You can and we do prevent medical errors. Look up any of the research on surgical checklists for a very specific example of preventing medical errors, but there are many others. We may not be able to prevent ALL medical errors, but we also can't prevent ALL mass shootings. There's room for improvement on both fronts.
Sadly he might just be thinking of antibiotic laced soaps. Not sure those have caused any resistance themselves, but they're shown to be no more effective that regular soap and we still made obscene amounts of the stuff. He's wrong but not quite crazy as you imply.
I think we computer people over simplify the task of taking a digital genome and remaking the original cell, but it should be simpler with a virus than just about anything else. I suspect we're safe from random weirdos recreating smallpox in a home lab for quite a while but I agree that the time is coming. As for your vaccination, high immune response lasts 3-5 years with decreasing immunity after that so your probably about as screwed as the rest of us.
One missed person is not enough for two big reasons. One, we have a concept of quarantine and how diseases spread, but more critically in this case the vaccine is a different virus. We can produce that virus, vaccinia, in whatever quantity we want without any access to smallpox. People use it for lab research now without a ton of precautions. Even unattenuated vaccinia only calls for biosafety level 2, I.e. the same level as e coli or various hep strains.
I'm feeding a troll here, but you know that the earliest "vaccine" for smallpox was.... smallpox. The fatality rate for that method, called variolation, was on the order of 1-2% and there are records of intentional smallpox infection to induce immunity going back to at least 1500. The first real vaccine was infection by cowpox and was introduced in 1796. I imagine your dates are a bit cherry picked, but the Jenner vaccine wouldn't have reached places like India and China immediately so I suspect they still used variolation at least for a while. A 2% death rate if enough people were "vaccinated" that way would drive up the death toll for that crappy vaccine but is still better than something like 35% death rates for natural infections.
Different viruses have different mutation rates. The major cause of that difference between influenza and smallpox is that one uses DNA and one uses RNA. RNA viruses are very sloppy copiers while DNA viruses like smallpox are much more consistent. If you look at the history of smallpox vaccination it mutates so slowly that over centuries the same vaccination methods, like accidental exposure to cow pox, are still effective.
That doesn't mean that there's no utility in keeping samples, but there's so little use in keeping them for vaccine development. If the strain mutated to the point that the old vaccine was useless you'd want samples of the new strain to develop the new vaccine. The old strain might be useful as a comparison, but probably for its genetic sequence that we have digitally stored anyway.
I don't know that they're fundamentally different on wanting to bypass encryption, but I do think that people who appear to believe:
the Government is a bunch of bumbling idiots where the free market can outperform it hands down, yet they also expect it able to perform things such as conspiracies where thousands of people are involved
tend to be Republicans. Actually they may identify as Libertarians, but probably not as Democrats who traditionally want to expand government reach to help some societal group. Not saying they even CAN help, but that's the reasoning. Of course, the politics surrounding fear of extremists tends to blur the lines quite hard. I think I'm in more danger of drowning in a bathtub that dying in a terrorist action but essentially no one worries about the death trap lurking in every bathroom.
I didn't see anything where it said the maximum salary would be $70k, just that the CEO was dropping his own to $70k. I've known a few businesses where a particularly valuable employee made more than the CEO/owner since the CEO/owner made money off stock or was already wealthy enough that the salary just didn't mean much to their overall numbers.
I'm with you on your points, but can you cite when Phil Zimmerman was arrested? He was investigated back in the 90s and it was dropped without an indictment in 96 according to wikipedia.
Ok, replying to myself because I realized that if you wanted to be even more specific you should probably know who has the disposable income AND how much time they spend on the highway. It sounds like this thing isn't made for stop-and-go traffic, although the summary could be misleading me (because why would I RTFA?). I would expect the buyers to be early adopter types who want the newest gadget, spend significant time in a highway environment, and have $10k to throw at the problem. That's seems like a profitable, but not terribly plentiful group especially if the buyers consider the liability questions at this point.
The specificity is odd, but I think you need to take the disposable income of the people who own the cars into account as well. A higher percentage of Audi drivers than Corolla drivers will shell out for this system. Depending on the difference in percentage it might still make sense to pick the Corolla but the math isn't quite as simple as car counting.
God this sounds familiar..... and that's because I wrote a PhD thesis about building a system to do something a lot like this. It involved a fairly mediocre web interface wrapping a database of trust relationships specified by end users. A trusts B for 0.7 and B trusts C for 0.6 then you can put together a trust level between A and C by multiplying those together with some user-tweakable distance dropoff. Those trust levels were then measured against the levels required for access to shared data. Maybe you would allow anyone with a 0.7 or higher to read a given document and a 0.9 or higher to contribute to it. It was an interesting idea, but man did I get tired of it by the end. If for some bizarre reason anyone wants to read bits of it google books has some indexed and I probably have a pdf laying around somewhere....
I figured it could be quite useful, but I was so fed up with the work in mid-2007 that I never looked back at it.
Lasers are light, not magic. If you block the laser beam you block its frequency (color) of light. If you block all the possible laser frequencies I'm pretty sure the cockpit isn't going to be nearly as transparent, which is kind of the point of the cockpit. Raising the awareness that this is stupid and dangerous is a perfectly reasonable way to address the problem. A conviction shouldn't ruin someone's life forever but it sure as hell should make them and anyone that hears about it not want to do this.
I completely agree that making a driverless car better than all drivers is near impossible, but making it better than the person on the cell phone who just pulled out of the McDonald's drive-through isn't a high a bar. I wish I could remember where I saw the statistic, but a large percentage of accidents are caused by a smallish percentage of drivers. Lets just get better than them and we'll be better off.
As for your examples, keep in mind that the reaction time of the computer vs you can make up for a lot of missing information. The computer may not be able to read the body language of the people on the corner, but it can react in a fraction of the time you would take and track more targets. I watched a friend hit a deer while he was watching the a different deer to make sure it didn't bound into traffic. Something with more independent "eyes" wouldn't have faced the same problem.
As a programmer I spoke from that perspective but code is certainly not all there is in many projects.
Geeks do tend to be too cute about names and I find that to be asinine, but that's up to the project owner. Most of the time the open source project I may pull in is not user-visible so the name is a lot less important.
As for the installing/configuring, I've hit good and bad on both the open and non-open sides. The open source side tends toward the hobbyist so it may be worse in this respect for a randomly selected project. For many of projects on Linux the package maintainer, whether it be for apt or yum or some other system, often takes care of a lot of the that particular ugliness. I have a far simpler time installing a working web server with [insert backend language here] and [sql backend] using open source tools on Linux than I do on Windows. In many cases its little more than one command "apt-get install X Y X" where it goes out and finds everything, confirms I really want this, and does all the work for me.
I'm not convinced that commercial apps I've used are not just as graphically ugly as many of the open source ones, but then again most of the projects I deal with are libraries so I have less emphasis on the graphical elements. From the big companies, like MS, Apple, or Google, yes the graphical polish will be nicer, but at least many open source projects can get someone to make english sentences in user-visible dialogs unlike my "Smart TV" manufacturer (panasonic).
As for documentation, in the open source world it sometimes does not exist and that is a *major* issue for those projects. In the close source world it make exist more often but can also be completely unusable. I'm not sure which is worse in the end. With one I know from the get-go that I'm on my own or that I have to use community resources. With the other I may struggle with it for a long time before I figure out that I should just ignore the docs and look at what the damn library is actually doing.... but I usually can't.
When it comes to testing I've had poorly tested from each side and no chance of fixing one but some change of fixing the other. There's shit on both sides of the fence.
I'd like to come back to my original point that financial incentives are not the only incentives. I spoke of code because I write code, not because it is the only important thing. It's the thing I can speak of from personal experience. I would argue that other types of work that are done because someone wants to impress others, wants to scratch their own itch, or just wants to help a good cause are often done better than tasks done strictly for pay. Hopefully we all have jobs doing what we love, but when that isn't the case there's a great chance that you do something better in your off time because you really give a shit about it.
Was there a good reason to structure your post as a series of snide questions and a flippant remark?
There are many other types of incentives and I have rarely done my best work for strictly financial ones. When contributing to an open source project you have to think that somewhere someone will look at the code you write and have the ability to publicly shame you if you do something truly stupid. Standing, respect, whatever you want to call it, is a big motivator for many people. If the same thing happens in many businesses there *may* be consequences, but often as long as it works well enough to collect the customer's money it ships. Personally, I've found more fugly code turds in various closed source projects than I've touched than in the open source world.
I think you're misreading or oversimplifying those situations. It's not that there was no demand, but it is more difficult to measure before the product exists. There was a market for smartphones and Apple 1) tends to put out good stuff, 2) already made iPods so they had some experience w making small portable consumer devices, 3) demonstrated that people loved the Apple branding of such things.
As for the Tesla, people plunked down $40000 reservations before the cars existed, and continue to do so for new models. If that doesn't show demand exists I'm not sure what would.
I'm sure there are products where there was no demand before the product existed, but I don't think you can call that the common case. Before a company sinks significant cost into designing a new product they generally run some numbers to measure the current market demand for the final product. Sure, some of that is wishful thinking, but some is evaluating current offerings and seeing how your product will be able to grab the customers. I'm sure Apple looked at BlackBerry devices before throwing their hat into the smartphone market and I'm sure Musk had more that just a hunch that rich people would buy a luxury electric car. However, when you have enough money, as did the creators of both these examples, the market analysis is a little less critical during the development stage just because a failed project isn't the end of the world. Apple could have kept going even if the iPhone flopped and Musk wouldn't run out of cash if Tesla crashed and burned during the roadster days.
I bet someone in battery manufacturing is looking as adding capacity now in anticipation of such events. This could be quite an opportunity for some manufacturer with a bit of foresight. As more companies make and sell more electric cars I doubt Tesla will be the only company hunting for more, cheaper, better.
The parasitic worm is Trichinella spiralis and is actually relatively easy to kill with cooking. Many pork dishes involve long slow cooking which renders them worms completely harmless. We eat feral pigs rural northern Florida without a problem. Granted, I pick preparations like slow cooker Cuban pork for any wild meat to be absolutely it's safe. As for exposure for the butcher, that's crap. The worm is embedded in a cyst that is dissolved by stomach acid so don't lick your hand while cleaning the animal but that's pretty much it for trichinosis. Of course it is an animal so there's always bacteria as well but you get different but other dangerous strains in our commercial meat supply too.
Considering the question of pilot reliance on automation, and the vast canyon of difference between the training they receive and that of a typical automobile operator, I fear this particular solution (self-driving cars) will only compound an existing problem.
I can see this being a problem if automation is introduced piecemeal, but if you go from what we have now to something where the human driver is not required your example tends to make the opposite point for me. Before you say that cutting the human out is insane keep in mind that when something goes wrong in a car the stakes are not nearly as high as in a plane, both for the number of people involved, and for the fact that if the car engine cuts out and it coasts to a stop the result is generally not a pile of flaming wreckage in a hillside.
Keep in mind that a safer self-driving car only has to be better than the current batch of drivers and that is not a terribly high bar. Given how little training the average driver is required to have and thereby how badly many (most?) people drive doesn't that make the problem of creating a self-driving car that improves safety on the roads even simpler? I think you're trying to say that humans will be even worse drivers when they usually rely on automated technology, but there's not a ton of room at the bottom now. Based on most of the drivers I've seen, once you clear the initial figuring-it-out phase of your teenage years there doesn't seem to be a huge correlation between how long a person has been driving and how safe a driver they are. Other factors including fatigue, anger, and distraction appear to play a *much* larger role in their abilities. Keep in mind that the computer suffers from none of those problems AND can't get drunk. Shoot, just removing human competitive and aggressive instincts from the road would make driving much safer even if other mistakes remained constant.
Insurance companies may be terrible at everything else, but they can run statistics and if there is a self-driving car that gets in only 10% of the accidents that a human driver does there will be price incentives to push people toward that. Of course the incentives will only reduce your insurance by 20%, thereby making the customer happy and giving the insurance company a tidy profit since they don't pay out nearly as much.
Slashdot Mirror
You can and we do prevent medical errors. Look up any of the research on surgical checklists for a very specific example of preventing medical errors, but there are many others. We may not be able to prevent ALL medical errors, but we also can't prevent ALL mass shootings. There's room for improvement on both fronts.
Wish I had mod points
Sadly he might just be thinking of antibiotic laced soaps. Not sure those have caused any resistance themselves, but they're shown to be no more effective that regular soap and we still made obscene amounts of the stuff. He's wrong but not quite crazy as you imply.
I think we computer people over simplify the task of taking a digital genome and remaking the original cell, but it should be simpler with a virus than just about anything else. I suspect we're safe from random weirdos recreating smallpox in a home lab for quite a while but I agree that the time is coming. As for your vaccination, high immune response lasts 3-5 years with decreasing immunity after that so your probably about as screwed as the rest of us.
One missed person is not enough for two big reasons. One, we have a concept of quarantine and how diseases spread, but more critically in this case the vaccine is a different virus. We can produce that virus, vaccinia, in whatever quantity we want without any access to smallpox. People use it for lab research now without a ton of precautions. Even unattenuated vaccinia only calls for biosafety level 2, I.e. the same level as e coli or various hep strains.
Cute fur
I'm feeding a troll here, but you know that the earliest "vaccine" for smallpox was.... smallpox. The fatality rate for that method, called variolation, was on the order of 1-2% and there are records of intentional smallpox infection to induce immunity going back to at least 1500. The first real vaccine was infection by cowpox and was introduced in 1796. I imagine your dates are a bit cherry picked, but the Jenner vaccine wouldn't have reached places like India and China immediately so I suspect they still used variolation at least for a while. A 2% death rate if enough people were "vaccinated" that way would drive up the death toll for that crappy vaccine but is still better than something like 35% death rates for natural infections.
Different viruses have different mutation rates. The major cause of that difference between influenza and smallpox is that one uses DNA and one uses RNA. RNA viruses are very sloppy copiers while DNA viruses like smallpox are much more consistent. If you look at the history of smallpox vaccination it mutates so slowly that over centuries the same vaccination methods, like accidental exposure to cow pox, are still effective.
That doesn't mean that there's no utility in keeping samples, but there's so little use in keeping them for vaccine development. If the strain mutated to the point that the old vaccine was useless you'd want samples of the new strain to develop the new vaccine. The old strain might be useful as a comparison, but probably for its genetic sequence that we have digitally stored anyway.
I don't know that they're fundamentally different on wanting to bypass encryption, but I do think that people who appear to believe:
the Government is a bunch of bumbling idiots where the free market can outperform it hands down, yet they also expect it able to perform things such as conspiracies where thousands of people are involved
tend to be Republicans. Actually they may identify as Libertarians, but probably not as Democrats who traditionally want to expand government reach to help some societal group. Not saying they even CAN help, but that's the reasoning. Of course, the politics surrounding fear of extremists tends to blur the lines quite hard. I think I'm in more danger of drowning in a bathtub that dying in a terrorist action but essentially no one worries about the death trap lurking in every bathroom.
I didn't see anything where it said the maximum salary would be $70k, just that the CEO was dropping his own to $70k. I've known a few businesses where a particularly valuable employee made more than the CEO/owner since the CEO/owner made money off stock or was already wealthy enough that the salary just didn't mean much to their overall numbers.
Remember it doesn't have to be perfect, it just has to be better than the bottom 50% of drivers.
That's great for the bottom 50% of drivers. No so great for those of us who've been driving for decades and never caused an accident.
Why would I want to travel in a self-driving car that drives worse than I do?
I and 80% of the other drivers on the road agree completely.
Damn, there's a hell of a good short story in there....
I'm with you on your points, but can you cite when Phil Zimmerman was arrested? He was investigated back in the 90s and it was dropped without an indictment in 96 according to wikipedia.
http://xkcd.com/605/
Ok, replying to myself because I realized that if you wanted to be even more specific you should probably know who has the disposable income AND how much time they spend on the highway. It sounds like this thing isn't made for stop-and-go traffic, although the summary could be misleading me (because why would I RTFA?). I would expect the buyers to be early adopter types who want the newest gadget, spend significant time in a highway environment, and have $10k to throw at the problem. That's seems like a profitable, but not terribly plentiful group especially if the buyers consider the liability questions at this point.
The specificity is odd, but I think you need to take the disposable income of the people who own the cars into account as well. A higher percentage of Audi drivers than Corolla drivers will shell out for this system. Depending on the difference in percentage it might still make sense to pick the Corolla but the math isn't quite as simple as car counting.
God this sounds familiar..... and that's because I wrote a PhD thesis about building a system to do something a lot like this. It involved a fairly mediocre web interface wrapping a database of trust relationships specified by end users. A trusts B for 0.7 and B trusts C for 0.6 then you can put together a trust level between A and C by multiplying those together with some user-tweakable distance dropoff. Those trust levels were then measured against the levels required for access to shared data. Maybe you would allow anyone with a 0.7 or higher to read a given document and a 0.9 or higher to contribute to it. It was an interesting idea, but man did I get tired of it by the end. If for some bizarre reason anyone wants to read bits of it google books has some indexed and I probably have a pdf laying around somewhere....
I figured it could be quite useful, but I was so fed up with the work in mid-2007 that I never looked back at it.
Lasers are light, not magic. If you block the laser beam you block its frequency (color) of light. If you block all the possible laser frequencies I'm pretty sure the cockpit isn't going to be nearly as transparent, which is kind of the point of the cockpit. Raising the awareness that this is stupid and dangerous is a perfectly reasonable way to address the problem. A conviction shouldn't ruin someone's life forever but it sure as hell should make them and anyone that hears about it not want to do this.
I completely agree that making a driverless car better than all drivers is near impossible, but making it better than the person on the cell phone who just pulled out of the McDonald's drive-through isn't a high a bar. I wish I could remember where I saw the statistic, but a large percentage of accidents are caused by a smallish percentage of drivers. Lets just get better than them and we'll be better off.
As for your examples, keep in mind that the reaction time of the computer vs you can make up for a lot of missing information. The computer may not be able to read the body language of the people on the corner, but it can react in a fraction of the time you would take and track more targets. I watched a friend hit a deer while he was watching the a different deer to make sure it didn't bound into traffic. Something with more independent "eyes" wouldn't have faced the same problem.
As a programmer I spoke from that perspective but code is certainly not all there is in many projects.
Geeks do tend to be too cute about names and I find that to be asinine, but that's up to the project owner. Most of the time the open source project I may pull in is not user-visible so the name is a lot less important.
As for the installing/configuring, I've hit good and bad on both the open and non-open sides. The open source side tends toward the hobbyist so it may be worse in this respect for a randomly selected project. For many of projects on Linux the package maintainer, whether it be for apt or yum or some other system, often takes care of a lot of the that particular ugliness. I have a far simpler time installing a working web server with [insert backend language here] and [sql backend] using open source tools on Linux than I do on Windows. In many cases its little more than one command "apt-get install X Y X" where it goes out and finds everything, confirms I really want this, and does all the work for me.
I'm not convinced that commercial apps I've used are not just as graphically ugly as many of the open source ones, but then again most of the projects I deal with are libraries so I have less emphasis on the graphical elements. From the big companies, like MS, Apple, or Google, yes the graphical polish will be nicer, but at least many open source projects can get someone to make english sentences in user-visible dialogs unlike my "Smart TV" manufacturer (panasonic).
As for documentation, in the open source world it sometimes does not exist and that is a *major* issue for those projects. In the close source world it make exist more often but can also be completely unusable. I'm not sure which is worse in the end. With one I know from the get-go that I'm on my own or that I have to use community resources. With the other I may struggle with it for a long time before I figure out that I should just ignore the docs and look at what the damn library is actually doing.... but I usually can't.
When it comes to testing I've had poorly tested from each side and no chance of fixing one but some change of fixing the other. There's shit on both sides of the fence.
I'd like to come back to my original point that financial incentives are not the only incentives. I spoke of code because I write code, not because it is the only important thing. It's the thing I can speak of from personal experience. I would argue that other types of work that are done because someone wants to impress others, wants to scratch their own itch, or just wants to help a good cause are often done better than tasks done strictly for pay. Hopefully we all have jobs doing what we love, but when that isn't the case there's a great chance that you do something better in your off time because you really give a shit about it.
Was there a good reason to structure your post as a series of snide questions and a flippant remark?
There are many other types of incentives and I have rarely done my best work for strictly financial ones. When contributing to an open source project you have to think that somewhere someone will look at the code you write and have the ability to publicly shame you if you do something truly stupid. Standing, respect, whatever you want to call it, is a big motivator for many people. If the same thing happens in many businesses there *may* be consequences, but often as long as it works well enough to collect the customer's money it ships. Personally, I've found more fugly code turds in various closed source projects than I've touched than in the open source world.
I think you're misreading or oversimplifying those situations. It's not that there was no demand, but it is more difficult to measure before the product exists. There was a market for smartphones and Apple 1) tends to put out good stuff, 2) already made iPods so they had some experience w making small portable consumer devices, 3) demonstrated that people loved the Apple branding of such things.
As for the Tesla, people plunked down $40000 reservations before the cars existed, and continue to do so for new models. If that doesn't show demand exists I'm not sure what would.
I'm sure there are products where there was no demand before the product existed, but I don't think you can call that the common case. Before a company sinks significant cost into designing a new product they generally run some numbers to measure the current market demand for the final product. Sure, some of that is wishful thinking, but some is evaluating current offerings and seeing how your product will be able to grab the customers. I'm sure Apple looked at BlackBerry devices before throwing their hat into the smartphone market and I'm sure Musk had more that just a hunch that rich people would buy a luxury electric car. However, when you have enough money, as did the creators of both these examples, the market analysis is a little less critical during the development stage just because a failed project isn't the end of the world. Apple could have kept going even if the iPhone flopped and Musk wouldn't run out of cash if Tesla crashed and burned during the roadster days.
Seriously: http://xkcd.com/605/
I bet someone in battery manufacturing is looking as adding capacity now in anticipation of such events. This could be quite an opportunity for some manufacturer with a bit of foresight. As more companies make and sell more electric cars I doubt Tesla will be the only company hunting for more, cheaper, better.
The parasitic worm is Trichinella spiralis and is actually relatively easy to kill with cooking. Many pork dishes involve long slow cooking which renders them worms completely harmless. We eat feral pigs rural northern Florida without a problem. Granted, I pick preparations like slow cooker Cuban pork for any wild meat to be absolutely it's safe. As for exposure for the butcher, that's crap. The worm is embedded in a cyst that is dissolved by stomach acid so don't lick your hand while cleaning the animal but that's pretty much it for trichinosis. Of course it is an animal so there's always bacteria as well but you get different but other dangerous strains in our commercial meat supply too.
Considering the question of pilot reliance on automation, and the vast canyon of difference between the training they receive and that of a typical automobile operator, I fear this particular solution (self-driving cars) will only compound an existing problem.
I can see this being a problem if automation is introduced piecemeal, but if you go from what we have now to something where the human driver is not required your example tends to make the opposite point for me. Before you say that cutting the human out is insane keep in mind that when something goes wrong in a car the stakes are not nearly as high as in a plane, both for the number of people involved, and for the fact that if the car engine cuts out and it coasts to a stop the result is generally not a pile of flaming wreckage in a hillside.
Keep in mind that a safer self-driving car only has to be better than the current batch of drivers and that is not a terribly high bar. Given how little training the average driver is required to have and thereby how badly many (most?) people drive doesn't that make the problem of creating a self-driving car that improves safety on the roads even simpler? I think you're trying to say that humans will be even worse drivers when they usually rely on automated technology, but there's not a ton of room at the bottom now. Based on most of the drivers I've seen, once you clear the initial figuring-it-out phase of your teenage years there doesn't seem to be a huge correlation between how long a person has been driving and how safe a driver they are. Other factors including fatigue, anger, and distraction appear to play a *much* larger role in their abilities. Keep in mind that the computer suffers from none of those problems AND can't get drunk. Shoot, just removing human competitive and aggressive instincts from the road would make driving much safer even if other mistakes remained constant.
Insurance companies may be terrible at everything else, but they can run statistics and if there is a self-driving car that gets in only 10% of the accidents that a human driver does there will be price incentives to push people toward that. Of course the incentives will only reduce your insurance by 20%, thereby making the customer happy and giving the insurance company a tidy profit since they don't pay out nearly as much.