If your code ever has the expression X / X and X is capable of being 0, you have a bug in your algorithm.
If you expect 0/0 = NaN, then it's not necessarily a bug.
You are correct there. A more accurate statement on my part should have been:
If your code ever has the expression X / X and X is capable of being 0, if your code does not treat this as an error and handle it properly then you have a bug in your algorithm.
Treating 0/0 as NaN is an example of handling the error properly, considering you are no longer treating the result as a valid number.
We had five bucks, and we decided to divide it up. After we were done dividing, the money was all gone. That is dividing by zero. It was divided into zero pieces.
Dividing something into zero pieces is not the same as dividing by zero. If you have no money left after dividing five bucks to five people, then the next attempt at dividing would be 0 / 5, not 0 / 0.
Programmers who primarily deal with integers and money units find it natural to have anything that doesn't make sense to default to zero.
Plenty of lazy programmers or those forced into using frameworks developed by other lazy programmers may be used to this. Perhaps even programmers forced to use primitive languages with no other options are used to it. But anyone else who knows what they are doing do not do this. This is similar to all of those developers who use NULL as if it means 0. Just because a lot of people do it doesn't make it any less idiotic.
So in the end the answer is to use high level languages, and use money objects with sensible defaults instead of floats and ints for money values.
Even low level languages have the capability of handling money correctly. It just requires more effort by the programmer, just like anything else done in a low level language. Considering all high level languages eventually have their code compiled down to the lowest level language understood by the CPU, low level languages are certainly capable of any calculation that a high level language is.
There are no specifics in that post. Just some vague mention of semantic vs syntactic content and a claim that it is so obvious you don't even need to explain it.
All I can make from your semantic/syntactic analogy is that the physiology of a human brain and its computational ability may equate to syntactic content, but with no inherent semantic meaning. The consciousness that is created by these computations is what provides the semantic meaning. aka I think therefore I am.
But there is no absolute reason why a manufactured machine with similar computational ability as the human brain couldn't also create a consciousness that could give itself the semantic meaning you are referencing.
You can never state that something is not sufficient just because no one has figured it out yet.
That it's logically impossible is reason enough! Consider for a moment a simple example: I claim that it is impossible to clear 5 lines simultaneously in a game of Tetris. Would you say that claim is nonsense and it's only a matter of time before someone figures it out? Of course not. You can clearly demonstrate that it is an impossibility. The same is true for computational approaches to AGI -- they are logically impossible.
Why do you believe in such silly nonsense when it's clear that those beliefs are pure fantasy?
You are correct that you can clearly demonstrate that clearing 5 lines simultaneously in Tetris is impossible. But that is because you can be very specific as to why: the longest piece is only 4 lines long. Notice I didn't say something vague like: it is logically impossible to clear 5 lines. I was incredibly specific.
If your answer to why computational approaches to AGI is logically impossible are not also as specific, then you are just spouting nonsense. Just saying it is impossible is not an acceptable answer, and neither are other deflections such as calling it silly nonsense or pure fantasy.
If you don't have specific reasons why it is impossible, which show a thorough understanding of why neuron interaction cannot be replicated by a Turing complete computation system, or why simulating neuron interaction is not enough to create AGI, then just stop replying.
I have no idea. Neither does anyone else. That doesn't change the fact that computation alone has been show to be insufficient. That's pretty well established.
You can never state that something is not sufficient just because no one has figured it out yet. Most likely 100% of all significant technological advances were preceded by people who tried and failed. Often the people who finally figure it out were among the people who had failed multiple times in the past.
To say something isn't possible you have to know a great deal about why it is not possible. For instance we know that silicon transistors cannot shrink indefinitely because of our knowledge of physical limitations such as the width of a silicon atom. We don't say that transistors cannot be 5 nm wide just because all attempts at making 5 nm wide transistors have failed; that would be stupid.
It could not possibly be established that computation alone cannot create general AI because we have not even approached computers with enough transistors to match the 100 trillion synapses in an average adult. Even once we reach that milestone it will take a large amount of research before people could claim it simply isn't possible, and even then there would be a good chance they are still wrong.
Realistically the only way to say it is impossible to create general AI with computation alone is to actually create a general AI by other means. Only then could we possibly understand the process enough to know for sure it couldn't be done another way. Although even in that case there is no guarantee someone won't come up with a more elegant and/or efficient way in the future which only uses computation.
Here is an update to this ruling found in another article:
Update: Uber pointed out that the ruling only applies to one driver. “Reuters’ original headline was not accurate. The California Labor Commission’s ruling is non-binding and applies to a single driver,” a spokesperson said. “Indeed it is contrary to a previous ruling by the same commission, which concluded in 2012 that the driver ‘performed services as an independent contractor, and not as a bona fide employee.’ Five other states have also come to the same conclusion. It’s important to remember that the number one reason drivers choose to use Uber is because they have complete flexibility and control. The majority of them can and do choose to earn their living from multiple sources, including other ride sharing companies.”
But in all seriousness, you still need someone there even on the freeway. Although, maybe three trucks can be in a train/convoy. But someone has to be there in case something goes wrong. Wheel falls apart, something comes unstrapped, etc. Monitoring it and having AAA come I don't think is as feasible as paying someone to sit there and deal with it when it happens.
As long as the truck can pull over if there is a problem, which it would need to even in your three trucks in a convoy solution, then this problem is easy to solve. You even mention the solution in your post: AAA. A commercial trucking version of AAA would have drivers to come fix any of these failures and get the trucks to their destinations. Who knows how many of these truckers they would need when this technology becomes ubiquitous, but its likely to be a very small percentage of the current trucking industry.
And struggling to maintain an outdated system in some kind of virtual environment isn't too risky?
Who said anything about upgrades/maintenance? Maybe he has to build it once, certify it once, deploy once on an embedded system.
If the job requirement says "25 years" then that's what he has to do. It wouldn't even be an unusual specification for military.
The only reason to need a consistent development environment is for upgrades and maintenance. If they never need to debug errors or add features then there is no reason to have a development environment at all. Just pass around binary executables.
Compilers improve their optimization algorithms over time. I've personally helped debug issues due to this on a number of occasions.
I assumed when the AC said new compilers alter the output of their algorithms, he meant the actual output of the functions, RPCs, services, etc. Optimization algorithms of the compilers shouldn't affect this unless they are broken. Or at least that would be my assumption. I would love clarification on how these compiler optimizations can cause different outputs from someone who has done this professionally though.
It does make sense that changes to math libraries would alter output because of differences in precision. Any other libraries, such as those which handle concurrency, would also open up possibilities for different behavior. I just don't see how compiler optimizations should make any difference.
While the summary is trying to make this some kind of huge rebuff of the President by democrats in Congress, the only serious problem with this bill is it was for too long of a time period. Obama is only in office for another 18 months, and this fast track authority would have extended years after he is gone. This vote had almost nothing to do with democrats not trusting Obama; it was them not trusting the unknown President who will take his mantle a couple years from now.
Technology and having it available does not imply that we have to use it.
At least so far I don't see mushroom clouds popping up left and right. And don't tell me we couldn't do that!
You are quite correct that not all technology will be used just because it exists. A better statement would be that all technologies which can be used to improve someone's utility will be used. Using nuclear weapons does not benefit a country when their opponents have them or if the political fallout would be too severe. And there are plenty of conventional weapons which are just as effective for most warfare requirements.
None of this has anything to do with consumer and corporate face recognition software. Augmented reality devices and consumer data mining are simply too useful to hold back. The entire lobbying industry would be a disgrace if they couldn't stop these regulations from impeding this progress because they will certainly be well funded.
But at least we can stop such sensitive data to be shared and processed en masse by both corporations and governments alike, by outlawing the practise.
An interesting side effect of this would be to give even more advantages to large companies. If this technology actually provides value, and small companies cannot buy this data from third parties, then only companies large enough to amass their own huge data sets have this advantage.
I am almost convinced the best route is to force large companies to open up their data (for a fee), similarly to how we forced certain utility providers to rent out their physical infrastructure. This lets a local grocer have the same data Walmart has. It also allows entrepreneurial AI companies to have access to the huge data sets companies like Facebook and Microsoft have access to.
Like encryption, those high powered IR LEDs to blind cameras, only attract attention -- until everyone is doing it. Once encryption is outlawed, only outlaws will use encryption. Similarly wearing IR LEDs wired into your hat.
If this ever became a trend, the device makers would just start IR cut filtering their cameras.
If Software developers made $250-500k a year, I guarantee there would be no shortage of CS graduates.
Of course there would be no shortage, because 80% of current United States CS-related jobs would be off-shored to countries without such outrageous salaries. If you inflate wages above what the market will bear (taking into account the international market), then any industry will suffer the same fate the US auto industry suffered last century.
Saying that NSA surveillance cannot prevent a 9/11 style attack is silly, although saying it most likely will not is far more accurate (IMHO). The New America Foundation think tank published a report claiming that of 225 investigations performed since 9/11, only 1.8% of them were initiated based on NSA bulk metadata surveillance and 4.4% of the cases were assisted by the NSA spying program. These are obviously low figures so by shear odds it is likely the next huge terror plot would not be caught by NSA metadata spying.
But considering PRISM allegedly cost $20 million of a total US intelligence budget of around $75 billion annually over the past few years, the NSA program could still be considered a bargain. Although probably not once you factor in $35 billion in lost GDP because of a new found lack of trust of USA companies abroad.
And once again, I was only playing devil's advocate and trying to explain the other side of the argument. I personally believe the costs to personal freedoms dwarf the benefits of all enhanced security related programs since 9/11.
Honest question. What is the "theoretical" benefit from the NSA spying? The U.S. gave up $35Bn (and, frankly, specific companies had the brunt of it), but is there "savings" because of our security?
I'm not trying to get into a political discussion of "NSA is over-stepping its bounds." I also realize that the "savings" is entirely implicit. But I do wonder if there are some other, immeasurable, benefits of the agency.
First off I am not defending the NSA's actions; I am just trying to give an honest answer to this question.
Since the stated goal is to protect America from attacks, looking at the financial costs of the 9/11 attacks is a good way to find the costs on the other side of the argument. According to the New York Times, the successful attacks on the World Trade Center had an immediate economic cost of $178 billion. This includes $24 billion for the value of life lost, using similar actuarial tables that insurance companies or wrongful death lawsuits would.
The $35 billion figure is over a 4 year period, so thats about $9 billion per year. With this reasoning, if the NSA PRISM program could prevent one 9/11 scale attack every 20 years, it could be argued that it is worth it. This does not count the actual cost of running the NSA operation though, but that allegedly only cost about $20 million so it barely factors in.
If you accept the argument that war is inevitable when the US is attacked like we were on 9/11, then the total cost of 9/11 could be closer to $3 trillion. If American was safe enough because of NSA spying that it didn't "need" to fight foreign wars, that would be a huge economic cost saver.
This obviously does not factor in the cost of our loss of freedom, but I am trying to play devil's advocate here.
"Protect your People" That's rich. Since when is Germany facing a problem of not protecting their people? I don't see anyone even threatening to invade German territory.
Perhaps I don't know my history very well, but isn't the US the only reason Germany isn't part of Russia right now? The Berlin Wall was in Germany, right?
The Ukraine is probably the best example right now of what happens when you don't have a large enough army or at least powerful allies. If the Ukraine was part of NATO there would be no warmongering in that country right now.
I am suggesting that getting an art history degree should not cost hundreds of thousands of dollars. And I do not fault anyone for trying to get an education.
Getting an art history degree does not cost hundreds of thousands of dollars. Certain art history degrees do, but that doesn't mean society should pay for any expensive educational endeavor a 20 year old wants to try. Education is an investment, even if your only planned returns are self improvement. And if that is the case, it should be treated as more of a vacation which you should make sure you can afford before you venture off.
Regardless of what has happened to social mobility in the last 30 years, it hasn't affected the author of this article because he is 57 years old. He went to college in the 80's when college was not nearly as expensive. I went to college at the turn of the century and even then it was cheap enough you could pay over half of your college expenses by working part time at minimum wage.
This guy is simply a sociopathic asshole who is just being provocative to get page views. He stopped paying his bills because he is an entitled prick, not because of the federal loan apparatus he is complaining about in the article. I have real sympathy for the problems younger millenials are having because of the rising price of college, and it is shameful for this author to exploit them like this.
Which part of the brain holds your conscious self? There is no scientific explanation for the phenomenon of consciousness - no theory about how it arises, not even a definition of what qualifies. You cannot transfer consciousness without know what it is and how it works.
Obviously science has not progressed far enough to know how to model the human brain in a computer, or else we would probably be doing it already. So I guess I concede that we can't transfer the human brain yet, but I never said we could. But it is silly to believe we won't figure this out eventually. I would be surprised if it takes us 50 years.
I was only responding to the idea that if you transfer your brain to another medium, the old you dies. This is potentially true, but very unlikely. People can lose large portions of their brain without dying, and if those portions were replaced with synthetic computing devices I don't think anyone would think the old them has died.
Worse case scenario would probably be like what happens to someone suffering from multiple sclerosis (MS), where their personality changes to the point where they seem like a new person. In this case they would likely become more capable instead of less (like an MS patient), but their personality is still likely to change considerably.
The thing everyone always ignores is that no matter what, eventually your brain dies. Whether in you or after you've been uploaded to a computer or another brain or what have you. And when that happens *THAT* you is dead. *YOU* still experience the pain of death. YOU still cease to exist. There is something out there with your memories and thoughts, but they are not you any more than a photo album or journal is you.
This is very easily solved as a concept, although the implementation will obviously be insanely difficult. As another poster mentioned, your brain cells are constantly dying already. You still feel like you probably because it happens so gradually. So the answer to replacing your brain is the same; do it gradually. Conceptually you would be hooking your brain to a helmet filled with electronics that slowly replace your brain functions. At the end of the process your brain is completely electronic and you are still you. This is the theory anyway.
If you consider this scenario to be the same as you experiencing death, then you have already died perhaps hundreds of times in your lifetime so far.
Re:Basically a manager's job is to make other peop
on
On Managing Developers
·
· Score: 2
...He has some decent starting points. For example, Basically a manager's job is to make other people more productive...
Well, if the summary starts off on such a wrong assumption, it can only get worse from there.
imo, one of a software development manager's jobs is to create an environment that allows the software developers to do their jobs. If the manager has to "make" them do their jobs or be more productive, then the wrong people are in the software developer jobs.
I agree the article starts off with a very poor assertion about the most important role of a manager, but I don't necessarily agree with your either. I like your change to what the article says, but I still don't think it is the most important part of a manager's job.
IMHO, a manager's job is to ensure their projects are a success. Regardless of which developer or business analysts messes up, the ultimate responsibility always lies on the manager. Many employees don't realize this because they never witness their boss getting yelled at by his boss, but when projects miss their due dates the developers are not the only ones in trouble. Developers probably have a more silo-ed view of the whole project so they can legitimately blame failures on factors outside of their control, but a manager can rarely do that. The buck stops with him.
Bad managers micromanage because they are afraid the job won't be done right and they know their ass in on the line. Good managers find a way of trusting but validating their senior level resources.
from a distance. Say "Make this do this" and stand back, let the kids do their magic (hell, being a manager doesn't require technical knowledge) and look forward to the result. Micromanaging is for kindergarten teachers. You have a team of devs because they know how to code and you probably don't, keep thy nose in thine own trough.
While I like the overall message here of not micromanaging, I think an important distinction needs to be made between telling developers how to do their work and constantly validating that developers are doing their work. The latter is counterproductive but the former is absolutely necessary. Junior level employees often don't understand one of the major differences between managers and non-managers: in the eyes of executives your managers are ultimately responsible for the project's success or failure, regardless of who actually messed up. So they absolutely cannot just sit back and wait for developers to be done with their work and hope for the best.
While I agree with this article's advice that a good manager will find ways to take grunt work away from developers, that is such a minor part of their role as manager. The most important, and most difficult, part of being a manager is finding a way of constantly validating the project's progress without being a major hindrance to the developers. There is no easy answer to this that I can give in a Slashdot post, but it is vital to the success of any project.
As a senior level developer, one task I take upon myself is ensuring that our non-technical personnel have a method of validating our work throughout the project. I prefer having these milestones at least every other week. This doesn't have to mean you create a releasable product every other week like some Agile methodologies suggest, but you should at least find a way that your project / product managers, business analysts, and other key stakeholders can validate the project is on track.
When I join a new company (like I just did four months ago) I am very clear with my managers that I will take on this responsibility, but it means they are expected to trust me to do my job as long as they are happy with these validations. So far it has been working well for both me and my employers over the years, although that only comes from experience with two employers since I became a senior level resource.
Slashdot Mirror
If your code ever has the expression X / X and X is capable of being 0, you have a bug in your algorithm.
If you expect 0/0 = NaN, then it's not necessarily a bug.
You are correct there. A more accurate statement on my part should have been:
If your code ever has the expression X / X and X is capable of being 0, if your code does not treat this as an error and handle it properly then you have a bug in your algorithm.
Treating 0/0 as NaN is an example of handling the error properly, considering you are no longer treating the result as a valid number.
Let me rewrite that as:
Rule 1: 0*X (where X = 1/0)
Rule 2: X/X (where X = 0)
I've had this argument often with APL fans who don't get that as obvious as 0/0 = 1 may sound, it's not mathematically sound.
If your code ever has the expression X / X and X is capable of being 0, you have a bug in your algorithm. It really is as simple as that.
We had five bucks, and we decided to divide it up. After we were done dividing, the money was all gone. That is dividing by zero. It was divided into zero pieces.
Dividing something into zero pieces is not the same as dividing by zero. If you have no money left after dividing five bucks to five people, then the next attempt at dividing would be 0 / 5, not 0 / 0.
Programmers who primarily deal with integers and money units find it natural to have anything that doesn't make sense to default to zero.
Plenty of lazy programmers or those forced into using frameworks developed by other lazy programmers may be used to this. Perhaps even programmers forced to use primitive languages with no other options are used to it. But anyone else who knows what they are doing do not do this. This is similar to all of those developers who use NULL as if it means 0. Just because a lot of people do it doesn't make it any less idiotic.
So in the end the answer is to use high level languages, and use money objects with sensible defaults instead of floats and ints for money values.
Even low level languages have the capability of handling money correctly. It just requires more effort by the programmer, just like anything else done in a low level language. Considering all high level languages eventually have their code compiled down to the lowest level language understood by the CPU, low level languages are certainly capable of any calculation that a high level language is.
If you don't have specific reasons why it is impossible
I've made that clear already. Did you miss it?
There are no specifics in that post. Just some vague mention of semantic vs syntactic content and a claim that it is so obvious you don't even need to explain it.
All I can make from your semantic/syntactic analogy is that the physiology of a human brain and its computational ability may equate to syntactic content, but with no inherent semantic meaning. The consciousness that is created by these computations is what provides the semantic meaning. aka I think therefore I am.
But there is no absolute reason why a manufactured machine with similar computational ability as the human brain couldn't also create a consciousness that could give itself the semantic meaning you are referencing.
You can never state that something is not sufficient just because no one has figured it out yet.
That it's logically impossible is reason enough! Consider for a moment a simple example: I claim that it is impossible to clear 5 lines simultaneously in a game of Tetris. Would you say that claim is nonsense and it's only a matter of time before someone figures it out? Of course not. You can clearly demonstrate that it is an impossibility. The same is true for computational approaches to AGI -- they are logically impossible.
Why do you believe in such silly nonsense when it's clear that those beliefs are pure fantasy?
You are correct that you can clearly demonstrate that clearing 5 lines simultaneously in Tetris is impossible. But that is because you can be very specific as to why: the longest piece is only 4 lines long. Notice I didn't say something vague like: it is logically impossible to clear 5 lines. I was incredibly specific.
If your answer to why computational approaches to AGI is logically impossible are not also as specific, then you are just spouting nonsense. Just saying it is impossible is not an acceptable answer, and neither are other deflections such as calling it silly nonsense or pure fantasy.
If you don't have specific reasons why it is impossible, which show a thorough understanding of why neuron interaction cannot be replicated by a Turing complete computation system, or why simulating neuron interaction is not enough to create AGI, then just stop replying.
What else other than computation is required?
I have no idea. Neither does anyone else. That doesn't change the fact that computation alone has been show to be insufficient. That's pretty well established.
You can never state that something is not sufficient just because no one has figured it out yet. Most likely 100% of all significant technological advances were preceded by people who tried and failed. Often the people who finally figure it out were among the people who had failed multiple times in the past.
To say something isn't possible you have to know a great deal about why it is not possible. For instance we know that silicon transistors cannot shrink indefinitely because of our knowledge of physical limitations such as the width of a silicon atom. We don't say that transistors cannot be 5 nm wide just because all attempts at making 5 nm wide transistors have failed; that would be stupid.
It could not possibly be established that computation alone cannot create general AI because we have not even approached computers with enough transistors to match the 100 trillion synapses in an average adult. Even once we reach that milestone it will take a large amount of research before people could claim it simply isn't possible, and even then there would be a good chance they are still wrong.
Realistically the only way to say it is impossible to create general AI with computation alone is to actually create a general AI by other means. Only then could we possibly understand the process enough to know for sure it couldn't be done another way. Although even in that case there is no guarantee someone won't come up with a more elegant and/or efficient way in the future which only uses computation.
Here is an update to this ruling found in another article:
Update: Uber pointed out that the ruling only applies to one driver. “Reuters’ original headline was not accurate. The California Labor Commission’s ruling is non-binding and applies to a single driver,” a spokesperson said. “Indeed it is contrary to a previous ruling by the same commission, which concluded in 2012 that the driver ‘performed services as an independent contractor, and not as a bona fide employee.’ Five other states have also come to the same conclusion. It’s important to remember that the number one reason drivers choose to use Uber is because they have complete flexibility and control. The majority of them can and do choose to earn their living from multiple sources, including other ride sharing companies.”
Computation alone is insufficient.
What else other than computation is required? What is it you think your brain's neurons are doing that cannot be replicated?
Maximum Homerdrive
But in all seriousness, you still need someone there even on the freeway. Although, maybe three trucks can be in a train/convoy. But someone has to be there in case something goes wrong. Wheel falls apart, something comes unstrapped, etc. Monitoring it and having AAA come I don't think is as feasible as paying someone to sit there and deal with it when it happens.
As long as the truck can pull over if there is a problem, which it would need to even in your three trucks in a convoy solution, then this problem is easy to solve. You even mention the solution in your post: AAA. A commercial trucking version of AAA would have drivers to come fix any of these failures and get the trucks to their destinations. Who knows how many of these truckers they would need when this technology becomes ubiquitous, but its likely to be a very small percentage of the current trucking industry.
And struggling to maintain an outdated system in some kind of virtual environment isn't too risky?
Who said anything about upgrades/maintenance? Maybe he has to build it once, certify it once, deploy once on an embedded system.
If the job requirement says "25 years" then that's what he has to do. It wouldn't even be an unusual specification for military.
The only reason to need a consistent development environment is for upgrades and maintenance. If they never need to debug errors or add features then there is no reason to have a development environment at all. Just pass around binary executables.
Compilers improve their optimization algorithms over time. I've personally helped debug issues due to this on a number of occasions.
I assumed when the AC said new compilers alter the output of their algorithms, he meant the actual output of the functions, RPCs, services, etc. Optimization algorithms of the compilers shouldn't affect this unless they are broken. Or at least that would be my assumption. I would love clarification on how these compiler optimizations can cause different outputs from someone who has done this professionally though.
It does make sense that changes to math libraries would alter output because of differences in precision. Any other libraries, such as those which handle concurrency, would also open up possibilities for different behavior. I just don't see how compiler optimizations should make any difference.
While the summary is trying to make this some kind of huge rebuff of the President by democrats in Congress, the only serious problem with this bill is it was for too long of a time period. Obama is only in office for another 18 months, and this fast track authority would have extended years after he is gone. This vote had almost nothing to do with democrats not trusting Obama; it was them not trusting the unknown President who will take his mantle a couple years from now.
Technology and having it available does not imply that we have to use it.
At least so far I don't see mushroom clouds popping up left and right. And don't tell me we couldn't do that!
You are quite correct that not all technology will be used just because it exists. A better statement would be that all technologies which can be used to improve someone's utility will be used. Using nuclear weapons does not benefit a country when their opponents have them or if the political fallout would be too severe. And there are plenty of conventional weapons which are just as effective for most warfare requirements.
None of this has anything to do with consumer and corporate face recognition software. Augmented reality devices and consumer data mining are simply too useful to hold back. The entire lobbying industry would be a disgrace if they couldn't stop these regulations from impeding this progress because they will certainly be well funded.
But at least we can stop such sensitive data to be shared and processed en masse by both corporations and governments alike, by outlawing the practise.
An interesting side effect of this would be to give even more advantages to large companies. If this technology actually provides value, and small companies cannot buy this data from third parties, then only companies large enough to amass their own huge data sets have this advantage.
I am almost convinced the best route is to force large companies to open up their data (for a fee), similarly to how we forced certain utility providers to rent out their physical infrastructure. This lets a local grocer have the same data Walmart has. It also allows entrepreneurial AI companies to have access to the huge data sets companies like Facebook and Microsoft have access to.
Like encryption, those high powered IR LEDs to blind cameras, only attract attention -- until everyone is doing it.
Once encryption is outlawed, only outlaws will use encryption. Similarly wearing IR LEDs wired into your hat.
If this ever became a trend, the device makers would just start IR cut filtering their cameras.
If Software developers made $250-500k a year, I guarantee there would be no shortage of CS graduates.
Of course there would be no shortage, because 80% of current United States CS-related jobs would be off-shored to countries without such outrageous salaries. If you inflate wages above what the market will bear (taking into account the international market), then any industry will suffer the same fate the US auto industry suffered last century.
Saying that NSA surveillance cannot prevent a 9/11 style attack is silly, although saying it most likely will not is far more accurate (IMHO). The New America Foundation think tank published a report claiming that of 225 investigations performed since 9/11, only 1.8% of them were initiated based on NSA bulk metadata surveillance and 4.4% of the cases were assisted by the NSA spying program. These are obviously low figures so by shear odds it is likely the next huge terror plot would not be caught by NSA metadata spying.
But considering PRISM allegedly cost $20 million of a total US intelligence budget of around $75 billion annually over the past few years, the NSA program could still be considered a bargain. Although probably not once you factor in $35 billion in lost GDP because of a new found lack of trust of USA companies abroad.
And once again, I was only playing devil's advocate and trying to explain the other side of the argument. I personally believe the costs to personal freedoms dwarf the benefits of all enhanced security related programs since 9/11.
Honest question. What is the "theoretical" benefit from the NSA spying? The U.S. gave up $35Bn (and, frankly, specific companies had the brunt of it), but is there "savings" because of our security?
I'm not trying to get into a political discussion of "NSA is over-stepping its bounds." I also realize that the "savings" is entirely implicit. But I do wonder if there are some other, immeasurable, benefits of the agency.
First off I am not defending the NSA's actions; I am just trying to give an honest answer to this question.
Since the stated goal is to protect America from attacks, looking at the financial costs of the 9/11 attacks is a good way to find the costs on the other side of the argument. According to the New York Times, the successful attacks on the World Trade Center had an immediate economic cost of $178 billion. This includes $24 billion for the value of life lost, using similar actuarial tables that insurance companies or wrongful death lawsuits would.
The $35 billion figure is over a 4 year period, so thats about $9 billion per year. With this reasoning, if the NSA PRISM program could prevent one 9/11 scale attack every 20 years, it could be argued that it is worth it. This does not count the actual cost of running the NSA operation though, but that allegedly only cost about $20 million so it barely factors in.
If you accept the argument that war is inevitable when the US is attacked like we were on 9/11, then the total cost of 9/11 could be closer to $3 trillion. If American was safe enough because of NSA spying that it didn't "need" to fight foreign wars, that would be a huge economic cost saver.
This obviously does not factor in the cost of our loss of freedom, but I am trying to play devil's advocate here.
"Protect your People" That's rich. Since when is Germany facing a problem of not protecting their people? I don't see anyone even threatening to invade German territory.
Perhaps I don't know my history very well, but isn't the US the only reason Germany isn't part of Russia right now? The Berlin Wall was in Germany, right?
The Ukraine is probably the best example right now of what happens when you don't have a large enough army or at least powerful allies. If the Ukraine was part of NATO there would be no warmongering in that country right now.
I am suggesting that getting an art history degree should not cost hundreds of thousands of dollars. And I do not fault anyone for trying to get an education.
Getting an art history degree does not cost hundreds of thousands of dollars. Certain art history degrees do, but that doesn't mean society should pay for any expensive educational endeavor a 20 year old wants to try. Education is an investment, even if your only planned returns are self improvement. And if that is the case, it should be treated as more of a vacation which you should make sure you can afford before you venture off.
Regardless of what has happened to social mobility in the last 30 years, it hasn't affected the author of this article because he is 57 years old. He went to college in the 80's when college was not nearly as expensive. I went to college at the turn of the century and even then it was cheap enough you could pay over half of your college expenses by working part time at minimum wage.
This guy is simply a sociopathic asshole who is just being provocative to get page views. He stopped paying his bills because he is an entitled prick, not because of the federal loan apparatus he is complaining about in the article. I have real sympathy for the problems younger millenials are having because of the rising price of college, and it is shameful for this author to exploit them like this.
Which part of the brain holds your conscious self?
There is no scientific explanation for the phenomenon of consciousness - no theory about how it arises, not even a definition of what qualifies.
You cannot transfer consciousness without know what it is and how it works.
Obviously science has not progressed far enough to know how to model the human brain in a computer, or else we would probably be doing it already. So I guess I concede that we can't transfer the human brain yet, but I never said we could. But it is silly to believe we won't figure this out eventually. I would be surprised if it takes us 50 years.
I was only responding to the idea that if you transfer your brain to another medium, the old you dies. This is potentially true, but very unlikely. People can lose large portions of their brain without dying, and if those portions were replaced with synthetic computing devices I don't think anyone would think the old them has died.
Worse case scenario would probably be like what happens to someone suffering from multiple sclerosis (MS), where their personality changes to the point where they seem like a new person. In this case they would likely become more capable instead of less (like an MS patient), but their personality is still likely to change considerably.
The thing everyone always ignores is that no matter what, eventually your brain dies. Whether in you or after you've been uploaded to a computer or another brain or what have you. And when that happens *THAT* you is dead. *YOU* still experience the pain of death. YOU still cease to exist. There is something out there with your memories and thoughts, but they are not you any more than a photo album or journal is you.
This is very easily solved as a concept, although the implementation will obviously be insanely difficult. As another poster mentioned, your brain cells are constantly dying already. You still feel like you probably because it happens so gradually. So the answer to replacing your brain is the same; do it gradually. Conceptually you would be hooking your brain to a helmet filled with electronics that slowly replace your brain functions. At the end of the process your brain is completely electronic and you are still you. This is the theory anyway.
If you consider this scenario to be the same as you experiencing death, then you have already died perhaps hundreds of times in your lifetime so far.
...He has some decent starting points. For example, Basically a manager's job is to make other people more productive...
Well, if the summary starts off on such a wrong assumption, it can only get worse from there.
imo, one of a software development manager's jobs is to create an environment that allows the software developers to do their jobs. If the manager has to "make" them do their jobs or be more productive, then the wrong people are in the software developer jobs.
I agree the article starts off with a very poor assertion about the most important role of a manager, but I don't necessarily agree with your either. I like your change to what the article says, but I still don't think it is the most important part of a manager's job.
IMHO, a manager's job is to ensure their projects are a success. Regardless of which developer or business analysts messes up, the ultimate responsibility always lies on the manager. Many employees don't realize this because they never witness their boss getting yelled at by his boss, but when projects miss their due dates the developers are not the only ones in trouble. Developers probably have a more silo-ed view of the whole project so they can legitimately blame failures on factors outside of their control, but a manager can rarely do that. The buck stops with him.
Bad managers micromanage because they are afraid the job won't be done right and they know their ass in on the line. Good managers find a way of trusting but validating their senior level resources.
from a distance. Say "Make this do this" and stand back, let the kids do their magic (hell, being a manager doesn't require technical knowledge) and look forward to the result. Micromanaging is for kindergarten teachers. You have a team of devs because they know how to code and you probably don't, keep thy nose in thine own trough.
While I like the overall message here of not micromanaging, I think an important distinction needs to be made between telling developers how to do their work and constantly validating that developers are doing their work. The latter is counterproductive but the former is absolutely necessary. Junior level employees often don't understand one of the major differences between managers and non-managers: in the eyes of executives your managers are ultimately responsible for the project's success or failure, regardless of who actually messed up. So they absolutely cannot just sit back and wait for developers to be done with their work and hope for the best.
While I agree with this article's advice that a good manager will find ways to take grunt work away from developers, that is such a minor part of their role as manager. The most important, and most difficult, part of being a manager is finding a way of constantly validating the project's progress without being a major hindrance to the developers. There is no easy answer to this that I can give in a Slashdot post, but it is vital to the success of any project.
As a senior level developer, one task I take upon myself is ensuring that our non-technical personnel have a method of validating our work throughout the project. I prefer having these milestones at least every other week. This doesn't have to mean you create a releasable product every other week like some Agile methodologies suggest, but you should at least find a way that your project / product managers, business analysts, and other key stakeholders can validate the project is on track.
When I join a new company (like I just did four months ago) I am very clear with my managers that I will take on this responsibility, but it means they are expected to trust me to do my job as long as they are happy with these validations. So far it has been working well for both me and my employers over the years, although that only comes from experience with two employers since I became a senior level resource.