Hasn't buying in to English Gentry been the ambition of the vulgar nouveau riche for at least a couple of centuries now? Back in the day, more than a few American industrial fortunes went into obtaining distressed-but-honorable lineages...
Libya used to be pretty good as well. In particular, Gaddafi's Amazonian Guard pushed him well into Bond Villian territory on style points alone, and his identifiable cult of personality and history of nefarious plotting were icing on the cake.
The new guys may or may not actually be less villainous; but they certainly are less colorful.
That, my friend, is a country where Bond villian rules.
Russia, on the other hand, pales in comparison.
You have to have some sort of ideosyncratic affectations and Rube Goldberg plots to go along with your general miasma of evil in order to be a 'Bond villian'. Having at least one henchman with identifiable qualities and a suitably snappy uniform for your expendable muscle is also a good idea. Neither Pakistan nor Saudi Arabia seem to remotely qualify on those counts. They get plenty of good, mundane, (halal)meat-and-potatoes evil done; but they don't really have any good cults of personality or sinister hijinks to go along with them. In post-Soviet Russia, you have a mediagenic stream of macho stunts by Putin, with occasional cuts to the Medvedev for comic relief, and the odd mysterious death among the living-in-posh-exile Russian plutocrats in London for a change of scene.
By contrast, neither Pakistan nor Saudi Arabia even rises to the level of having personally identifiable villains. Anonymous crowds of angry guys with scruffy beards? Sure, plenty of those; but no larger-than-life characters to work with.
I suspect that the big problem with that analogy is that data collection(unlike electrical wiring) is a substantially adversarial field.
There is a certain amount of tension, (fast, cheap, good, pick any two, and the usual buyer/seller desire to not leave money on the table); but the buyer and the seller both share roughly the same ideal, though they may deviate from it out of laziness, cheapness, or incompetence.
With data collection, the purely security/architectural aspects are somewhat similar; but there is the more fundamental problem that data collection is frequently not for the good of the collected. There is only the merest pretence of aligned interests, and it mostly is a matter of what the collector can get away with.
I imagine that the concern(if the 'complex systems' guys are on the case) is less a matter of a classic Malthusian mismatch between population growth and agricultural productivity, especially since developments in contraceptive technology post-malthus seem to suggest that people don't actually like shoving out babies until they start to starve, when an alternative exists; but about the same sort of problems that plague highways:
If you examine a road system under good conditions, you can get extremely favorable figures on carrying capacity and mean travel time. However, if a localized problem occurs, or even if a relatively small number of drivers do the wrong thing, you can get major slowdowns that ripple around and take a while to clear, even once their cause is resolved.
On a large scale, food production technology is in fine shape, and people don't actually seem to like outstripping their food supply; but that hardly precludes messy localized incidents(including some that might have the unfortunate side effect of reducing agricultural productivity where they occur, which could be a problem...)
The solution to these problems is to pit them against one another.
Simply cultivate a quantity of desperate people with nothing to lose who are willing to shoot pesky trespassers in exchange for a small cut of your food. Getting the implementation just right can be tricky, but this(along with appeals to the authority of the invisible friends of the powers that be) has been a fundamental part of human civilization for pretty much all of human history...
Does anybody know how much that would matter on this particular test?
No commercially available system(and quite likely no machine system yet developed) can actually parse natural languages especially well; but if the only Japanese is just boilerplate 'Name', 'Date', 'Solve for X and show your work', that won't really matter. If the test is larded with cunningly phrased word problems, by contrast...
Given that 'memory' is perhaps the area of AI we've had the most success with(not any of the fancy salience-based selective tricks; but quantity has a quality all its own), normal function might well be indistinguishable from the vast majority of human cheating.
I do not "reject" science as my socialist detractors may claim. Rather, I merely withhold my currency from the marketplace of ideas in order to incentivize the production of science more in line with today's consumer preferences!
Come back when team medicine has advanced to the point where the other guy being held liable is sufficient to scrape my central nervous system off the steering column... Ideally, bring a population of replacement humans who don't have a hilariously dodgy risk-discounting algorithm as a matter of empirical fact...
I'm told that, sometime after the great war; but before 2277, the crawler transporter is extensively modified to serve as a (slightly) mobile command post for Enclave remnant forces operating in the capitol wasteland area...
I wonder if future firmware revisions will incorporate all sorts of deeply-unsettling sudden change of direction capabilities based on being able to swap 'front' and 'rear' limb roles in short order...
There's still inertia to worry about; but something that moves forward or backward with equal speed and ease could pull off some interesting tricks.
I imagine that the mere existence of Raven, and his 'thermonuclear second strike is just a stroke away' deterrence policy, would keep the proliferation types too busy attempting to find underwear not sodden with human filth and pure fear to be worried about a few stray RTGs...
I suspect that much of the interest is also because of our persistent desire to more efficiently perform rescue operations and/or slaughter the locals in some downright hostile terrain.
Just as a pick-and-place provided with precise instructions and reels of neatly packaged and identical components can out-assemble a factory worth of nimble-fingered children; but couldn't beat a single freshman nerd at 'dig through the junk box and breadboard something', wheeled vehicles run like a bat out of hell on the terrain we lovingly build for them; but work increasingly poorly outside of that. At the cost of size and weight, larger wheels and/or tracks can muscle the problem a bit; but there are limits.
Legs, on the other hand, are mediocre at moving fast over well behaved terrain; but scrambling up mildly alarming slopes composed of loose rubble is practically routine...
CC 'non-commercial' strikes me as actually overwhelmingly different, in terms of objectives and in terms of the interactions being planned for(or against) than the more software-focused GPL/LGPL/BSD/MIT crowd:
With the software-focused licenses, the thinking(regardless of which camp you fall into on the questions of which interests you value most highly) is really about the relationship between the original developer, intermediate bundlers/distributors, and end users. Some licenses(ie. BSD) prefer to impose basically no restrictions beyond attribution on the intermediate users, considering it sufficient that the original developer can do what they want, and the end user(while they may or may not have any access to the guts of whatever binaries they get from the intermediate bundlers) does have access to the original project. LGPL is more aggressive about protecting the interests of the original project; by requiring intermediate distributors of works that include the original to make their changes available; but doesn't go much further than BSD in terms of watching over the end user. GPL explicitly demands that the end user's interest in access to the code be preserved, and the AGPL and GPL3 address the same interests in the context of 'cloud' and 'tivoized' scenarios.
However, across all of that, 'commercial/noncommercial' isn't an area of distinction. Obviously, these licenses have an effect on the viability of certain commercial schemes; but they have no explicit objectives in that area: If you can find a way to make money by selling software that is available in source form under the GPL3, rock on. If your freedom to add whatever proprietary features to FreeBSD can't save you from being crushed by generic upstream FreeBSD, sucks to be you.
By contrast, the whole idea of 'commercial' vs. 'noncommercial' reflects an explicit focus on matters that the software focused licenses leave implicit or simply consider an irrelevant detail to be worked out by what people can actually manage to succeed or fail at making money on. The closest analog, perhaps, is the bit in GPL compliance where it applies only if you distribute, not if you use internally. "Commercial/noncommercial" is a sort of much broader extension of that, positing that there is a 'noncommercial' area of culture-as-something-that-people-do that is separate from, and not merely a 'cottage industry' scaled version of the 'commercial' culture-as-something-that-media-industries-do, which is a monetary phenomenon and distinct both in degree and in kind from the noncommercial flavor.
In the context of software, I have to admit that I like the conceptual frameworks of the software-oriented licenses much better. This isn't to say that software cannot be a cultural thing; but most software is tools, complex tools, and it seems both appropriate and pragmatic to think about tools in terms of "How can we get good tools?" and "How can we ensure that our labor in producing tools ends up putting tools in the hands we want them to end up in?" OSS/FOSS licensing is a rather novel technique, born of the peculiar economics of software, for answering these questions; but it's actually a fairly conservative conceptual framework.
If anything, the 'Commercial'/'Noncommercial' framework is much more radical: The implicit assertion that there are areas of cultural output that are simply not amenable to resolution along the lines of market commerce(while certainly supportable, if Team Anthropologist fancies a trip back from the tribal regions to inform us that people who don't even have currency somehow manage to have music...); but it isn't necessarily something that would be a natural fit in ye olde contemporary consumer capitalist economies of the early 21st century... By contrast, CC-sharealike, CC-attribution, and similar are practically analogous to the software-oriented concepts, just written to cover non-software more cleanly, in that they don't posit this commercial/noncommercial distinction; but are more focused on either the GPL-style protection of the downstream user and the upstream developer, or the BSD style protection of attribution without significant interference in the distribution process.
Remember when people were proud to own USA items? Perhaps it is time to bring that back.
The minute "people" are willing to spend $1500 on a phone that currently costs $550, you'll see iPhones built right here in the good ol' USA. You can be the first. What's that? Not interested? Oh, sorry, never mind then, hypocrite.
And Apple(or any other hypothetical vendor) wouldn't just pocket the extra grand and continue production by the means that allowed them to hit the $500 price point why exactly?
I suspect that electromagnetic projectile propulsion gets a great deal more interesting if you enjoy the benefit of having nothing but a couple of meters of superconductive feed rail between you and your nuclear reactor(s), which I assume is why the navy is messing around with them. It's in the smaller scale areas where little packets of chemicals have been reliably killing people for several centuries now, and microfusion cells are still a Fallout 3 inventory item...
The one nice shortcut with the math is that most of the things humans are interested in are (comparatively) large gravity wells.
That still leaves you with the somewhat hairy problem of not falling in to the biggest star in the area when your millenia ship fires up its ancient engines as it approaches the target; but at least the dartboard in China is a powerful magnet, so there is a small envelope for near misses.
Personally, I'd be worried about the limits(both in terms of 'with our present technology' and in terms of fundamental thermodynamics) of materials science.
Even if kept very cold, complex chemical structures degrade over time(with a little help from any radiation zipping around in the endless void, naturally). Building machines, or preserving biological samples, such that they will be viable in 100,000 years or more, when the ship finally drifts to its target, could be a bit tricky...
If memory serves, 'junk' was some unfortunate-but-persistent description of non-coding regions(which are, indeed, the great majority of the genome); but that work on what exactly the regions that don't code do do has advanced considerably since then...
My concern with 3d printing(for some niches, for others it isn't an issue) is not so much with resolution and detail; but with materials. Unless team matter-compiler gets up to speed, 3d printing will be limited to materials that are amenable to the various high resolution localized binding methods. That's hardly a useless list(especially if you are willing to perform a second stage, as with the 'stainless steel' printing processes that use a temporary adhesive to bind the steel particles and then a molten brass impregnation step to produce a mixed steel/brass final product, or the various ceramic variations that bind the clay powder and glaze and then go through a firing step, or high resolution wax printing as a starting material for lost-wax casting in silver or other suitable casting media); but it does significantly limit your ability to manipulate certain very convenient materials, especially if you want the '3d printing should just be file -> object' thing, rather than treating 3d printing as one stage in a process incorporating other material processing techniques.
Some of the DLP-based photocure systems are already quite impressive(full resolution of the DLP projector in X and Y, limited by the opacity of the medium and the precision of the drive system on the Z axis); but they limit you entirely to putting out pieces of photocure resin. Even the very pricey selective laser sintering stuff only handles materials that take gracefully to being sintered. Again, this is hardly to say that these techniques are useless, they are far from it; but manufacturers don't use all kinds of (often rather horrid or deeply inconvenient) processes that don't fit nicely on the desktop just because they enjoy suffering; but because some useful materials are touchy and do not play nicely without a great deal of coddling. I don't doubt that the visible grains-and-ridges that currently identify cheaper 3d prints will go away, just as resolution issues have largely disappeared in printing; but things like annealing, heat treatment, various surface plating and finishing techniques, and similar not-desk-friendly techniques aren't going to be replicated until overwhelmingly more advanced control of deposition at very low levels becomes possible.
Life also gets interesting when we can print keys. To your house, your car, your safe deposit box....
It isn't polite to mention it, and people have set up a degree of obfuscation by using blind codes with proprietary conversion books or software; but you can reproduce (basic) keys with just the bitting codes and an appropriate blank. If you want EZ-while-U-wait, you'll need a key cutting machine; but a set of files and some calipers will work, if you don't mind building some character in the process.
If you have physical access to the original(rather than just a photo or set of bitting codes) various seriously cheesy casting techniques should also suffice to produce a working copy(plaster and any low or moderate temperature metal strong enough to take it, silicone and classier thermoplastics, probably a bar of soap and some JB Weld, if you are feeling improvisational...)
Machine vision algorithms that can infer bitting codes just from adequate photos of the rentacop's big bundle 'o keys are a new and exciting trick; but reproducing low cost keys is not the world's biggest news.
Slashdot Mirror
Hasn't buying in to English Gentry been the ambition of the vulgar nouveau riche for at least a couple of centuries now? Back in the day, more than a few American industrial fortunes went into obtaining distressed-but-honorable lineages...
Libya used to be pretty good as well. In particular, Gaddafi's Amazonian Guard pushed him well into Bond Villian territory on style points alone, and his identifiable cult of personality and history of nefarious plotting were icing on the cake.
The new guys may or may not actually be less villainous; but they certainly are less colorful.
That, my friend, is a country where Bond villian rules.
Russia, on the other hand, pales in comparison.
You have to have some sort of ideosyncratic affectations and Rube Goldberg plots to go along with your general miasma of evil in order to be a 'Bond villian'. Having at least one henchman with identifiable qualities and a suitably snappy uniform for your expendable muscle is also a good idea. Neither Pakistan nor Saudi Arabia seem to remotely qualify on those counts. They get plenty of good, mundane, (halal)meat-and-potatoes evil done; but they don't really have any good cults of personality or sinister hijinks to go along with them. In post-Soviet Russia, you have a mediagenic stream of macho stunts by Putin, with occasional cuts to the Medvedev for comic relief, and the odd mysterious death among the living-in-posh-exile Russian plutocrats in London for a change of scene.
By contrast, neither Pakistan nor Saudi Arabia even rises to the level of having personally identifiable villains. Anonymous crowds of angry guys with scruffy beards? Sure, plenty of those; but no larger-than-life characters to work with.
I suspect that the big problem with that analogy is that data collection(unlike electrical wiring) is a substantially adversarial field.
There is a certain amount of tension, (fast, cheap, good, pick any two, and the usual buyer/seller desire to not leave money on the table); but the buyer and the seller both share roughly the same ideal, though they may deviate from it out of laziness, cheapness, or incompetence.
With data collection, the purely security/architectural aspects are somewhat similar; but there is the more fundamental problem that data collection is frequently not for the good of the collected. There is only the merest pretence of aligned interests, and it mostly is a matter of what the collector can get away with.
I'm sure that Manifest Destiny can be pointed north, as well as west...
I imagine that the concern(if the 'complex systems' guys are on the case) is less a matter of a classic Malthusian mismatch between population growth and agricultural productivity, especially since developments in contraceptive technology post-malthus seem to suggest that people don't actually like shoving out babies until they start to starve, when an alternative exists; but about the same sort of problems that plague highways:
If you examine a road system under good conditions, you can get extremely favorable figures on carrying capacity and mean travel time. However, if a localized problem occurs, or even if a relatively small number of drivers do the wrong thing, you can get major slowdowns that ripple around and take a while to clear, even once their cause is resolved.
On a large scale, food production technology is in fine shape, and people don't actually seem to like outstripping their food supply; but that hardly precludes messy localized incidents(including some that might have the unfortunate side effect of reducing agricultural productivity where they occur, which could be a problem...)
The solution to these problems is to pit them against one another.
Simply cultivate a quantity of desperate people with nothing to lose who are willing to shoot pesky trespassers in exchange for a small cut of your food. Getting the implementation just right can be tricky, but this(along with appeals to the authority of the invisible friends of the powers that be) has been a fundamental part of human civilization for pretty much all of human history...
Does anybody know how much that would matter on this particular test?
No commercially available system(and quite likely no machine system yet developed) can actually parse natural languages especially well; but if the only Japanese is just boilerplate 'Name', 'Date', 'Solve for X and show your work', that won't really matter. If the test is larded with cunningly phrased word problems, by contrast...
Given that 'memory' is perhaps the area of AI we've had the most success with(not any of the fancy salience-based selective tricks; but quantity has a quality all its own), normal function might well be indistinguishable from the vast majority of human cheating.
"Lack of speed can be overcome. In the worst case by patience."
What do you think the crawler will be busy doing between now and 2277?
I do not "reject" science as my socialist detractors may claim. Rather, I merely withhold my currency from the marketplace of ideas in order to incentivize the production of science more in line with today's consumer preferences!
Come back when team medicine has advanced to the point where the other guy being held liable is sufficient to scrape my central nervous system off the steering column... Ideally, bring a population of replacement humans who don't have a hilariously dodgy risk-discounting algorithm as a matter of empirical fact...
I'm told that, sometime after the great war; but before 2277, the crawler transporter is extensively modified to serve as a (slightly) mobile command post for Enclave remnant forces operating in the capitol wasteland area...
I wonder if future firmware revisions will incorporate all sorts of deeply-unsettling sudden change of direction capabilities based on being able to swap 'front' and 'rear' limb roles in short order...
There's still inertia to worry about; but something that moves forward or backward with equal speed and ease could pull off some interesting tricks.
I imagine that the mere existence of Raven, and his 'thermonuclear second strike is just a stroke away' deterrence policy, would keep the proliferation types too busy attempting to find underwear not sodden with human filth and pure fear to be worried about a few stray RTGs...
I suspect that much of the interest is also because of our persistent desire to more efficiently perform rescue operations and/or slaughter the locals in some downright hostile terrain.
Just as a pick-and-place provided with precise instructions and reels of neatly packaged and identical components can out-assemble a factory worth of nimble-fingered children; but couldn't beat a single freshman nerd at 'dig through the junk box and breadboard something', wheeled vehicles run like a bat out of hell on the terrain we lovingly build for them; but work increasingly poorly outside of that. At the cost of size and weight, larger wheels and/or tracks can muscle the problem a bit; but there are limits.
Legs, on the other hand, are mediocre at moving fast over well behaved terrain; but scrambling up mildly alarming slopes composed of loose rubble is practically routine...
CC 'non-commercial' strikes me as actually overwhelmingly different, in terms of objectives and in terms of the interactions being planned for(or against) than the more software-focused GPL/LGPL/BSD/MIT crowd:
With the software-focused licenses, the thinking(regardless of which camp you fall into on the questions of which interests you value most highly) is really about the relationship between the original developer, intermediate bundlers/distributors, and end users. Some licenses(ie. BSD) prefer to impose basically no restrictions beyond attribution on the intermediate users, considering it sufficient that the original developer can do what they want, and the end user(while they may or may not have any access to the guts of whatever binaries they get from the intermediate bundlers) does have access to the original project. LGPL is more aggressive about protecting the interests of the original project; by requiring intermediate distributors of works that include the original to make their changes available; but doesn't go much further than BSD in terms of watching over the end user. GPL explicitly demands that the end user's interest in access to the code be preserved, and the AGPL and GPL3 address the same interests in the context of 'cloud' and 'tivoized' scenarios.
However, across all of that, 'commercial/noncommercial' isn't an area of distinction. Obviously, these licenses have an effect on the viability of certain commercial schemes; but they have no explicit objectives in that area: If you can find a way to make money by selling software that is available in source form under the GPL3, rock on. If your freedom to add whatever proprietary features to FreeBSD can't save you from being crushed by generic upstream FreeBSD, sucks to be you.
By contrast, the whole idea of 'commercial' vs. 'noncommercial' reflects an explicit focus on matters that the software focused licenses leave implicit or simply consider an irrelevant detail to be worked out by what people can actually manage to succeed or fail at making money on. The closest analog, perhaps, is the bit in GPL compliance where it applies only if you distribute, not if you use internally. "Commercial/noncommercial" is a sort of much broader extension of that, positing that there is a 'noncommercial' area of culture-as-something-that-people-do that is separate from, and not merely a 'cottage industry' scaled version of the 'commercial' culture-as-something-that-media-industries-do, which is a monetary phenomenon and distinct both in degree and in kind from the noncommercial flavor.
In the context of software, I have to admit that I like the conceptual frameworks of the software-oriented licenses much better. This isn't to say that software cannot be a cultural thing; but most software is tools, complex tools, and it seems both appropriate and pragmatic to think about tools in terms of "How can we get good tools?" and "How can we ensure that our labor in producing tools ends up putting tools in the hands we want them to end up in?" OSS/FOSS licensing is a rather novel technique, born of the peculiar economics of software, for answering these questions; but it's actually a fairly conservative conceptual framework.
If anything, the 'Commercial'/'Noncommercial' framework is much more radical: The implicit assertion that there are areas of cultural output that are simply not amenable to resolution along the lines of market commerce(while certainly supportable, if Team Anthropologist fancies a trip back from the tribal regions to inform us that people who don't even have currency somehow manage to have music...); but it isn't necessarily something that would be a natural fit in ye olde contemporary consumer capitalist economies of the early 21st century... By contrast, CC-sharealike, CC-attribution, and similar are practically analogous to the software-oriented concepts, just written to cover non-software more cleanly, in that they don't posit this commercial/noncommercial distinction; but are more focused on either the GPL-style protection of the downstream user and the upstream developer, or the BSD style protection of attribution without significant interference in the distribution process.
Remember when people were proud to own USA items? Perhaps it is time to bring that back.
The minute "people" are willing to spend $1500 on a phone that currently costs $550, you'll see iPhones built right here in the good ol' USA. You can be the first. What's that? Not interested? Oh, sorry, never mind then, hypocrite.
And Apple(or any other hypothetical vendor) wouldn't just pocket the extra grand and continue production by the means that allowed them to hit the $500 price point why exactly?
I suspect that electromagnetic projectile propulsion gets a great deal more interesting if you enjoy the benefit of having nothing but a couple of meters of superconductive feed rail between you and your nuclear reactor(s), which I assume is why the navy is messing around with them. It's in the smaller scale areas where little packets of chemicals have been reliably killing people for several centuries now, and microfusion cells are still a Fallout 3 inventory item...
The NHS should begin a program of providing him with a homeopathic salary. The less they pay him, the more motivated he will become!
The one nice shortcut with the math is that most of the things humans are interested in are (comparatively) large gravity wells.
That still leaves you with the somewhat hairy problem of not falling in to the biggest star in the area when your millenia ship fires up its ancient engines as it approaches the target; but at least the dartboard in China is a powerful magnet, so there is a small envelope for near misses.
Personally, I'd be worried about the limits(both in terms of 'with our present technology' and in terms of fundamental thermodynamics) of materials science.
Even if kept very cold, complex chemical structures degrade over time(with a little help from any radiation zipping around in the endless void, naturally). Building machines, or preserving biological samples, such that they will be viable in 100,000 years or more, when the ship finally drifts to its target, could be a bit tricky...
If memory serves, 'junk' was some unfortunate-but-persistent description of non-coding regions(which are, indeed, the great majority of the genome); but that work on what exactly the regions that don't code do do has advanced considerably since then...
My concern with 3d printing(for some niches, for others it isn't an issue) is not so much with resolution and detail; but with materials. Unless team matter-compiler gets up to speed, 3d printing will be limited to materials that are amenable to the various high resolution localized binding methods. That's hardly a useless list(especially if you are willing to perform a second stage, as with the 'stainless steel' printing processes that use a temporary adhesive to bind the steel particles and then a molten brass impregnation step to produce a mixed steel/brass final product, or the various ceramic variations that bind the clay powder and glaze and then go through a firing step, or high resolution wax printing as a starting material for lost-wax casting in silver or other suitable casting media); but it does significantly limit your ability to manipulate certain very convenient materials, especially if you want the '3d printing should just be file -> object' thing, rather than treating 3d printing as one stage in a process incorporating other material processing techniques.
Some of the DLP-based photocure systems are already quite impressive(full resolution of the DLP projector in X and Y, limited by the opacity of the medium and the precision of the drive system on the Z axis); but they limit you entirely to putting out pieces of photocure resin. Even the very pricey selective laser sintering stuff only handles materials that take gracefully to being sintered. Again, this is hardly to say that these techniques are useless, they are far from it; but manufacturers don't use all kinds of (often rather horrid or deeply inconvenient) processes that don't fit nicely on the desktop just because they enjoy suffering; but because some useful materials are touchy and do not play nicely without a great deal of coddling. I don't doubt that the visible grains-and-ridges that currently identify cheaper 3d prints will go away, just as resolution issues have largely disappeared in printing; but things like annealing, heat treatment, various surface plating and finishing techniques, and similar not-desk-friendly techniques aren't going to be replicated until overwhelmingly more advanced control of deposition at very low levels becomes possible.
As with electric cars and aircraft, the power density of boring, smelly chemical fuels are just stubbornly competitive with electric tricks...
It's a pity, because they are much more entertaining; but it's persistently the case.
Life also gets interesting when we can print keys. To your house, your car, your safe deposit box....
It isn't polite to mention it, and people have set up a degree of obfuscation by using blind codes with proprietary conversion books or software; but you can reproduce (basic) keys with just the bitting codes and an appropriate blank. If you want EZ-while-U-wait, you'll need a key cutting machine; but a set of files and some calipers will work, if you don't mind building some character in the process.
If you have physical access to the original(rather than just a photo or set of bitting codes) various seriously cheesy casting techniques should also suffice to produce a working copy(plaster and any low or moderate temperature metal strong enough to take it, silicone and classier thermoplastics, probably a bar of soap and some JB Weld, if you are feeling improvisational...)
Machine vision algorithms that can infer bitting codes just from adequate photos of the rentacop's big bundle 'o keys are a new and exciting trick; but reproducing low cost keys is not the world's biggest news.