...it had better have stars and monoliths. I was a fan of many of his books - Islands in the Sky, 2001, 2010, Rendezvous with Rama. They were brilliant, detailed, imaginative and really achieved what they set out to. Some of his other stuff - Cradle, 2061, Imperial Earth, and the later Rama books - didn't really appeal to me in the same way.
In terms of his factual writings, I have many of his articles that were written for Wireless World, including the letter and two follow-up articles on geostationary satellites. Those three in particular can be found on the web - many people have scanned them in. They're well worth reading. He was a highly skilled writer on technical stuff. Technical writers today should pay attention to them and learn.
Anyone here own a car? Maintain it? Use only a single screwdriver for everything - including hammering, removing bolts, replacing the oil filter....?
Different languages in the same basic family are designed for different, albeit similar, purposes. The similarities should not mislead you. Each will be good for the parameters for which they are designed, each will be bad for almost anything outside of those parameters. Smalltalk, Java and C++ belong to the object-based family (although Smalltalk is the only OO language of the three regarded as pure, Java is mostly OO and C++ is extremely hybrid). Because all are object-based, all are designed for problem-spaces concerned with containers, encapsulation, localizing and tightly-controlled visibility.
On the other hand, C and C++ also belong to a common family. In this case, both support C-style programming and syntax. This is useful when you want to use compilers with a lot of research into techniques behind them. C++ at its slowest was faster than any object-based or object-oriented programming language that required compilers that could learn little or nothing from previous generations of compiler theory. (Experimental languages that follow novel designs and syntax often fall into this category.)
Other examples of languages that fall into a common family and are obviously different: C-shell and BASH. Pascal and Occam. 80x86 and MIPS. Perl and Python. PHP and Javascript. Fortran and BASIC.
Unless you're looking for work today, what they want on Monster today is unimportant. If you are looking for work next week, then your resume needs to reflect the job requirements next week - today's requirements only apply for today. This is where risk analysis comes in. You can't know what's going to be important next week, next year, five years from now. No language lasts forever and it's fair to assume that eventually what is fashionable now will go the way of Cobol and Fortran - into very specialist and limited communities. But which languages will die off? And which languages will replace them?
BCPL evolved into B, which evolved into C. BCPL is no longer used, so technically died off, but C and C-related languages are not only in widespread use, the family is still growing. C#, D, Objective C... Anyone who learned BCPL (so long as they're not dead from old age) could understand the basics of D without ever touching a textbook on it, by virtue of it being a linear descendent. As such, it's less a die-off and more an evolution, which requires minimal training to be able to get right.
Erlang, Haskell, Forth, LISP, Ruby... These are not direct descendents and require additional understanding to program well in. The greater the difference in methodology, abstractness, style and notation, the greater the learning curve to become proficient in that language.
This is where it becomes important. You can't start learning the language the day you start a new job. You can't really start a month before, unless the learning curve only requires you put in a month. You have to determine how likely each language is to be needed and when by, and then make sure you learn sufficiently in advance to compete for those jobs once those positions do become available. Unless it's far enough in advance that you'll have retired, in which case you can probably disregard it except for hobbyist purposes.
I'd agree, except that compilers even 20K is size can do an awful lot if they're well-written. Now, 8K compilers... (They do exist, but I can't think of one that's any good. There must be some, though - the PDP-8 and PDP-11 weren't known for having masses of RAM.)
Trying something new is very very important, as is the need to be generalized. Specialists can do very well in their specific field, but if that field dies off (even temporarily, as happened with Java in the DotCom crash) then the person is unmarketable until the field revives or they learn new skills. Generalists can always find employment somewhere, even if they never reach the same giddy heights, the versatility and flexibility for good pay, versus high risk at high pay.
The DMCA cripples reverse-engineering and software analysis. There was another "unified copyright" bill around the same time that forbid the review, or publication of any review, of software without the permission of the company selling it. Never heard if that made it into law. If it did, then yes, Felton's review would be illegal. If it didn't, but the review process would necessarily violate the DMCA, then it is also illegal.
Is this stupid? Yes. Is this hostile to the interests of Americans? Yes. Was it voted for by a Congressman you have re-elected since? That I don't know, obviously, but it's something you may want to check on before this election. If you're not planning onvoting for whoever is currently elected, you might also want to find out the views of the opponents, particularly if the region is tech-savvy enough (or even tech-phobic enough) to be suspicious of voting machines. The candidate you're looking at might enjoy playing around with Sequoia's attitude problem as a (minor, to them) campaign issue.
I'll add to that by saying that the "good person cannot do bad" idea results in an action being "good" because of who it was done by and an identical action by someone else as "bad" for the same reason. In such cases, the action becomes good or bad because of who did it, not because of what it is.
There are other variants on the theme. There are those who start by assuming that they are "good" and therefore all who disagree with them or are in conflict with them are necessarily "bad", as they are doing "un-good" things. This probably covers just about any major figure in any cult and quite a lot of the criminally insane. Cultures based on personal honour tend to extend the good person/bad person idea to say that bad consequences of a "good" action must be because the person the consequence happened to was bad.
Images can't be filtered for content except by inspection. Computers cannot analyze pictures in a subjective way. For that matter, there are enough infamous examples of humans not being able to analyze pictures reliably to make any form of filtering unreliable.
Why is this important? Surely freedom is the more important part of the argument. No. Freedom is itself subjective. There is no objective definition of freedom, because that is dependent on what a given culture values. The freedom of a group is inherently in conflict with the freedom of an individual, so the more you value one, the less you value the other. If your society valued the freedom of the group, then to you, freedom means group freedom. Individuality reduces that group fredom and thus (to such a society) results in less freedom, not more. Now, it is also possible to have societies that value neither group freedom nor individual freedom. Some might value leadership freedom.
In all of these cases, the technological argument is untouched and unaltered. It is not subjective, it doesn't vary by how you define freedom, how you define an obscene image, or indeed how you define an image. It simply is. It doesn't depend on technological sophistication, money available, attitudes, social niceties, etc. It is consistant, universal, unalterable, inescapable.
THAT is why arguments over freedom make no sense - whose definition? Yours or your opponents? - and why technological arguments should always be used instead where it's not just a matter of time, but a matter of what technology can accomplish. Never, ever give credence to the impossible by implying fluffy non-real things are a stronger argument.
Oregon has a disasterous reputation when it comes to education. State schools are not doing well and the department of education at the state level has a reputation for being gridlocked. Whether or not this reflects events on the ground, it hurts Oregon to be seen this way. It hurts the migration of business in, it hurts those who are already in Oregon, and it depresses those entering or in the system. I've lived in Oregon for many years now and have yet to hear a single good word about the schools here. To the extent that it is true, it's far from necessary. Oregon has some of the brightest people in the world, has an amazing reputation for creativity and expression, and has all the natural, historic and modern resources a teacher could possibly want for any art, science or social science class you could imagine.
What would - or, indeed, could - you do to improve the reputation and quality of Oregon schools, to produce the kind of skilled yet generalized mind so badly needed?
Contextual Information: Oregon could - and should - be one of the leading States in America in education and should be regarded in much the same way as the Oxbridge or former JMB regions in England are - centres of excellence. The old JMB region (which covered five educational districts in the northwest of England) is probably the closer of the two - they are roughly from the same era, the industries past and present are roughly the same, and the resources available are comparable. There, the comparison seems to end. The Universities in Cheshire, Lancashire and Derbyshire are ranked amongst the best in the world. The schools are generally highly regarded and the state-run schools are considered comparable with the old grammar (ie: private) schools. Not equal, in terms of intake - the grammar schools are very select - but certainly comparable in terms of what they do with who they take. Even if the only real difference between them and Oregon's system was on of image and attitude, Oregon's system is broken and theirs isn't. The fix exists, apply it.
I've personally never seen the conflict myself. If you salvage twice as much, provided it is done efficiently, then you reduce the cost of mining and ore processing (both of which are expensive), which must lead to a net reduction in cost of materials, which in turn must lead to a saving for the consumer. Efficient environmentalism is expensive to start, but would (on the longer-term) work out cheaper.
The same is true on pollution controls. A lot of pollution is generated by the processing method itself, so improving the method must reduce the pollution. Since creating pollution consumes resources (materials and power) that the company has paid for, it would seem to follow that the less pollution a company creates, the cheaper it would be. Now, there's only so far you can go with this, and the research to get any further than what is presently done isn't cheap and will give declining returns, so filters and overheads are inevitable, but it does lead me to believe that environmentalism is cheaper than we're being led to believe.
(Since we can reasonably assume companies want to make money and therefore want to lower overheads, we can assume that a lot of the research and development necessary is beyond a lot of private industry at this time, and/or there's not enough incentive to handle the high initial costs. Maybe the role of government can be to fund some of the work, perhaps provide X-prize-like challenges, and once things are cost-effective to deploy, offer tax cuts on business so that it's practical to switch methods and add an inefficiency tax to make it impractical not to.)
An accurate method won't give you a value to test against, it'll give you a set of values for each pixel, where the value of that pixel contributes towards the values in the remaining sets for all the other pixels. You do not get a yes/no answer. The best you can do is a maybe/no. Simply put, let's just take the case of multi-generational diffuse reflections. Each object in the scene contributes something to the lighting of every other object, directly or indirectly. By adding or removing even a single object, the diffuse reflections for all pixels become wrong. You then have to calculate the inverse of a one-way function for each pixel, and then calculate the intersection of all the results of that, and then regenerate the scene, which gives you a really bad exponential problem. Whereas, if the photograph is complete, if you take enough pixels to develop an idea of the system, you can constrain the number of possible values of the remaining pixels sufficiently to be able to test if any given pixel is in or out of that set in very short order. It won't prove the image authentic, but it will be able to prove it inauthentic, if it is, on many occasions. How many depends on how constrained the system is. It can never be constrained to 1 possible solution, but you can certainly constrain it enough to have a high level of confidence in an image being genuine if no fakery is detected.
This is why I think it would be better to have virtual cores and physical hyperthreading. You have as many compute elements as possible, all of which are available to all virtual cores. The number of virtual cores presented could be set equal to the number of threads available, equal to the number of register sets the processor could describe in internal memory, or to some number decided by some other aspect of the design. Each core would see all compute elements, and would use them as needed for out-of-order operatons. The primary idea would be to hide the multithreading of the chip from the OS, yet take advantage of being able to multithread. In addition to that, however, if one core can't exploit multiple threads but another core can exploit many, then you don't waste compute elements or slow things down by not making resources available.
(Since a compute element is designed for one specific task, you end up with a maximum number of supportable virtual cores equal to the number of pools times the number of elements in each pool. The minimum number of cores would be determined by the maximum number of threads generated by any instruction supported. If the CPU was really smart, it could "hotplug" CPUs to increase and reduce the number of cores that appear to the operating system, so that if there's a heavy, sustained use of the threading, the CPU doesn't try to overcommit resources.)
Well, not if you wanted any kind of performance. A very crude model of electronics is to think of voltage as the difference in the amount of charge between two points and the current as the rate at which units of charge are being shovelled through the system. Think of it as lumps of coal and the size of shovel being used. Ten times the voltage with one tenth the current would mean it would take a hundred times as long to do the shovelling. Smaller voltages work with smaller currents, if the two are comparable, because the reductions cancel out.
This method, incidentally, has been five years away for several years. I think that other methods of reducing power output on chips (Gallium Arsonide, asychronous solid-state circuits, reduced resistivity in the interconnects) and on computers overall (better caching on drives, smarter offloading, etc) will prove to be the way to go.
This would presumably work best with HDR images, such as JPEG2000, OpenEXR and the like, as very few image capturing devices could be reliable at 48bpp. Actually, the "ideal" might well be to calculate the degree of randomness in the low-order bits of the image and then pick an encryption algorithm whose typical degree of apparent randomness was closest to the degree of randomness in the image. There may be ways of using compression prior to encryption, or data recovery bits + deliberate corruption to further massage the encrypted data so that statistical treatment would find no abnormalities.
Detecting discrepencies between light sources is something that can be written as an algorithm and should increase in complexity linearly with the product of the number of light sources and number of pixels. Determining the correct value for pixels is something that must be written as a herustic and will increase in complexity exponentially with the number of objects (including light sources and non-visible objects) in the system. A herustic is not guaranteed to finish. Ever. Even if it always did, because the problem is exponential, far more images can be checked than can be fixed up.
The disasterous city-to-city races (in which driver and spectator fatalities tended to be high) showed what safety these cars had at any speed. In the Brighton historic run (cars later than 1906 not permitted), roughly 2/3rds of the cars finish a very gruelling race, which shows that reliability was respectable for the day but not at the level of modern cars. Borrowing ideas from the engine seems reasonable enough, but it would need to be heavily modernised.
I think it would be sufficient to cross Wolverine with Rogue. No further genetic engineering should be required after that.
Seriously, horses are a Really Bad Idea for general transportation. The reason they were abandoned for that purpose has far less to do with speed and far more to do with cost in money and cost in time to maintain horses, which is essentially what you said. Also, you need different breeds of horses for different types of work. You wouldn't use a shire horse for rapid transit, a dartmoor pony for heavy loads, or a modern racehorse for farming. That makes any kind of genetic engineering hard, as you now have to solve the problem potentially once for every breed, depending on the genetic distance between them.
Well, for smallpox, those milking cows or drinking unpasteurised milk would have been exposed to cowpox, which is similar enough that it provides a level of immunity. Indeed, it is by observing this that vaccinations were invented. However, the entire world isn't descended from milkmaids or lactose-tolerent individuals. Rather, smallpox is a survivable disease. Not all who catch it die, although survival rates aren't high. It is also a relatively fast-acting disease. Couple that with primitive isolation techniques, pockets of the disease will rapidly burn out. All told, there was enough low-level exposure to confer limited resistance to the general population, which is why smallpox - when taken to isolated comunities around the world - was so devastating.
The other diseases have similar backgrounds. Lots of low-level exposure, a non-zero survival rate, similar diseases which could assist in resistance being built up, etc. HIV belongs to a category of disease for which none of these apply. Even deactivating the virus has proved useless - it somehow manages to reactivate itself.
There is no resistance to it. Not a single person has survived exposure to the virus. The few supposed exceptions turned out not to be. The body cannot adjust to it. HIV is a polymorphic virus that mutates almost every replication. There is no evolutionary pressure to be resistant to it, because there is no survival rate. Same as there's no build-up in antibiotic-resistant bacteria when medication is taken correctly and appropriately. Resistant people in Africa or anywhere else is a nice fiction but should be left in Neuromancer.
Can tell you what to be afraid of, they can only tell you what they are afraid of. Personally, I favour the approach of knowing more about a subject before dismissing it, and I regard this new problem as potentially very significant. However, there is a huge difference between what is possible and what actually happens. H5N1 is a possible threat (and can therefore be prevented before it becomes actual). AIDS is an actual disaster and, no matter what is ever achieved, the best that can be done is preventing that disaster from becoming worse. You can't prevent it from being a tragedy at this point. Global warming - well, it depends on who you talk to. James Lovelock - one of the world's premiere environmental scientists - thinks prevention is now impossible. Others, just as notable, think it is. World wheat collapses? A lot of land has been cleared for beef (and, these days, corn) around the world, well outside the affected areas, so I'm inclined to think that action is still possible......but only if it's taken seriously as a possibility. It's when nobody cares that things become a disaster.
As for the over-medicated culture we live in, that's still about not caring. If people cared, they wouldn't avoid. If they didn't avoid, they wouldn't need over-medicating. Avoiding by apathy or by drugs is the same thing.
...if there's a viable alternative. If changing ISP isn't possible or practical, it will modify behaviours of neither users nor the ISPs. It'll simply result in unstable service and unstable customer bases, helping no-one but (because it's distributed across all ISPs) hurting no-one who can make any difference.
Slashdot Mirror
In terms of his factual writings, I have many of his articles that were written for Wireless World, including the letter and two follow-up articles on geostationary satellites. Those three in particular can be found on the web - many people have scanned them in. They're well worth reading. He was a highly skilled writer on technical stuff. Technical writers today should pay attention to them and learn.
Different languages in the same basic family are designed for different, albeit similar, purposes. The similarities should not mislead you. Each will be good for the parameters for which they are designed, each will be bad for almost anything outside of those parameters. Smalltalk, Java and C++ belong to the object-based family (although Smalltalk is the only OO language of the three regarded as pure, Java is mostly OO and C++ is extremely hybrid). Because all are object-based, all are designed for problem-spaces concerned with containers, encapsulation, localizing and tightly-controlled visibility.
On the other hand, C and C++ also belong to a common family. In this case, both support C-style programming and syntax. This is useful when you want to use compilers with a lot of research into techniques behind them. C++ at its slowest was faster than any object-based or object-oriented programming language that required compilers that could learn little or nothing from previous generations of compiler theory. (Experimental languages that follow novel designs and syntax often fall into this category.)
Other examples of languages that fall into a common family and are obviously different: C-shell and BASH. Pascal and Occam. 80x86 and MIPS. Perl and Python. PHP and Javascript. Fortran and BASIC.
BCPL evolved into B, which evolved into C. BCPL is no longer used, so technically died off, but C and C-related languages are not only in widespread use, the family is still growing. C#, D, Objective C... Anyone who learned BCPL (so long as they're not dead from old age) could understand the basics of D without ever touching a textbook on it, by virtue of it being a linear descendent. As such, it's less a die-off and more an evolution, which requires minimal training to be able to get right.
Erlang, Haskell, Forth, LISP, Ruby... These are not direct descendents and require additional understanding to program well in. The greater the difference in methodology, abstractness, style and notation, the greater the learning curve to become proficient in that language.
This is where it becomes important. You can't start learning the language the day you start a new job. You can't really start a month before, unless the learning curve only requires you put in a month. You have to determine how likely each language is to be needed and when by, and then make sure you learn sufficiently in advance to compete for those jobs once those positions do become available. Unless it's far enough in advance that you'll have retired, in which case you can probably disregard it except for hobbyist purposes.
Trying something new is very very important, as is the need to be generalized. Specialists can do very well in their specific field, but if that field dies off (even temporarily, as happened with Java in the DotCom crash) then the person is unmarketable until the field revives or they learn new skills. Generalists can always find employment somewhere, even if they never reach the same giddy heights, the versatility and flexibility for good pay, versus high risk at high pay.
Is this stupid? Yes. Is this hostile to the interests of Americans? Yes. Was it voted for by a Congressman you have re-elected since? That I don't know, obviously, but it's something you may want to check on before this election. If you're not planning onvoting for whoever is currently elected, you might also want to find out the views of the opponents, particularly if the region is tech-savvy enough (or even tech-phobic enough) to be suspicious of voting machines. The candidate you're looking at might enjoy playing around with Sequoia's attitude problem as a (minor, to them) campaign issue.
There are other variants on the theme. There are those who start by assuming that they are "good" and therefore all who disagree with them or are in conflict with them are necessarily "bad", as they are doing "un-good" things. This probably covers just about any major figure in any cult and quite a lot of the criminally insane. Cultures based on personal honour tend to extend the good person/bad person idea to say that bad consequences of a "good" action must be because the person the consequence happened to was bad.
Why is this important? Surely freedom is the more important part of the argument. No. Freedom is itself subjective. There is no objective definition of freedom, because that is dependent on what a given culture values. The freedom of a group is inherently in conflict with the freedom of an individual, so the more you value one, the less you value the other. If your society valued the freedom of the group, then to you, freedom means group freedom. Individuality reduces that group fredom and thus (to such a society) results in less freedom, not more. Now, it is also possible to have societies that value neither group freedom nor individual freedom. Some might value leadership freedom.
In all of these cases, the technological argument is untouched and unaltered. It is not subjective, it doesn't vary by how you define freedom, how you define an obscene image, or indeed how you define an image. It simply is. It doesn't depend on technological sophistication, money available, attitudes, social niceties, etc. It is consistant, universal, unalterable, inescapable.
THAT is why arguments over freedom make no sense - whose definition? Yours or your opponents? - and why technological arguments should always be used instead where it's not just a matter of time, but a matter of what technology can accomplish. Never, ever give credence to the impossible by implying fluffy non-real things are a stronger argument.
What would - or, indeed, could - you do to improve the reputation and quality of Oregon schools, to produce the kind of skilled yet generalized mind so badly needed?
Contextual Information: Oregon could - and should - be one of the leading States in America in education and should be regarded in much the same way as the Oxbridge or former JMB regions in England are - centres of excellence. The old JMB region (which covered five educational districts in the northwest of England) is probably the closer of the two - they are roughly from the same era, the industries past and present are roughly the same, and the resources available are comparable. There, the comparison seems to end. The Universities in Cheshire, Lancashire and Derbyshire are ranked amongst the best in the world. The schools are generally highly regarded and the state-run schools are considered comparable with the old grammar (ie: private) schools. Not equal, in terms of intake - the grammar schools are very select - but certainly comparable in terms of what they do with who they take. Even if the only real difference between them and Oregon's system was on of image and attitude, Oregon's system is broken and theirs isn't. The fix exists, apply it.
The same is true on pollution controls. A lot of pollution is generated by the processing method itself, so improving the method must reduce the pollution. Since creating pollution consumes resources (materials and power) that the company has paid for, it would seem to follow that the less pollution a company creates, the cheaper it would be. Now, there's only so far you can go with this, and the research to get any further than what is presently done isn't cheap and will give declining returns, so filters and overheads are inevitable, but it does lead me to believe that environmentalism is cheaper than we're being led to believe.
(Since we can reasonably assume companies want to make money and therefore want to lower overheads, we can assume that a lot of the research and development necessary is beyond a lot of private industry at this time, and/or there's not enough incentive to handle the high initial costs. Maybe the role of government can be to fund some of the work, perhaps provide X-prize-like challenges, and once things are cost-effective to deploy, offer tax cuts on business so that it's practical to switch methods and add an inefficiency tax to make it impractical not to.)
Heh! The first viral Elite-player-detector Sig functions wonderfully... :)
An accurate method won't give you a value to test against, it'll give you a set of values for each pixel, where the value of that pixel contributes towards the values in the remaining sets for all the other pixels. You do not get a yes/no answer. The best you can do is a maybe/no. Simply put, let's just take the case of multi-generational diffuse reflections. Each object in the scene contributes something to the lighting of every other object, directly or indirectly. By adding or removing even a single object, the diffuse reflections for all pixels become wrong. You then have to calculate the inverse of a one-way function for each pixel, and then calculate the intersection of all the results of that, and then regenerate the scene, which gives you a really bad exponential problem. Whereas, if the photograph is complete, if you take enough pixels to develop an idea of the system, you can constrain the number of possible values of the remaining pixels sufficiently to be able to test if any given pixel is in or out of that set in very short order. It won't prove the image authentic, but it will be able to prove it inauthentic, if it is, on many occasions. How many depends on how constrained the system is. It can never be constrained to 1 possible solution, but you can certainly constrain it enough to have a high level of confidence in an image being genuine if no fakery is detected.
There is no NSA. The NSA does not exist. They are not run by CmdrTaco as they do not exist to be run. There are no Macra! (Sorry, wrong series.)
(Since a compute element is designed for one specific task, you end up with a maximum number of supportable virtual cores equal to the number of pools times the number of elements in each pool. The minimum number of cores would be determined by the maximum number of threads generated by any instruction supported. If the CPU was really smart, it could "hotplug" CPUs to increase and reduce the number of cores that appear to the operating system, so that if there's a heavy, sustained use of the threading, the CPU doesn't try to overcommit resources.)
This method, incidentally, has been five years away for several years. I think that other methods of reducing power output on chips (Gallium Arsonide, asychronous solid-state circuits, reduced resistivity in the interconnects) and on computers overall (better caching on drives, smarter offloading, etc) will prove to be the way to go.
There are twice as many jokes of that kind every 18 months.
This would presumably work best with HDR images, such as JPEG2000, OpenEXR and the like, as very few image capturing devices could be reliable at 48bpp. Actually, the "ideal" might well be to calculate the degree of randomness in the low-order bits of the image and then pick an encryption algorithm whose typical degree of apparent randomness was closest to the degree of randomness in the image. There may be ways of using compression prior to encryption, or data recovery bits + deliberate corruption to further massage the encrypted data so that statistical treatment would find no abnormalities.
Detecting discrepencies between light sources is something that can be written as an algorithm and should increase in complexity linearly with the product of the number of light sources and number of pixels. Determining the correct value for pixels is something that must be written as a herustic and will increase in complexity exponentially with the number of objects (including light sources and non-visible objects) in the system. A herustic is not guaranteed to finish. Ever. Even if it always did, because the problem is exponential, far more images can be checked than can be fixed up.
The disasterous city-to-city races (in which driver and spectator fatalities tended to be high) showed what safety these cars had at any speed. In the Brighton historic run (cars later than 1906 not permitted), roughly 2/3rds of the cars finish a very gruelling race, which shows that reliability was respectable for the day but not at the level of modern cars. Borrowing ideas from the engine seems reasonable enough, but it would need to be heavily modernised.
Seriously, horses are a Really Bad Idea for general transportation. The reason they were abandoned for that purpose has far less to do with speed and far more to do with cost in money and cost in time to maintain horses, which is essentially what you said. Also, you need different breeds of horses for different types of work. You wouldn't use a shire horse for rapid transit, a dartmoor pony for heavy loads, or a modern racehorse for farming. That makes any kind of genetic engineering hard, as you now have to solve the problem potentially once for every breed, depending on the genetic distance between them.
The other diseases have similar backgrounds. Lots of low-level exposure, a non-zero survival rate, similar diseases which could assist in resistance being built up, etc. HIV belongs to a category of disease for which none of these apply. Even deactivating the virus has proved useless - it somehow manages to reactivate itself.
There is no resistance to it. Not a single person has survived exposure to the virus. The few supposed exceptions turned out not to be. The body cannot adjust to it. HIV is a polymorphic virus that mutates almost every replication. There is no evolutionary pressure to be resistant to it, because there is no survival rate. Same as there's no build-up in antibiotic-resistant bacteria when medication is taken correctly and appropriately. Resistant people in Africa or anywhere else is a nice fiction but should be left in Neuromancer.
As for the over-medicated culture we live in, that's still about not caring. If people cared, they wouldn't avoid. If they didn't avoid, they wouldn't need over-medicating. Avoiding by apathy or by drugs is the same thing.
...in the BBC terror-drama "Edge of Darkness", although a little artistic license was taken.
...if there's a viable alternative. If changing ISP isn't possible or practical, it will modify behaviours of neither users nor the ISPs. It'll simply result in unstable service and unstable customer bases, helping no-one but (because it's distributed across all ISPs) hurting no-one who can make any difference.
Only Cowboy Neil's alive. He's imagining the rest of us.