You can. It's the underpinning of NEMO (NEtwork MObility), provided the means by which Telebit routers allowed you to make network segmentation totally invisible to the routing protocol, is fundamental to IPv4/IPv6 mapping, is key to creating private networks, and is built in to the notion of transient addressing schemes. It's one thing if people don't want to use the mechanisms that exist, but it's another to imagine that non-use is the same as non-presence. That's more than a bit unfair.
Addressing is this teeny tiny eenie weenie ittie bittie fragment of the changes involved in IPv6. I wish people would stop going on about it, it's an utterly insignificant component. And even if it were important, addressing is heirarchical by design (provided you use automatic addressing) and the bulk of problems involving it were considered solved by the 6Bone group at the time the protocol went native on the backbone. Routing on IPv6 is far simpler than on IPv4. It's also faster, because routing tables can be much smaller, which in turn is largely because there are far fewer special cases to consider.
But if you do want to delve into addressing, why not consider the greater range of multicast addresses? Or the fact that the automatic addressing scheme is ideal for mobile networking? Or the fact that automatic configuration eliminates many of the problems with network administrating? Or the fact that sparse address tables are easier to maintain?
Hell, if you only want to consider the addressing aspect, why not be relieved that TUBA was abandoned as the IP-ng protocol?
Yes, it took me a few months, running 2.4.20 and the IPv6 patches, back in 1996. Since then, the software has improved, support in applications is so much better, and many grey areas have been cleaned up. It would probably take a few days to migrate a network of reasonable size today. Maybe a week at most.
(By comparison, it took about 1.5 years for the US Navy to switch from one e-mail system to a more secure alternative, due to reliability issues, security problems and brain-dead contracting.)
...this is important (beyond the address count issue) for the Feds specifically:
IPv6 has better security provisions within the protocol itself, making the usual run of D- through to F- on Federal security audits less likely.
The protocol incorporates many of the features back-engineered into IPv4 as standard, producing a cleaner design with fewer compromises and fewer flaws
Built-in support for protocol expansion means future updates should have less impact and be adoptable faster
Automatic configuration means fewer errors and less maintenance
Alignment of entries in the header means potentially greater throughput
Skript Kiddies will end up jumping off bridges as they won't know what to do
Software contracting firms are located in regions in which elections are due, creating excellent opportunities on both sides of the table
There have been some changes, if not drastic ones. Blended Wing Body and Waverider airframe designs offer a major shift from conventional designs. Basically, in the case of Waveriders, the aircraft skims on top of its own supersonic shockwave. For fuels, ramjets and scramjets would be better off with a hydrogen fuel rather than a conventional fuel - much greater range of speeds, for example. There may be drag-reduction ideas that can be borrowed from the various racing worlds where it's a lot easier and cheaper to experiment wildly.
What about massively different paradigms? Well, it's no secret that British Aerospace and Boeing are researching anti-gravity. Chances are, they won't succeed at that, but rather that there might be useful spin-off ideas. If there was an efficient way of ionising the incoming air, you could replace a chemical reaction engine with a linear accelerator, which means you could use energy sources with better originating-mass-to-useful-energy ratios, the engines would likely be a lot quieter, and pollution would be reduced. On the other hand, even just the first step is very very hard.
Are there other avenues? Maybe. Zepplin-style airships are being re-examined, particularly for non-time-critical journeys. Ornithopters are being evaluated. Flying disks are being re-evaluated. A claim for first flight from a Welsh minister talked of yet other experimental airframes that have never been repeated or developed. There is therefore room.
Since the core is software-defined, there should be relatively little difficulty in defining a bus abstraction and then having modules for HyperTransport, PCI Express, or any other bus they care to support, where perhaps only one module can be included at "compile-time". Why go for something this complex? It makes things future-proof. If a new bus or a new revision of a bus becomes the "in-thing", you only change the module. The chip architecture remains fundamentally the same. (Both HyperTransport and PCI Express have been through multiple revisions, with next to nobody using the latest. They are likely to be revised extensively in the future. Laying out and validating a CPU is non-trivial, so the less you need to do to support such changes, the better.)
I looked at similar data rates when working as a programmer for CERN's EUROGAM project out at Daresbury Laboratory. That was using a 68040-based VME crate running VxWorks for the real-time stuff and a Sparcstation for user-side processing. (During a test run, I ended up killing the nuclear structure facility's internal network, with the data collection and storage software running around on the CPU giggling.)
Modern high-end processors should be able to highly sophisticated real-time processing on the data without working up a sweat. A 68040 is how old? Yes, you could do quad CPU VME processor cards back then - can't remember if we did. I seem to remember you could even do 8-way. Even so, those were hardly gigahertz processors with modern FPU cores. What more should software be capable of today?
You cannot have awe without wonder, you cannot have wonder without curiosity and a delight in asking. The best scientists are, in many ways, childlike in their need to ask and discover. The same is true of geeks, inventors and other such individuals that society routinely calls "mad", "eccentric", "crazy" or "abnormal". Which is why geniuses are also usually outcasts from the very societies that revere them.
Is discovery possible without such qualities? I'd say no - you wouldn't think of asking enough questions until you hit on the key, fundamental questions, let alone gotten around to answering them. Which is why "normal" people don't usually make key discoveries. Not all science costs billions, some can cost next to nothing. But who is going to try for even the nearly-free discoveries if they're not hyped up about knowing the answers? Discoveries don't make one rich, often have no impact on the quality-of-life, won't interest any friends at the social clubs, and won't grab the interest of hotly desired individuals. That only leaves those people who discover because discovering is what they do and who they are. In other words, the people who have all the curiosity and tenacity of a child but also have the maturity and resources to do something more than bug others with questions.
Wonder would cease to be in short supply if the child-like qualities that make or break the discovery power of a person were nurtured and not crippled by societies and schooling. There seems little point in learning how to do something if doing so destroys your ability to learn why you learned it.
Then I guess one area that open source hardware nerds can work on is to rework the FPU to be faster. If people can load the files into an FPGA and get comparable results on maths-heavy software to a full hardware implementation of the T2 as it stands, I would imagine the scientific computing folk would go for the FPGA solution as it would be cheaper so they could build more nodes for the same amount of money.
Sun is a lot of things, some unprintable, but stupid isn't one of them. If it can be shown that a T2 with stronger maths will sell better than the T2 as-is, then you will see a T2 with stronger maths in very short order.
There are other things Sun could include in the processor. I am, to this day, a devoted fan of direct CPU-to-CPU channels for multi-processor systems. Inmos' Transputer let you build hypercubes of processors as large as you liked without scaling issues. Xyron's ZOTS also seems an interesting technology, even if nobody uses it at this time. The potential wishlist of things that could be added without wrecking the design is large. Given that the core is GPLed, it would seem to make sense to experiment with some of those ideas. See what would actually work in practice, with the possibility that some vendor (not necessarily Sun) will chase the idea and turn it into a reality.
The Ghost of Infocom cries out in the wind.
The Ghost of Infocom moves north.
The Ghost of Infocom enters.
The Ghost of Infocom hits Blizzard with the long sword.
Neither. "Not acceptable" in the sense that the Open Source paradigm approaches total ineffectiveness when the number of eyes approaches zero. Also "not acceptable" in the sense that communities are no stronger than the weakest element within them; by having high quality software that the community may theoretically use but in practice can't/won't, the community is in a state of wholly unnecessary weakness.
Personality is important, although arguably it should not be. Crick and Watson did minimal work, but virtually owned DNA research on the basis of a personality cult. Bill Gates and Linus Torvalds have nothing in common, except they too had strong personalities that turned cult projects into international software celebrities. Linux did not achieve results by merit any more than Windows did. If merit mattered, Linux and the *BSDs would own the server market entirely and the desktop market would be split equally between Linux and Mac OS/X. (You would also never see a salesperson with more polish than a shoe-shine specialist. They wouldn't need it if the product mattered.)
We're not in that world, sadly. And yes, I mean sadly. The greatest geeks typically are on the Autistic Spectrum and have one or more quirks that place them outside "normal" society. This is why society has history books riddled with disaffected, socially-abandoned, neglected, ignored/abused geniuses. Those aren't by chance, those are because thinkers and doers are almost always alien to society as a whole.
This is why it's important for the personality figures to run with projects by such individuals, and why I blame the lack of such people for the failure of such projects. The stereotypical loner "mad scientist" doesn't need a manager, they need a promoter. A few can both do and promote, but those are exceptionally rare. There are probably a hundred Teslas for every Eddison.
OpenBSD could be light-years ahead of where it is in development and deployment, but companies are wary of firebrands, no matter how good they are. Stupid, yes - quality should matter more than glamour - but a product of the human mind. Since efforts to debug the mind have largely floundered, the lone geniuses will need to turn to the personality cults for help or be doomed to senseless obscurity.
(Some would prefer senseless obscurity, but I go back to society on that. The tools available are the limiting factor on what society can do, and a withheld tool is just as damaging as a withheld bugfix, for the same reasons. It is just as morally wrong to cripple the distribution of a valuable tool as it is to cripple the tool itself.)
The crypto software and FFT software especially so, but maintenance isn't always as hot. That's hardly DJB's fault - they are public domain and nobody has run with them. On the other hand, it is not acceptable that his software is not being properly distributed, promoted or documented. Nor is it acceptable that he allows his personality quirks to interfere with the primary purpose of getting code into active circulation.
I have no problem with employees giving up their privacy... so long as ALL managers realize they are also employees (and so must give up their privacy too), that said managers lead from the front (ie: give up their privacy first) and prove that they have done so by publishing significant personal information - including protected information - on a public site. Ideally, those who are at the most senior ranks of NASA should do so likewise, along with those members of Homeland Security responsible for oversight.
When it is only the low-level employees who are vulnerable, then it is an instrument of terror, not an instrument to prevent terror. Even mid-levels are not really enough, although it is better. Management, not underlings, are the primary cause of security leaks and the primary cause of poor security practices. They are therefore the ones in greatest need of scrutiny. Not only to prevent these flaws, but also to prevent those managers from misusing knowledge to cover up their ineptitude.
You're missing a few other optimizations. I suggest modifying the makefile to delete everything on the system, pipe out a short "hello world"-type program that reports that the system is out of memory, and compile just that.
This is impossible. I'm now up to two posts by circletimessquare on Slashdot that I've publicly agreed with. Well, four more to go and I can cash in on some free tickets to Milliways.
Seriously, the header information states that this is generated by a PHP script, not by an e-mail client. Now, I can't expect the judge to interpret SMTP headers, but provided this was pointed out, it beggars belief that the judge could rule the e-mail as personal on any kind of "obviousness" test.
Ok, I'll buy your argument. My rationale for QC was partly based on the heavy need for floating-point. An analog computer can represent a floating-point number as accurately as you like - just reduce the noise to the required level. You can't quite do this using multi-state logic, but you can approximate it if you have enough discrete states. Quantum computers use quantum states which gives you the multi-state. It was also based on the premise that for a given level of complexity of network, there will be N different networks that can subdivide the problem-space meaningfully given the same level of training. A classical computer starts in a random state and then works to the nearest state through the training. If you don't get the final state you want, you re-randomize and re-try. A quantum computer would (in theory) be able to represent all valid networks and then discard the ones that aren't useful, which means you'd be able to get a deterministic result as there would be no random element.
Hmmm. It may be possible, especially if it's basic pattern recognition and other such image processing techniques (eg: contrast stretching). I guess I'm thinking in terms of signals to be sent to the rest of the brain, in the sense that if we know the bit pattern being sent to the rest of the brain for a given input, all we want to do is duplicate that output. We don't really care what the reason is for that bit pattern, so long as we can generate it. That's a classic neural net training set, if ever I saw one - you can log the output for the input and just replay that to produce something that is functionally the same.
If we do the image processing directly, in code, we'd get something that is (a) faster, (b) more compact, and (c) probably much less prone to the errors of interpretation that happen in the visual cortex. However, getting the output into a suitable format is going to be much tougher, as a lookup table would not make sense. It would be too big. That is probably going to be harder than the image processing itself, as the mappings are not necessarily going to be logical, they will merely be whatever evolved for the different connections as the brain has evolved.
The artificial limbs offer some hope that the brain is far more logical than I fear, but it should be kept in mind that that's output from the motor neurons to the nerves that control the muscles. That part is going to be relatively basic signaling, but it's not clear to me that the rest of the brain is as simple. If it is, then this will indeed be solved in 15 years or less.
The reason I'm guessing quantum computing is that neural networks don't scale very well on traditional computers. A few thousand, maybe a few tens of thousands, of neurons can be put into a non-linear back-propagating neural network, but that's getting on for the upper limit for a computer small enough to be portable. You could build larger self-training networks, or trainable networks on big iron, but the former can't be used when you want specific inputs to generate very specific outputs, and the latter would only be useful if you don't plan on moving much.
However, the visual cortex is probably in the order of a few tens of million neurons. Since the compute power needed is (at best) superlinear, that would mean a computer over a thousand times as powerful as the best portable machines today, even if you take the very best case possible and have a truly simple neural net. If the simple net is enough, then the computing power will be available in around 15-20 years, according to Moore's Law. If we're looking at a complex net and exponential increases in CPU requirements, then you're looking at an exponential increase in CPU requirements and that'd require another 1,500-1,800 years of development.
If the simple case holds up, you don't need quantum computers and the technology will exist by the time the medical industry is at the point of being able to make use of it. In the complex case, traditional computers won't be ready for so long that you can forget about that avenue. The only way you could do it is with quantum computers, which will bring the time requirement down substantially - maybe to again just a few decades - once the technology matures enough for this kind of work. My guess is that the visual cortex is so complex and so difficult to replicate that this will be closer to what actually happens in practice.
...may be achievable for certain forms of blindness. Neurologists can decipher the signals sent over the optic nerve in the case of a cat. If that advances far enough to induce signals, then damage to the lens or retina could be bypassed. We've certainly got light-sensitive devices with the resolution and speed.
Damage to the optic nerve would be tougher, as it's hard to induce a signal in something that's dead. That would require placing the signal directly onto the brain's visual cortex, which would be a much tougher problem. Theoretically solvable, though, as modern 9 Tesla MRIs can actually observe individual neurons firing and implants can monitor signals. Based on a crude timeline I've drawn up for the average time from theory to practice, we should see this sort of technology emerge into highly specialist settings in about 18 years and into practical (though not necessarily approved) medical applications in about 48 years.
Damage to the visual cortex - or the claiming of the neurons that would normally go for that towards some other function - would require a real-time simulation, plus induction of a signal at multiple points. We're about 50 years away from a quantum computer with that level of capability, and probably another 20 from being able to use quantum computers in such a setting. Even then, the brain's wiring would only be suitable until some time in the teens, when it is not just flexible but actually growing. By the age of 24 or so, it starts to die back. It can still be programmed some, according to the latest research, but it's nowhere near the same level.
Slashdot Mirror
You can. It's the underpinning of NEMO (NEtwork MObility), provided the means by which Telebit routers allowed you to make network segmentation totally invisible to the routing protocol, is fundamental to IPv4/IPv6 mapping, is key to creating private networks, and is built in to the notion of transient addressing schemes. It's one thing if people don't want to use the mechanisms that exist, but it's another to imagine that non-use is the same as non-presence. That's more than a bit unfair.
But if you do want to delve into addressing, why not consider the greater range of multicast addresses? Or the fact that the automatic addressing scheme is ideal for mobile networking? Or the fact that automatic configuration eliminates many of the problems with network administrating? Or the fact that sparse address tables are easier to maintain?
Hell, if you only want to consider the addressing aspect, why not be relieved that TUBA was abandoned as the IP-ng protocol?
(By comparison, it took about 1.5 years for the US Navy to switch from one e-mail system to a more secure alternative, due to reliability issues, security problems and brain-dead contracting.)
What about massively different paradigms? Well, it's no secret that British Aerospace and Boeing are researching anti-gravity. Chances are, they won't succeed at that, but rather that there might be useful spin-off ideas. If there was an efficient way of ionising the incoming air, you could replace a chemical reaction engine with a linear accelerator, which means you could use energy sources with better originating-mass-to-useful-energy ratios, the engines would likely be a lot quieter, and pollution would be reduced. On the other hand, even just the first step is very very hard.
Are there other avenues? Maybe. Zepplin-style airships are being re-examined, particularly for non-time-critical journeys. Ornithopters are being evaluated. Flying disks are being re-evaluated. A claim for first flight from a Welsh minister talked of yet other experimental airframes that have never been repeated or developed. There is therefore room.
Since the core is software-defined, there should be relatively little difficulty in defining a bus abstraction and then having modules for HyperTransport, PCI Express, or any other bus they care to support, where perhaps only one module can be included at "compile-time". Why go for something this complex? It makes things future-proof. If a new bus or a new revision of a bus becomes the "in-thing", you only change the module. The chip architecture remains fundamentally the same. (Both HyperTransport and PCI Express have been through multiple revisions, with next to nobody using the latest. They are likely to be revised extensively in the future. Laying out and validating a CPU is non-trivial, so the less you need to do to support such changes, the better.)
Modern high-end processors should be able to highly sophisticated real-time processing on the data without working up a sweat. A 68040 is how old? Yes, you could do quad CPU VME processor cards back then - can't remember if we did. I seem to remember you could even do 8-way. Even so, those were hardly gigahertz processors with modern FPU cores. What more should software be capable of today?
Is discovery possible without such qualities? I'd say no - you wouldn't think of asking enough questions until you hit on the key, fundamental questions, let alone gotten around to answering them. Which is why "normal" people don't usually make key discoveries. Not all science costs billions, some can cost next to nothing. But who is going to try for even the nearly-free discoveries if they're not hyped up about knowing the answers? Discoveries don't make one rich, often have no impact on the quality-of-life, won't interest any friends at the social clubs, and won't grab the interest of hotly desired individuals. That only leaves those people who discover because discovering is what they do and who they are. In other words, the people who have all the curiosity and tenacity of a child but also have the maturity and resources to do something more than bug others with questions.
Wonder would cease to be in short supply if the child-like qualities that make or break the discovery power of a person were nurtured and not crippled by societies and schooling. There seems little point in learning how to do something if doing so destroys your ability to learn why you learned it.
Sun is a lot of things, some unprintable, but stupid isn't one of them. If it can be shown that a T2 with stronger maths will sell better than the T2 as-is, then you will see a T2 with stronger maths in very short order.
There are other things Sun could include in the processor. I am, to this day, a devoted fan of direct CPU-to-CPU channels for multi-processor systems. Inmos' Transputer let you build hypercubes of processors as large as you liked without scaling issues. Xyron's ZOTS also seems an interesting technology, even if nobody uses it at this time. The potential wishlist of things that could be added without wrecking the design is large. Given that the core is GPLed, it would seem to make sense to experiment with some of those ideas. See what would actually work in practice, with the possibility that some vendor (not necessarily Sun) will chase the idea and turn it into a reality.
Where's the sticky-back plastic? If she's going to pull a Valerie Singleton, she needs sticky-back plastic! And where's the one she made earlier?
The Ghost of Infocom cries out in the wind.
The Ghost of Infocom moves north.
The Ghost of Infocom enters.
The Ghost of Infocom hits Blizzard with the long sword.
Personality is important, although arguably it should not be. Crick and Watson did minimal work, but virtually owned DNA research on the basis of a personality cult. Bill Gates and Linus Torvalds have nothing in common, except they too had strong personalities that turned cult projects into international software celebrities. Linux did not achieve results by merit any more than Windows did. If merit mattered, Linux and the *BSDs would own the server market entirely and the desktop market would be split equally between Linux and Mac OS/X. (You would also never see a salesperson with more polish than a shoe-shine specialist. They wouldn't need it if the product mattered.)
We're not in that world, sadly. And yes, I mean sadly. The greatest geeks typically are on the Autistic Spectrum and have one or more quirks that place them outside "normal" society. This is why society has history books riddled with disaffected, socially-abandoned, neglected, ignored/abused geniuses. Those aren't by chance, those are because thinkers and doers are almost always alien to society as a whole.
This is why it's important for the personality figures to run with projects by such individuals, and why I blame the lack of such people for the failure of such projects. The stereotypical loner "mad scientist" doesn't need a manager, they need a promoter. A few can both do and promote, but those are exceptionally rare. There are probably a hundred Teslas for every Eddison.
OpenBSD could be light-years ahead of where it is in development and deployment, but companies are wary of firebrands, no matter how good they are. Stupid, yes - quality should matter more than glamour - but a product of the human mind. Since efforts to debug the mind have largely floundered, the lone geniuses will need to turn to the personality cults for help or be doomed to senseless obscurity.
(Some would prefer senseless obscurity, but I go back to society on that. The tools available are the limiting factor on what society can do, and a withheld tool is just as damaging as a withheld bugfix, for the same reasons. It is just as morally wrong to cripple the distribution of a valuable tool as it is to cripple the tool itself.)
The crypto software and FFT software especially so, but maintenance isn't always as hot. That's hardly DJB's fault - they are public domain and nobody has run with them. On the other hand, it is not acceptable that his software is not being properly distributed, promoted or documented. Nor is it acceptable that he allows his personality quirks to interfere with the primary purpose of getting code into active circulation.
When it is only the low-level employees who are vulnerable, then it is an instrument of terror, not an instrument to prevent terror. Even mid-levels are not really enough, although it is better. Management, not underlings, are the primary cause of security leaks and the primary cause of poor security practices. They are therefore the ones in greatest need of scrutiny. Not only to prevent these flaws, but also to prevent those managers from misusing knowledge to cover up their ineptitude.
You're missing a few other optimizations. I suggest modifying the makefile to delete everything on the system, pipe out a short "hello world"-type program that reports that the system is out of memory, and compile just that.
Seriously, the header information states that this is generated by a PHP script, not by an e-mail client. Now, I can't expect the judge to interpret SMTP headers, but provided this was pointed out, it beggars belief that the judge could rule the e-mail as personal on any kind of "obviousness" test.
Ok, I'll buy your argument. My rationale for QC was partly based on the heavy need for floating-point. An analog computer can represent a floating-point number as accurately as you like - just reduce the noise to the required level. You can't quite do this using multi-state logic, but you can approximate it if you have enough discrete states. Quantum computers use quantum states which gives you the multi-state. It was also based on the premise that for a given level of complexity of network, there will be N different networks that can subdivide the problem-space meaningfully given the same level of training. A classical computer starts in a random state and then works to the nearest state through the training. If you don't get the final state you want, you re-randomize and re-try. A quantum computer would (in theory) be able to represent all valid networks and then discard the ones that aren't useful, which means you'd be able to get a deterministic result as there would be no random element.
If we do the image processing directly, in code, we'd get something that is (a) faster, (b) more compact, and (c) probably much less prone to the errors of interpretation that happen in the visual cortex. However, getting the output into a suitable format is going to be much tougher, as a lookup table would not make sense. It would be too big. That is probably going to be harder than the image processing itself, as the mappings are not necessarily going to be logical, they will merely be whatever evolved for the different connections as the brain has evolved.
The artificial limbs offer some hope that the brain is far more logical than I fear, but it should be kept in mind that that's output from the motor neurons to the nerves that control the muscles. That part is going to be relatively basic signaling, but it's not clear to me that the rest of the brain is as simple. If it is, then this will indeed be solved in 15 years or less.
Damn, you can't fit lions before tigers and bears because they're social. Gah! Talk about life not imitating art.
However, the visual cortex is probably in the order of a few tens of million neurons. Since the compute power needed is (at best) superlinear, that would mean a computer over a thousand times as powerful as the best portable machines today, even if you take the very best case possible and have a truly simple neural net. If the simple net is enough, then the computing power will be available in around 15-20 years, according to Moore's Law. If we're looking at a complex net and exponential increases in CPU requirements, then you're looking at an exponential increase in CPU requirements and that'd require another 1,500-1,800 years of development.
If the simple case holds up, you don't need quantum computers and the technology will exist by the time the medical industry is at the point of being able to make use of it. In the complex case, traditional computers won't be ready for so long that you can forget about that avenue. The only way you could do it is with quantum computers, which will bring the time requirement down substantially - maybe to again just a few decades - once the technology matures enough for this kind of work. My guess is that the visual cortex is so complex and so difficult to replicate that this will be closer to what actually happens in practice.
Doctor, conducting evaluation for prosthetic forehead: Are you half-Klingon on your mother's side or your father's side?
Damage to the optic nerve would be tougher, as it's hard to induce a signal in something that's dead. That would require placing the signal directly onto the brain's visual cortex, which would be a much tougher problem. Theoretically solvable, though, as modern 9 Tesla MRIs can actually observe individual neurons firing and implants can monitor signals. Based on a crude timeline I've drawn up for the average time from theory to practice, we should see this sort of technology emerge into highly specialist settings in about 18 years and into practical (though not necessarily approved) medical applications in about 48 years.
Damage to the visual cortex - or the claiming of the neurons that would normally go for that towards some other function - would require a real-time simulation, plus induction of a signal at multiple points. We're about 50 years away from a quantum computer with that level of capability, and probably another 20 from being able to use quantum computers in such a setting. Even then, the brain's wiring would only be suitable until some time in the teens, when it is not just flexible but actually growing. By the age of 24 or so, it starts to die back. It can still be programmed some, according to the latest research, but it's nowhere near the same level.
...computers must indeed covert these pulses.
...Barney was the first intelligent robot and came very close to destroying humanity...
It has reached 80%? When did that happen?