the service) When the service goes down you can't do anything at all. Even once they iron the bugs out, the annual downtime will probably be more than you've experienced with owning the device the games are played on.
Overall, I think the advantages overwhelmingly outweigh the disadvantages. I think consoles with their crazy hardware architectures are going to die away, relegated to the dustbins of history. In the future, all games will be PC compatible. They'll still release local copies of some games for hardcore gamers to run on their PCs, especially of multiplayer only CD key requiring games (like first person shooters)
What I am curious about is why this technology is being deployed on a wide scale now. Cogeneration, where a heat engine's waste heat is used to heat a structure has existed for a long time. There's no reason that natural gas generator/heater couldn't have been installed in your basement in 1970. It would have made your house more efficient then much as it would now. So what has changed over 40 years that make the arguments for/against shift in favor of doing it? The biggest change I can think of is maybe better communications makes it easier for the power company to remotely control the generator. (since it wouldn't do any good to only have a generator in your basement for supplying power to your own house, wouldn't get enough return on investment...that power needs to be sold/credited to other users as well)
Advertising may not be a permanent phenomenon. Right now, believe it or not, economists thing that advertising is a signal to consumers that a product meets minimum levels of quality. It's possible that the information technology will make product "reputation" far more important than advertising, making that component of branding less important. Right now, for instance, ASUS motherboards have a tremendous reputation for quality due to years of positive experiences. Yet, the company has never made a television advertisement or hired a celebrity as far as I know.
Truth is, a little harmless digital stereotyping never hurt anybody. I look forward to living in a future where the advertisements on the street are video screens and they adapt their message to who they think is walking by. That's the kind of world people wrote about in science fiction decades ago, or put into movies like Blade Runner. This kind of thing has been dreamed about for decades, and thanks to the hard work of thousands of people, is finally possible.
Sure, it's not really that "useful" a technological improvement...kind of evil almost...but it sure is cool.
Yes and no. The problem is that right now the underlying first technology was never developed. We have no cheap way to get into orbit at all. So, in order to send a crew to mars, we would have to burn up as much wealth as would be produced by the entire lives of about a million people. That is, about a million people would have to work their entire lives in order to get around 10-20 people to Mars for a couple years. Or about 100,000 folks to get some astronauts to the moon. (yes, I know, the labor is spread out among more people than that, but it's simplest to equate to man-lifetimes)
It isn't just money - it's the labor.
Laser launch, for instance, could potentially radically change that equation, such that a measily few "man decades" of labor is consumed for every person we launch into orbit.
Committing an armed robbery with a firearm has sent people to prison for 75 years or more. That's not the average sentence, but the criminal justice system is like a slot machine. Not worth it for a 1/5 share of 20k or less..
Agreed. SpaceX could probably have built at least ONE fcking prototype rocket ship for 3 billion dollars. They spend a few hundred million for their first successful launch. Instead we've got dick for all that money. Even if SpaceX couldn't have finished a mega heavy lift booster, I bet they could have flown a modular part of a larger design for 3 billion bucks.
Yes, 3 billion dollars of taxpayer money has been blown. However, the decision to make is : will the gains from FUTURE spending exceed FUTURE costs? We don't factor in the 3 billion already spent in this decision. Alas, it's impossible to quantify gains since a few moon rocks and some pretty pictures don't have a readily assignable value. I'd say no, because I think the 20 billion or whatever a working Ares rocket line would cost could be better spent on other areas of space exploration. 20 billion would pay for a lot of unmanned missions, or could be used to develop a cheaper way to get to orbit (such as lasers or an EM accelerator or something)
That's my point. It's not worth spending more than a day or two, at most, looking for the sub. Any longer than that and it would be more cost effective spending the money on a new one.
How much do you think it costs per day to send a ship out to look for the robot submarine?
The submarine is self guided, with no tether, and runs on batteries. The sea is a very harsh environment...one has to expect to lose a few. $100,000 isn't really that much money, and I suspect that there recovery efforts will be fairly limited.
Scramjet tech is worthless. It's not a very good weapon - a scramjet is going to have one heck of a heat signature and probably can't be very stealthy. Not just from the exhaust...at Mach 6 the entire aircraft/missile is going to be glowing red from heat. Also, air to air missiles (like the Patriot) that use rockets already go that fast.
Second, it's worthless for getting stuff into orbit. The reason is simple - the reason a rocket costs so darn much has nothing to do with fuel. It has to do with complexity - it's very expensive to make something as complicated as a rocket work under all the stresses of a launch. A scramjet just worsens the problem. It's not the fuel or the size of the tankage that makes the rockets that SpaceX builds cost so much. It has to do with building the rocket well enough that it makes it and doesn't fail (again). A scramjet engine is evidently incredibly complex to make work, and is just another pork barrel project of the air force.
So, you do think that if we did succeed in developing some kind of machine intelligence comparable to humans in flexibility (but far faster in raw speed) then the 'singularity' would happen? Again, I think the singularity would be pretty short lived, since once you have converted all the mass in our solar system to optimized molecular components, what more can you really do? Nevertheless, the technology available to these posited beings would be fairly close to godlike from our human perspectives.
I know it would be hard for a being that thought that fast to communicate with meatspace humans without some tweaks. But, among other things, said being would be able to adjust his/her 'clockspeed' as necessary if there really is nothing to do for 2.7 million years. And, said being could spend a few million years gradualling improving itself to efficiently use it's extreme congitive performance and to multitask and to think in ways that humans cannot.
Another bloody obvious fact. It took nearly 3 billion years for evolution to reach the era before homo sapiens, which was perhaps 100,000 years ago. And then we evolved. Evolution is able to make more and more complex structures as time goes on because more and more unique biological systems are available for evolution to work on. A similar accelerating trend. So, 100,000 years since humans, a couple thousand years since the start of recorded history, and a century or two since the start of the industrial revolution. 30 years since the development of the integrated circuit. It kind of looks like progress is accelerating.
Reason for this is that as more and more correct ideas are discovered, those ideas make possible new science and technology. Once we accelerate our own brains, the pace of progress will accelerate to near infinity (well, temporarily, for a period of time until science and technology hits the hard limits allowed by physics)
I think the singularity will involve the developement of some sort of machine version of human intelligence (even if it's a human being scanned into a computer), and will involve incredibly rapid progress (since the machine will be able to think billions of times quicker than us) until our successors have hit the limits. Probably molecular technology that has converted all of the mass in our solar system into technological systems.
Actually, yeah, biology does exactly that, ultimately, bond by bond. I think at near absolute zero and supplied with electricity, in a non aqueous vacuum chamber, the whole process would be dramatically more efficient.
That only works in science fiction. You can't create electricity to power the lasers just from ambient heat, you need a temperature differential and are limited by Carnot for efficiency. You'd end up adding more heat to the earth running the lasers. Mirrors would actually work.
Here's how the mirror plan would work. Nuclear fission plants (or solar arrays) would power an array of about 10 billion dollars worth of solid state lasers. (at current prices, available today). The lasers would probably use LEDs to pump doped fiber optics, producing very cheap laser energy.
The capsules containing the mirrors would be kicked into the air using a catapault and then the bottom of the capsule would be vaporized using the lasers to create thrust. The laser array alone would insert the mirror capsules into orbit...tehre would be minimal to no onboard thrusters needed.
That's how you'd launch one every minute (need several arrays) over a 30 year period.
Human beings didn't design the technology in themselves, so that is irrelevent.
I'm saying the 2009 machine is more technologically advanced than the 1989 machine because it is more complex and had more thought put into it by human beings. Even if you are using it to run a web browser version of a word processor that was available in 1989.
Semantics. Technology is continuing to advance at an accelerating pace. That is bloody obvious, and you're simply trying to redefine "technology" in a way that makes this fact untrue. Over a longer timespan, say the past 1000 years, it is far more obvious how fast technological advancement is accelerating.
Other ideas : the same molecular tools could be used to create a complete copy of a frozen human brain. They would use the information gained about each synapse to create a molecular computer that emulated the dead person's mind. Presto, "AI". Singularity would happen within a year or so after this point.
Ok, so what if you had a machine that consisted of a flat plane of tiny nanoscale manipulator arms, built atom by atom using an atomic force microscope the first time. Essentially, the machine would be supplied with feedstock consisting of each pure element. The head containing the nanoscale manipulators would be able to lay down specific atoms at low temperature, and with itnermediates move electrons to create the chemical bonds. So, atom by atom the head could build up a physical object. Most notably, it could make a copy of itself.
Now, this nanoscale manufacturing head might be incredibly complex and might run very slow. However, if it could make a copy of itself, you'd have exponential growth on your side. After you make one, you could double the number of heads you have every cycle. If it took a week to manufacture a new head using an existing one, you'd have 4.5 quadrillion of these heads within a year.
I'm assuming a nanoscale manufacturing head is only a few hundred thousand atoms thick.
There is one problem : it would be very, very energy inefficient to make objects this way. Of course, if the nanoscale manufacturing heads also spend time making solar cells, you could generate enough energy to keep the whole process going.
For instance, you could send a single nanoscale manufacturing head, containers of starter feedstock, a nuclear reactor, ore processor, and a simple mining machine to the moon. The machine would gather lunar ore, which would be vaporized and separated using a mass spectrometer into individual element feedstock. The nanoscale manufacturing head would start making a second nanoscale manufacturing head... Those two heads would then make copies of themselves....eventually, you'd have enough heads that you would now need more energy to power them with. You'd put some of the heads to work making solar panels.
Eventually, you'd need more ore processors and minining tools. Again, some of the heads would be put to work making those...
Anyways, it's hard to say how fast the whole process would go without knowing how fast a nanoscale manufacturing head could work. Biological cell proteins are basically nanotech replicators, and a bacterial cell can replicate a complete copy of itself within an hour. So a week for a human designed one seems like a conservative estimate. I'd say within 1 to 5 years one could cover the entire lit lunar surface with solar panels and mining machines. At this point you could send up molecular blueprints for anything anyone on earth wanted manufactured, and within limits everyone on the planet would be able to get what they wanted. (obviously these factories would also manufacture reentry vehicles to get the finished products back to earth)
The weasel word is that it depends on how you define complexity. It is true for all things.
A solar pumped electric grid would eliminate a number of nasty problems with current electricity production. One is pollution, of course. Another is a dependence on fuel that occurs in finite deposits that are becoming increasingly labor intensive to recover. (since we can only get so much energy from hydro, and nuclear power is currently MORE expensive than solar)
A bulldozer requires a society to build it and maintain it that is vastly more complex than a simple camp of fruit pickers.
Yes, a modern PC is radically different. There are MANY more layers of abstraction in the code you are running right now...from BIOS layers that are nearly as complex as DOS was in 1989, to the OS kernel, GUI, browser, and now virtual machines running in the browser that are as complex as whole applications were in 1989. The main block components may look the same, but there's a LOT more in that CPU than there was in 1989, and a lot more memory cells in the DRAM as well.