My personal solution (which I do not expect to fit everyone) was to get an Asus Chrome book. I had a work Mac at home, and it started going wrong. I realised I would have no internet access of my own if it went, so I looked at getting a tablet, or something. I could get a ChromeBook for £230 at the time, which was cheaper than any equivalent tablet before you added the case and the keyboard. I have been using it for my home and commuting stuff every since. No camera, so you can't accidentally broadcast pictures. Outlook works best for mail. I tend to use my Mac at home for work e-mails, and home e-mails as well, because it is on a table, and has a better keyboard and screen; and use the ChromeBook when commuting (wi-fi on the coach), or at home when I am not working.
My original intent was not to separate work and home stuff. It happened by accident. But I like it.
Now, this is the weird thing. We know the people have managed to get new skills in later life. We know that damaged brains have managed to repurpose neighbouring regions of the brain. It is possible that this was done by the re-arrangement of existing cells. However, when almost every other part of the body, excepting the teeth and the lens of the eye, continuously regrows and repairs itself, what was the evidence that the brain does not?
I don't think this is a problem with science. Good science is constantly re-checking things that we 'know'. I think this is a problem with the reporting of science. Somehow 'you grow no new braincells after your teens' has become something that 'scientists have proved': it has become a meme, and now it reproduces so fast, it may never be slain.
Oumauamua could be an outgassing comet-like body. Or, if it is a solar sail, and outgassing, in which case any acceleration we see must be explained by photonics pressure alone. This calculation give the very low densities given in the post.
In one astronomy post web group (which is hardly a scientific survey) the original argument that Oumauamua was a solar sail were accepted as "promoting the discussion", but there were demands that early papers for the scientific community replying on arXiv were not properly peer reviewed and so should be stricken from the record. This is weird: not only were the authors and arCiv doing a reasonable job against a set of unsubstantiated claims, you would also not expect a peer-reviewed article to come as quickly. So, in effect we are allowing people to post wacky propositions about alien spacecraft, but not allowing rational criticism from the body of experts whoo is most likely to have a useful contribution.
Astronomy is a long game. Have your little victories winning short-term arguments this way. The stars will be much where they were when you have long gone. And whoever was right will still be right. We can all wait.
Kodak begat Polaroid,
and Polaroid begat Xerox,
and Xerox begat Apple..
It's weird because these were all agile companies. Perhaps you have to start afresh.
Yep. Those panda guys are real go-getters. No-wonder so many of them burn out. Live fast, die young, eh?
Actually the real issue is a bit more complex than some comments suggest. Anyone who has owned a greyhound will tell you how little exercise they need. 10 minutes of going bonkers a day is fine. Cheetahs can run fast but will do nothing for hours if the don't need to. If you can go into low power mode, then you don't have to go hunting as often, with all the risks,
I am still getting links to "Financial experts agree that the Euro may crash in months", "Analysts say EU likely do break up on the next decade", and stuff like that. No citations, no experts that I've ever heard of. And why are they targeting a strong European like me with this junk? unless they are sending it to everyone.
It would have to learn how to do the chemistry at factory scales to do that. This is possible but unlikely to be economic.
If anything, the interesting work would be to go in the other direction. It is possible to make lab-on-a-chip units where the reagents are sealed in and each experiment is done in a small drop. This means you can repeat each experiment many times and see whether the results are repeatable. If your process is catalysed by some low concentration impurity, then that effect will vary with the number of molecules of that impurity, and you would expect a larger scatter in the yield.
Remember, the Viking lander experiments were designed to work on a planet that no-one had been to before. We expected some solid carbon from meteorites, which ought to be pretty inert even when wartmed. We might hope for a sign of organic (that's chemical organic, though it could be life too). The perchlorates under the surface were a total surprise to everyone. So, it's not the case that the people who designed the experiments were (a) fools and should be fired, or (b) geniuses who were right all along and should be honoured. This is just science doing its thing. Nice to see the Viking results make sense, even though it didn't tell us much.
I have not worked on face recognition, but I have worked on text recognition, which has some of the same problems. You have several recognition algorithms that work. You can make some super-algorithm by polling all of the better ones. Oddly, the performance of the super-algorithm is measurably worse if you take away the worst one out of ten. It turns out that diversity is an important part of recognition. Sometimes the errors are trivial - I heard from someone who's lab had a face location system that had been trained on internet pictures of faces and cats, and could identify everyone in the lab except for one guy with a beard who was always a cat.
Let us compare this with our baseline, which is what we do. We recognise people but we come with a lot of faults and biases. We are able to see small differences in people like us, and are less able to distinguish people who are different. We can have our memories altered after the event. We cannot even run proper tests the way we can with computer recognition systems. People usually know when they are being tested. So, the baseline is pretty poor, too.
We can fix this. I see no reason why computer recognition systems should not be superior to humans. But this software is going to need training and development, and this process should be open if we are to be sure it does not have any built-in biases. Software does not spring fully-formed from the brow of Alan Turing, and to expect it to work from day one is unrealistic. To rely on a boxed solution from some private contractor is also unrealistic, and also dangerous. If we automatically identify any development in recognition systems as Orwellian evil, then I am afraid that's what we are going to get.
The best you can probably do for a future historian is to record as complete a picture as you can without filtering. If you have space, and I don't see why you don't, you could chuck in a copy of the entire web, warts and all. It would, initially, be helpful to have a snapshot of how a few random families actually lived, so that the historians could get some context before wading in. And some digital library might give a general opinion of what we knew and believed. But any filtering process, even if it improves the quality of the record, must also add some bias. The whole picture, or if we can't manage that, then a random snapshot, may be more helpful.
There have been lots of posts demanding computer security or privacy, but little in the way of actual suggestions. Here's one that went out of fashion: the executive terminal.
The PDP-11 computer could have several users. Each user had their own slice of the computer memory to play with. Another user could not address stuff in your space by setting a pointer off the end of their own space, because the memory was paged, in 4-kilobyte chunks if I remember correctly. The operating system could not work that way, because it assigns who could see what memory. There have to be additional instructions to run the memory paging. These instructions could only be accessed by the person using the executive terminal tty0 - physically plugged into the unique tty0: serial port on the processor board. You did this to kill off programs that had hung, and the sort of things you would use 'sudo' for in Linux. There was a Texas thermal printer and keyboard which was often used for tty0, so every keystroke and every character that was replied was permanently recorded on paper.
Almost every machine today has a Von Neumann architecture with one 64-bit addressable space of memory that is shared by everyone. All users, remote and local, have equal access. The modern equivalent of the executive terminal might be a laptop where I could get superuser privileges, but only on commands typed into the laptop keyboard.
This is a lovely idea. I wonder whether we could also come up with a digital filter that could take out Parkinson's tremors when using an ordinary mouse. I want to work on it.
I used to sit next to someone who had RSI so used to Dragon Dictate (as it was then) to do a lot of his keyboard work, which included writing code. On bad dates he could use it to control the mouse too. This is not easy, so it depends on how determined your mother-in-law is. If she has computer skills but has just lost the manual control, then a Dragon product may do the job. If she hasn't handled a computer in some time, so she will be getting used to a new computer, a new OS, and everything else being in the wrong place as well as this new tool, all at once; then it is a big ask.
I was replying to another post, and the Apple spelling correct automatically replaced 'elections' with 'electric actions'. It may have started already.
I have considered systems where you could vote for and against people. My local elections usually some up with about 8 candidates, where the best are faceless, and the worst are the nasty sort of nationalist. I can't pick one Iike, but there are certainly ones I hate. People with strong opinions for one person would vote for them and blackball everyone else. So, being able to vote for and against any candidate almost reduces to approval voting.
Approval voting also is easy to understand, and gives a sensible result in one vote. If there is a simpler system without the problems of the single vote for, I can't see it.
This may be one of the more far sighted solutions.
We have seen Google mis-identifying photos, and IBM's chatbot gown Nazi. However, it would be sensible if we were to try and get some AI system, and get it to act as though it was a member of parliament. It should be capable of reading letters from constituents, and political correspondence, and from expert committees, and to understand the differences between them. People could influence it by writing to it. It might be able to ask questions in return. From this we would learn the weakness of such a system. It would have to recognise all forms of influence, and rank them so a lengthy reply from one user might be as influential as a slogan repeated by ten thousand. It should be able to explain why it voted in a particular way, if it knew.
At some point, people will point to the Rt. Hon. Watson M.P, and say that their voting patterns have been consistently sound, and the time has come to award them an actual vote in the Commons. Then two votes...
Apple were the shiny Linux-in-a-box for us for many years...
My company provides Linux systems for the motion picture industry. We have to specify everything in the machine build and in the Linux installation to make sure it worked. We also needed a portable solution for developers and engineers. We tried using Dell laptops, but the build was not consistent - we found that our application would work on one laptop, but would not work on a slightly later version of the same machine, built to the same nominal standard. The Mac laptops at the time had a very consistent build - if you were in the field in most parts of the world and your laptop broke, you could go into a store, get a new one, install our software, and it would just work.
This was really helpful. We are prepared to pay the Apple markup to avoid the sysadmin woes for the field operators. Unfortunately, there are two meanings for the phrase 'it just works' and Apple seem to be moving towards the other one. This is a great pity because Apple were helping Linux even if they weren't exactly your regular Linux distro. Interesting times...
This. Plus bring some of your attitude with you. The people who write the job adverts are probably HR drones or recruiting consultants, and not the people who will interview you, or work with you.
I remember being given a coding test as part of a job interview. One of the questions involved parsing a long expression in 'C' without brackets. My reply was (a) no-one ought to write code like this; (b) I can probably remember or guess the precedence for the weirder operators, such as bitwise XOR, but I wouldn't trust my memory when it is all on page 53 of K&R (the yellow, crumbly one); and finally (c) my best answer, which may have been right.
There are a number of word-values here that are peculiar to mathematics, and probably giving the wrong opinion for a general audience. To many people, myself included, this reads a bit like the cold fusion debacle. But it isn't. With cold fusion there ought to be an experiment that others can repeat or not as the came may be, but many critical factors were omitted. With mathematics, the written analysis is everything.
So, let's have a go at re-drafting the summary. Blum feels they have got somewhere with P != NP. He is unlikely to just trust his own intuition, so the work has probably had several reviewers by the time he feels confident enough to go public. At which point he probably still feels there may be errors that more eyes will spot, but right or wrong, this is something that people should be looking at.
And look they do. If he has really slain this monster, it will probably take a year or so of hard looking before most mathematicians will provisionally accept it. And even then they will keep poking.
I remember a very young kid's program on the BCC Acorn computer that asked you to draw and colour a house for a girl on the screen. You could chose the colours for the door and the windows by putting in numbers. So a child of mine started putting in the numbers. Then we wondered, if they specified numbers from one to seven, what happened when you put in eight, what happened at sixteen, what happened with minus one, what happened with two and a half? To our delight we got flashing colours, we got some of the original colours, and we could guess at the rules within the program.
Education is different for everyone. Sometimes it works best when your are not blindly following the rules, but asking why they exist, and what happens when you don't; preferably in some non-fatal environment. Something like Windows (this just got autocorrected to Sindows - really!, but I am on an Apple, which is morally no better) everything is supposed to work provided you obey the rules. On Linux installs, you get the same drag and drop interface, but sometimes the tool you want doesn't work well on your installation. You have to fish about for an alternative. Or you can find out what you can do by going on the web. You meet others who have met the same problem. It ends up being a scavenger hunt. You may even find out how to fix it yourself, and post your solution. Hell, you can even teach yourself some programming.
This is not an alternative to learning how to format your Windows documents correctly. The finder and fixer skills you get in keeping your machine running as you want it to run, will probably mean working out how to do CSS is less of a pain in the pancreas.
The development is entirely reasonable, and follows the normal pattern of risk reduction in an emerging technology.
Oil drilling started off on land, then it moved to shallow waters, then it moved to deep waters. The technology developed on the easiest sites then moved to the harder sites when there were not enough easy sites.
Some of the first wind farms in the UK were on Scottish islands. Not only did they have plenty of wind, the inhabitants of the islands used diesel generators, which were over four times the cost of mainland electricity; so the site was likely to be profitable even if it used first-generation parts. It reduced the risk of a highly visible venture site being unprofitable, and blocking the chance of making others.
Putting windmills in shallow water is quite like building them on land. You sink piles into shallow water or boggy land. The existing technologies can be used with longer piles, but in the end another solution is going to be cheaper. The cheaper solution is going to be anchored platforms. The people with the most experience of these are oil companies.
Instead of asking, "why an oil company?", ask "why not all oil companies with offshore drilling?". The US division - fossil fuels are freedom, wind and renewables are tree-hugging socialism - does not hold in Europe.
If you have a missile chasing a missile, the persuer ought to take a shorter course. This is not the case if it overshoots, so the last part of the interception should be from behind rather than from the side. The interceptor will probalby explode and destroy the target with shrapnel, as closing the last bit is hard. Another, trickier but possible option is trying to hit the target from the side or head on. You can make your rocket lighter and more agile if you omit the explosives and shrapnel, and try to impact the target, using your relative kinetic energy to do the job.
ICBS are different again. They have several phases. You have the boost phase when it is ascending. Hit it with a laser and there is a lot of propellant and stuff that will destroy it for you. But the boost phase may be only minutes long. Then there is the ballistic phase: the warhead is moving largely under its own momentum. It is hard to destroy. You can make it pretty laser-resistant with a shiny foil and an ablative layer. If it has multiple warheads, you want to get it before it divides, or before they have separated much. Finally, there is the re-entry phase. Any decoy warheads will be lighter and not survive re-entry. You know the real ones, but you have left it pretty late.
I think it makes a lot of sense to dig out your last cold war experiments, and find out whether they still work.
What he means is what he says - you can have an AI helper that you can talk to, and replies in a seemingly intelligent manner, without having all the self awareness, and the OMG! Bots are going to murder us in our beds!! nonsense we have had from Steven Hawking, or the singularity quasi-rapture predictions of Kurzweil et al. The first implementation of AI's is probably going to be something that will be guided by humans rather than replacing them. Cars that drive themselves, but will come safely to a stop or ask for help if they see something they cannot understand. TV remotes that can respond to speech, and find something you might like to watch without flipping channels. Less "Open pod bay doors, Hal" "I'm sorry, Dave. I cannot do that", and more "Shit! Did I forget to turn off the oven, Hal?" "Yes, you did, Dave. I turned it down a bit there was nothing in it. Would you like me to turn it off?
There is a lot of AI bull about. Oddly enough, I don't think this particular article was an example. But maybe I'm wrong.
This is an entirely sensible thing to do. You might have a game that uses the gyroscope. Embedded within that game there might be a rogue application that also uses the gyroscope data to measure the tilt as a result of using the keyboard, and report that along with your high score, or whatever to some game sever. If you have some security mode so when you are entering a password, it disables keyboard sharing, screen grabs, the camera (looking for reflections in you glasses) and the microphone (in case you say the numbers aloud) it must also disable the gyroscope and accelerometer. It does not matter that the process is not reliable - all we need to know is that is could work, so we can put in a fix.
Too often we learn what is possible after the event. We have to pay ingenious people to report weaknesses in the system. If you don't then the only ones working on it are the ones with something to gain.
Slashdot Mirror
My personal solution (which I do not expect to fit everyone) was to get an Asus Chrome book. I had a work Mac at home, and it started going wrong. I realised I would have no internet access of my own if it went, so I looked at getting a tablet, or something. I could get a ChromeBook for £230 at the time, which was cheaper than any equivalent tablet before you added the case and the keyboard. I have been using it for my home and commuting stuff every since. No camera, so you can't accidentally broadcast pictures. Outlook works best for mail. I tend to use my Mac at home for work e-mails, and home e-mails as well, because it is on a table, and has a better keyboard and screen; and use the ChromeBook when commuting (wi-fi on the coach), or at home when I am not working.
My original intent was not to separate work and home stuff. It happened by accident. But I like it.
Now, this is the weird thing. We know the people have managed to get new skills in later life. We know that damaged brains have managed to repurpose neighbouring regions of the brain. It is possible that this was done by the re-arrangement of existing cells. However, when almost every other part of the body, excepting the teeth and the lens of the eye, continuously regrows and repairs itself, what was the evidence that the brain does not?
I don't think this is a problem with science. Good science is constantly re-checking things that we 'know'. I think this is a problem with the reporting of science. Somehow 'you grow no new braincells after your teens' has become something that 'scientists have proved': it has become a meme, and now it reproduces so fast, it may never be slain.
Also for your consideration...
gnuplot : user-friendly as a cornered rat, but powerful. I use it for roughing out functions for work, and producing .eps graphs for papers.
octave: freeware Matlab. I use this a lot for prototyping.
Oumauamua could be an outgassing comet-like body. Or, if it is a solar sail, and outgassing, in which case any acceleration we see must be explained by photonics pressure alone. This calculation give the very low densities given in the post.
In one astronomy post web group (which is hardly a scientific survey) the original argument that Oumauamua was a solar sail were accepted as "promoting the discussion", but there were demands that early papers for the scientific community replying on arXiv were not properly peer reviewed and so should be stricken from the record. This is weird: not only were the authors and arCiv doing a reasonable job against a set of unsubstantiated claims, you would also not expect a peer-reviewed article to come as quickly. So, in effect we are allowing people to post wacky propositions about alien spacecraft, but not allowing rational criticism from the body of experts whoo is most likely to have a useful contribution.
Astronomy is a long game. Have your little victories winning short-term arguments this way. The stars will be much where they were when you have long gone. And whoever was right will still be right. We can all wait.
Kodak begat Polaroid, and Polaroid begat Xerox, and Xerox begat Apple.. It's weird because these were all agile companies. Perhaps you have to start afresh.
Yep. Those panda guys are real go-getters. No-wonder so many of them burn out. Live fast, die young, eh?
Actually the real issue is a bit more complex than some comments suggest. Anyone who has owned a greyhound will tell you how little exercise they need. 10 minutes of going bonkers a day is fine. Cheetahs can run fast but will do nothing for hours if the don't need to. If you can go into low power mode, then you don't have to go hunting as often, with all the risks,
I am still getting links to "Financial experts agree that the Euro may crash in months", "Analysts say EU likely do break up on the next decade", and stuff like that. No citations, no experts that I've ever heard of. And why are they targeting a strong European like me with this junk? unless they are sending it to everyone.
It would have to learn how to do the chemistry at factory scales to do that. This is possible but unlikely to be economic.
If anything, the interesting work would be to go in the other direction. It is possible to make lab-on-a-chip units where the reagents are sealed in and each experiment is done in a small drop. This means you can repeat each experiment many times and see whether the results are repeatable. If your process is catalysed by some low concentration impurity, then that effect will vary with the number of molecules of that impurity, and you would expect a larger scatter in the yield.
This is a pretty good summary.
Remember, the Viking lander experiments were designed to work on a planet that no-one had been to before. We expected some solid carbon from meteorites, which ought to be pretty inert even when wartmed. We might hope for a sign of organic (that's chemical organic, though it could be life too). The perchlorates under the surface were a total surprise to everyone. So, it's not the case that the people who designed the experiments were (a) fools and should be fired, or (b) geniuses who were right all along and should be honoured. This is just science doing its thing. Nice to see the Viking results make sense, even though it didn't tell us much.
I have not worked on face recognition, but I have worked on text recognition, which has some of the same problems. You have several recognition algorithms that work. You can make some super-algorithm by polling all of the better ones. Oddly, the performance of the super-algorithm is measurably worse if you take away the worst one out of ten. It turns out that diversity is an important part of recognition. Sometimes the errors are trivial - I heard from someone who's lab had a face location system that had been trained on internet pictures of faces and cats, and could identify everyone in the lab except for one guy with a beard who was always a cat.
Let us compare this with our baseline, which is what we do. We recognise people but we come with a lot of faults and biases. We are able to see small differences in people like us, and are less able to distinguish people who are different. We can have our memories altered after the event. We cannot even run proper tests the way we can with computer recognition systems. People usually know when they are being tested. So, the baseline is pretty poor, too.
We can fix this. I see no reason why computer recognition systems should not be superior to humans. But this software is going to need training and development, and this process should be open if we are to be sure it does not have any built-in biases. Software does not spring fully-formed from the brow of Alan Turing, and to expect it to work from day one is unrealistic. To rely on a boxed solution from some private contractor is also unrealistic, and also dangerous. If we automatically identify any development in recognition systems as Orwellian evil, then I am afraid that's what we are going to get.
The best you can probably do for a future historian is to record as complete a picture as you can without filtering. If you have space, and I don't see why you don't, you could chuck in a copy of the entire web, warts and all. It would, initially, be helpful to have a snapshot of how a few random families actually lived, so that the historians could get some context before wading in. And some digital library might give a general opinion of what we knew and believed. But any filtering process, even if it improves the quality of the record, must also add some bias. The whole picture, or if we can't manage that, then a random snapshot, may be more helpful.
There have been lots of posts demanding computer security or privacy, but little in the way of actual suggestions. Here's one that went out of fashion: the executive terminal.
The PDP-11 computer could have several users. Each user had their own slice of the computer memory to play with. Another user could not address stuff in your space by setting a pointer off the end of their own space, because the memory was paged, in 4-kilobyte chunks if I remember correctly. The operating system could not work that way, because it assigns who could see what memory. There have to be additional instructions to run the memory paging. These instructions could only be accessed by the person using the executive terminal tty0 - physically plugged into the unique tty0: serial port on the processor board. You did this to kill off programs that had hung, and the sort of things you would use 'sudo' for in Linux. There was a Texas thermal printer and keyboard which was often used for tty0, so every keystroke and every character that was replied was permanently recorded on paper.
Almost every machine today has a Von Neumann architecture with one 64-bit addressable space of memory that is shared by everyone. All users, remote and local, have equal access. The modern equivalent of the executive terminal might be a laptop where I could get superuser privileges, but only on commands typed into the laptop keyboard.
Same thing, different article. Not available yet, I believe...
https://blogs.microsoft.com/tr...
This is a lovely idea. I wonder whether we could also come up with a digital filter that could take out Parkinson's tremors when using an ordinary mouse. I want to work on it.
I used to sit next to someone who had RSI so used to Dragon Dictate (as it was then) to do a lot of his keyboard work, which included writing code. On bad dates he could use it to control the mouse too. This is not easy, so it depends on how determined your mother-in-law is. If she has computer skills but has just lost the manual control, then a Dragon product may do the job. If she hasn't handled a computer in some time, so she will be getting used to a new computer, a new OS, and everything else being in the wrong place as well as this new tool, all at once; then it is a big ask.
I was replying to another post, and the Apple spelling correct automatically replaced 'elections' with 'electric actions'. It may have started already.
Approval voting gets my vote.
I have considered systems where you could vote for and against people. My local elections usually some up with about 8 candidates, where the best are faceless, and the worst are the nasty sort of nationalist. I can't pick one Iike, but there are certainly ones I hate. People with strong opinions for one person would vote for them and blackball everyone else. So, being able to vote for and against any candidate almost reduces to approval voting.
Approval voting also is easy to understand, and gives a sensible result in one vote. If there is a simpler system without the problems of the single vote for, I can't see it.
This may be one of the more far sighted solutions.
We have seen Google mis-identifying photos, and IBM's chatbot gown Nazi. However, it would be sensible if we were to try and get some AI system, and get it to act as though it was a member of parliament. It should be capable of reading letters from constituents, and political correspondence, and from expert committees, and to understand the differences between them. People could influence it by writing to it. It might be able to ask questions in return. From this we would learn the weakness of such a system. It would have to recognise all forms of influence, and rank them so a lengthy reply from one user might be as influential as a slogan repeated by ten thousand. It should be able to explain why it voted in a particular way, if it knew.
At some point, people will point to the Rt. Hon. Watson M.P, and say that their voting patterns have been consistently sound, and the time has come to award them an actual vote in the Commons. Then two votes...
Apple were the shiny Linux-in-a-box for us for many years...
My company provides Linux systems for the motion picture industry. We have to specify everything in the machine build and in the Linux installation to make sure it worked. We also needed a portable solution for developers and engineers. We tried using Dell laptops, but the build was not consistent - we found that our application would work on one laptop, but would not work on a slightly later version of the same machine, built to the same nominal standard. The Mac laptops at the time had a very consistent build - if you were in the field in most parts of the world and your laptop broke, you could go into a store, get a new one, install our software, and it would just work.
This was really helpful. We are prepared to pay the Apple markup to avoid the sysadmin woes for the field operators. Unfortunately, there are two meanings for the phrase 'it just works' and Apple seem to be moving towards the other one. This is a great pity because Apple were helping Linux even if they weren't exactly your regular Linux distro. Interesting times...
This. Plus bring some of your attitude with you. The people who write the job adverts are probably HR drones or recruiting consultants, and not the people who will interview you, or work with you.
I remember being given a coding test as part of a job interview. One of the questions involved parsing a long expression in 'C' without brackets. My reply was (a) no-one ought to write code like this; (b) I can probably remember or guess the precedence for the weirder operators, such as bitwise XOR, but I wouldn't trust my memory when it is all on page 53 of K&R (the yellow, crumbly one); and finally (c) my best answer, which may have been right.
There are a number of word-values here that are peculiar to mathematics, and probably giving the wrong opinion for a general audience. To many people, myself included, this reads a bit like the cold fusion debacle. But it isn't. With cold fusion there ought to be an experiment that others can repeat or not as the came may be, but many critical factors were omitted. With mathematics, the written analysis is everything.
So, let's have a go at re-drafting the summary. Blum feels they have got somewhere with P != NP. He is unlikely to just trust his own intuition, so the work has probably had several reviewers by the time he feels confident enough to go public. At which point he probably still feels there may be errors that more eyes will spot, but right or wrong, this is something that people should be looking at.
And look they do. If he has really slain this monster, it will probably take a year or so of hard looking before most mathematicians will provisionally accept it. And even then they will keep poking.
I remember a very young kid's program on the BCC Acorn computer that asked you to draw and colour a house for a girl on the screen. You could chose the colours for the door and the windows by putting in numbers. So a child of mine started putting in the numbers. Then we wondered, if they specified numbers from one to seven, what happened when you put in eight, what happened at sixteen, what happened with minus one, what happened with two and a half? To our delight we got flashing colours, we got some of the original colours, and we could guess at the rules within the program.
Education is different for everyone. Sometimes it works best when your are not blindly following the rules, but asking why they exist, and what happens when you don't; preferably in some non-fatal environment. Something like Windows (this just got autocorrected to Sindows - really!, but I am on an Apple, which is morally no better) everything is supposed to work provided you obey the rules. On Linux installs, you get the same drag and drop interface, but sometimes the tool you want doesn't work well on your installation. You have to fish about for an alternative. Or you can find out what you can do by going on the web. You meet others who have met the same problem. It ends up being a scavenger hunt. You may even find out how to fix it yourself, and post your solution. Hell, you can even teach yourself some programming.
This is not an alternative to learning how to format your Windows documents correctly. The finder and fixer skills you get in keeping your machine running as you want it to run, will probably mean working out how to do CSS is less of a pain in the pancreas.
The development is entirely reasonable, and follows the normal pattern of risk reduction in an emerging technology.
Oil drilling started off on land, then it moved to shallow waters, then it moved to deep waters. The technology developed on the easiest sites then moved to the harder sites when there were not enough easy sites.
Some of the first wind farms in the UK were on Scottish islands. Not only did they have plenty of wind, the inhabitants of the islands used diesel generators, which were over four times the cost of mainland electricity; so the site was likely to be profitable even if it used first-generation parts. It reduced the risk of a highly visible venture site being unprofitable, and blocking the chance of making others.
Putting windmills in shallow water is quite like building them on land. You sink piles into shallow water or boggy land. The existing technologies can be used with longer piles, but in the end another solution is going to be cheaper. The cheaper solution is going to be anchored platforms. The people with the most experience of these are oil companies.
Instead of asking, "why an oil company?", ask "why not all oil companies with offshore drilling?". The US division - fossil fuels are freedom, wind and renewables are tree-hugging socialism - does not hold in Europe.
Yes, and no...
If you have a missile chasing a missile, the persuer ought to take a shorter course. This is not the case if it overshoots, so the last part of the interception should be from behind rather than from the side. The interceptor will probalby explode and destroy the target with shrapnel, as closing the last bit is hard. Another, trickier but possible option is trying to hit the target from the side or head on. You can make your rocket lighter and more agile if you omit the explosives and shrapnel, and try to impact the target, using your relative kinetic energy to do the job.
ICBS are different again. They have several phases. You have the boost phase when it is ascending. Hit it with a laser and there is a lot of propellant and stuff that will destroy it for you. But the boost phase may be only minutes long. Then there is the ballistic phase: the warhead is moving largely under its own momentum. It is hard to destroy. You can make it pretty laser-resistant with a shiny foil and an ablative layer. If it has multiple warheads, you want to get it before it divides, or before they have separated much. Finally, there is the re-entry phase. Any decoy warheads will be lighter and not survive re-entry. You know the real ones, but you have left it pretty late.
I think it makes a lot of sense to dig out your last cold war experiments, and find out whether they still work.
He just means "non-customer facing", right?
I think you are just grumpy this morning.
What he means is what he says - you can have an AI helper that you can talk to, and replies in a seemingly intelligent manner, without having all the self awareness, and the OMG! Bots are going to murder us in our beds!! nonsense we have had from Steven Hawking, or the singularity quasi-rapture predictions of Kurzweil et al. The first implementation of AI's is probably going to be something that will be guided by humans rather than replacing them. Cars that drive themselves, but will come safely to a stop or ask for help if they see something they cannot understand. TV remotes that can respond to speech, and find something you might like to watch without flipping channels. Less "Open pod bay doors, Hal" "I'm sorry, Dave. I cannot do that", and more "Shit! Did I forget to turn off the oven, Hal?" "Yes, you did, Dave. I turned it down a bit there was nothing in it. Would you like me to turn it off?
There is a lot of AI bull about. Oddly enough, I don't think this particular article was an example. But maybe I'm wrong.
This is an entirely sensible thing to do. You might have a game that uses the gyroscope. Embedded within that game there might be a rogue application that also uses the gyroscope data to measure the tilt as a result of using the keyboard, and report that along with your high score, or whatever to some game sever. If you have some security mode so when you are entering a password, it disables keyboard sharing, screen grabs, the camera (looking for reflections in you glasses) and the microphone (in case you say the numbers aloud) it must also disable the gyroscope and accelerometer. It does not matter that the process is not reliable - all we need to know is that is could work, so we can put in a fix.
Too often we learn what is possible after the event. We have to pay ingenious people to report weaknesses in the system. If you don't then the only ones working on it are the ones with something to gain.