Many people have worked on the area which they think AI is lacking in, but until we understand the brain (or even come up with a good definition of "intelligence") I don't think we'll get very far. But who knows, keep on looking!
It could also be the case that the wonders of the human brain are not so much a result of clever "design" (no pun intended) as it is of raw processing power and brute force.
Taken to its extreme, it could very well be the case that strong AI, at least to some degree, is a kind of emergent property of computational power. Obviously, a super-powerful super-computer wouldn't automatically become intelligent as soon as it boots up, but it could be the case that tricky AI problems turn out to be much simpler to solve given enough training data (the Internet?) and enough processing power. Take this for example; very impressive results from a fairly simple method (IMHO) fed with enough data and CPU-cycles.
Such algorithms are to simple to constitute examples of intelligence, you might say, and I would probably agree. But still, the images seem intelligently made. How can we be sure that what we perceive as intelligence in others isn't just all of their impressions and experiences filtered through some reasonably simple but massively parallel computational process? If that is the case, I think that more powerful computers will be the key that triggers the coming break-troughs in AI research.
(Disclaimer: IANANS (I Am Not A NeuroScientist), so please correct me if you are one and I am way off the mark.)
When your 18, go right ahead and make the 53 ch4R@ct3R password to lock your machine up, until then, accept the fact that you are the child and we are the parent, and you don't get root access or personal and private encryption, you ask the IT department (dad).
That's right, teach them early. You can't be trusted (even by your own parents) and it is only natural for the "authorities" to monitor what you're doing.
Seriously though, I'm not saying a 7 year old should be able to keep the parents out. But it's not like a 17 year old isn't going to keep secrets from you and have a private life anyway. And what kind of signals do you think you're sending by requiring full access to their computer?
Kids don't learn from what you tell them, they learn from your actions.
Now, if the consuming public provide a massive negative reaction to this behavior, then the shareholders would be rightfully demanding that Amazon and other companies not play the patent game. But we all know how thoughtful most consumers are.
Well, if we are talking about optimization, thoughtful customers maximizing their own situations individually would probably still find it suboptimal to buy from another supplier, if Amazon is the cheapest one. The kind of behaviour you ask for requires cooperation (from a game theoretical point of view, anyway).
Only as you combine a large number of similar files does the probability of correctly decoding the signatures of the components decrease. However by that time, you've added a large amount of noise to the audio file and it will probably sound bad anyway, so no-one will want to download it.
I'm not following you here. Wouldn't combining multiple noisy files rather reduce the noise? I mean, that is how you generally improve signal to noise ratio: collect more samples of the same data and combine them.
As far as I know, with nanokernel you usually mean a kernel in which even the scheduler not included. For example, you could have an OS running multiple schedulers at once. (Perhaps that is why you called your suggestion a "Scheduler Nanokernel"...?) See for example the Wikipedia article on nanokernels. Exokernels is a similar concept.
Firefox without favourites? Without history? [...] Most braindead idea I have heard all week.
Especially when you consider that the resource footprint of keeping a list of links in memory is utterly insignificant in this context. Talk about premature optimisation...
You are correct. However, reason in combination with a bias, such as Occam's razor, can get you somewhere if you already have assumed some basic moral rules and values.
Taking your example about killing neighbours; suppose there is a specific ethnically different neighbour called Jane Doe that I happen to care about and don't want to see killed. Suppose also that I dislike nepotism. I might then argue that, given that I think it is wrong to kill Jane Doe, it should also be wrong to kill any other ethnically different neighbour. Differentiating between them would introduce additional and unwarranted complexity and result in a less succinct system of values.
Of course, in a strict relativistic sense there is never anything preventing you from having completely different moral rules for each particular situation you are in (e.g. "Killing neighbours is morally right, unless their last name is Doe"). But such a system is quite useless, since it can never tell you how you should act, and it eliminates every possibility of even having discussion about morality.
In fact, the exact same things could be said about science, a field which we generally associate strongly with reason and reality, since its foundation, empiricism, in itself is an arbitrary principle. Reason does not give us science, yet it is usually said that science is grounded in reason.
Nontheless, to claim that either science or morality is straightforward is rather presumptuous, I would say.
Who would have thought it neccessary to catalog "a technique of allowing customers to make online purchases with a single click" as prior art?
I wonder, perhaps it would be possible to create a script generating all those trivial ideas. Obviously, it wouldn't be able to enumerate more complex "inventions", but at least we could clear off all of those "Use known technique A in conjunction with known technique B to accompilsh C". Recursion could be used to cover even more.
Taken to an extreme, we could create a search engine that automatically generates 1000 ideas or so based on the search term. But I suppose that would be pushing it a little.
Not if you did it the right way at the beginning. [...] "What are these stupid macros, why I can't just write unsigned int instead of that ugly looking DWORD."
Or you could do it the ISO standard way and write uint32_t. Unless you're running Visual Studio, of course, which doesn't provide the necessary headers.
Given that they're using my software in a device which they are locking down with cryptographic keys, and given that I don't want them to put my software in such a locked-down device, logically the correct thing to do is to license my software to them in a way that stops them from using it in such a device. This is a means appropriate to the end, is it not?
And you are certainly allowed to do that. However, unless your software is superb and unique, I doubt it will affect the course of events on a greater scale very much. To me, it is more of a moral standpoint than a pragmatic one, which is OK too of course.
Anyway, where if anywhere do you think [the fight] does belong?
Well, for starters, I'd like to see it in public debate and in the development of hardware standards.
Personally, I think that the impact the GPLv3 will have is primarily to attract attention to the issue.
"If you use my software, for free, to create a device which you then sell back to me at a steep price, I'd better bloody well be able to put any further modifications I make to that software onto said device.[...]"
To meta-paraphrase your paraphrase, that is essentially to say: "If you use my software, and put it on a device which you sell, that device must not be of kind X". A similar requirement would be: "You may not use my software in weapons", or "You may not use it in polluting industries" and so on.
While I find the cause to prevent vendors from locking down consumer hardware with cryptographic keys sound, I'm not yet convinced that the fight really belongs in a software license.
2. Start issuing SecurID tokens (or similar) to bank customers. This would take care of the simpler keyloggers and phishing attacks.
My bank uses cryptographic tokens, where you are required to "sign" any transaction by entering the same data (amounts and account numbers) into the token, which generates a confirmation code. Man-in-the-middle attacks (even if they should control my own computer) then become very difficult, since I have to enter all the numbers manually into the token. That is the only banking security model I have put any trust in so far.
But if you come along and *insert* ads on my pages and thus benefit from my work, I have no choice but to sue. That is copyright violation. Period. They are costing the content provider money.
There was actually a case in Sweden last year where the directors Claes Eirksson and Vilgot Sjöman successfully sued Sweden's largest commercial TV station TV4 after it had shown two of their films with interruptions for commercials. In the ruling the court concluded that the interruptions were an infringement of the moral right of the creators, since the station didn't have an express permission to insert them. I imagine a similar argument could be made for web sites.
And who claimed that "the public" tells the police what to do?
The public telling the police what to do is scary enough for me. Considering how quickly people often seem to pass judgement in situations where the only information they have is tabloid headlines, and how firmly they believe they are right. Don't get me wrong, the police must be under democratic control, but at least some degree of separation is needed, or else we end up with an ochlocracy.
Capitalism is a system of economic freedom, but "economic freedom" does not require that you adopt ruthless unethical business practices
You seem to suggest that there is a choice involved. Well, there is at the level of single companies, but it doesn't have any significance on a greater scale. Business practices evolve according to what is most efficient (i.e. profitable). If it is more profitable to be ruthless, the ruthless will survive and those who choose to be "ethical" will die out. It's no different from biological evolution, only faster. Nobody controls it.
Re:Of course you can talk about it at interviews
on
Google's Evil NDA
·
· Score: 1
If I were to hire you, and the first thing you did was to ignore some agreements you had with you former employer, how do you think that would affect my ability to trust you, being your new employer?
It may be that, despite the age of the universe, the conditions for intelligent life took a long time to come together.
But that would mean that we, by chance, happen to be born during the first wave of intelligent life, which seems very improbable if intelligent life will continue to develop throughout the universe. Unless, of course, the first intelligent life forms kill off or somehow prevent any other subsequent intelligent life, or if the universe for some other reason has only a narrow time window in which intelligent life can appear.
Re:7 centuries isn't feasible for humans
on
Interstellar Ark
·
· Score: 1
Yes. What I actually meant to say was "relativistic anti-matter rocket [...] would be constructed say 80 years after the ark had left". I mean, just building the ark would take decades, so obviously it isn't going to happen soon, if ever.
Re:7 centuries isn't feasible for humans
on
Interstellar Ark
·
· Score: 1
only if you're too mean to stop and pick up the ark in your fancy rocket on the way there.
There are two things that prevent you from doing that, even if you're kind. First, the anti-matter rocket is likely to be much smaller than the ark, so the best you could to would be to stop by and just say hello, basically. Secondly, the idea with the relativistic rocket is to accelerate half the journey and decelerate the other half. If you want to stop in the middle, you'd have to repeat that cycle twice, and since you then would spend less of the distance at near light speed it would take many times longer to reach the destination.
Re:7 centuries isn't feasible for humans
on
Interstellar Ark
·
· Score: 1
Also, imagine the irony if alternative two, the relativistic anti-matter rocket, would be constructed in say 80 years from now. When the ark finally arrives, they find that humans have lived there for 600 years already.
After RTFA I realised that:
a) 22 travel time was in the traveller's reference frame, so it would take longer from the arks perspective, and
b) the summary is incorrect -- 22 years is the time it would take to cross the entire galaxy.
Regardless, the anti-matter rocket would still overtake the ark easily, and arrive at Epsilon Eridani long before the ark. So my point still stands.
Re:7 centuries isn't feasible for humans
on
Interstellar Ark
·
· Score: 1
Besides, how would you select the crew
Actually, who would even want to go? Sure, arriving at Epsilon Eridani would be interesting, but unless we find a way to extend our life spans dramatically, you're basically condemning yourself, your children, grand children, great grand children etc. to live their entire lives on a space ship.
Also, imagine the irony if alternative two, the relativistic anti-matter rocket, would be constructed in say 80 years from now. When the ark finally arrives, they find that humans have lived there for 600 years already.
Ok, I realise a large portion of Slashdot readers are libertarians or similar, but why exactly was this moderated Flaimbait? It is a controversial suggestion, but not completely unthinkable. Especially since that is sort of how it works with library books in some countries already.
I see this as a board decision alone. While the community would have an uproar, the organization would survive. The vast majority of their "clients" would never realize the difference.
The problem is that with Wikipedia, the community is much of the organisation. While it would survive in some form, it would be severely decimated, losing a large portion of the most frequent and important contributors. The vast majority would perhaps not notice any difference immediately, but Wikipedia would have a tough time not deteriorating significantly over time.
It is curious how everyone seem to think they can just ignore ads, and still corporations find that it repays spending billions buying them.
Also, with ads you still pay for it all, through increased prices the products that are advertised. The only difference is that some of that money is wasted producing the ad (so you actually pay more), and that you have to be bothered watching it. (This is especially accentuated with TV commercials, which is why I have never understood how people can stand ad funded TV stations to such a degree.)
Corporate TV has: 24, House, Heroes, The Office, The Daily Show, Battlestar Galactica, Grey's Anatomy, Scrubs, The Wire, and lots of shows that used to be good like The Simpsons and Lost.
Yes, corporate TV has: entertainment, entertainment, entertainment, entertainment, entertainment, entertainment, entertainment, entertainment, entertainment, entertainment, and entertainment.
Slashdot Mirror
It could also be the case that the wonders of the human brain are not so much a result of clever "design" (no pun intended) as it is of raw processing power and brute force.
Taken to its extreme, it could very well be the case that strong AI, at least to some degree, is a kind of emergent property of computational power. Obviously, a super-powerful super-computer wouldn't automatically become intelligent as soon as it boots up, but it could be the case that tricky AI problems turn out to be much simpler to solve given enough training data (the Internet?) and enough processing power. Take this for example; very impressive results from a fairly simple method (IMHO) fed with enough data and CPU-cycles.
Such algorithms are to simple to constitute examples of intelligence, you might say, and I would probably agree. But still, the images seem intelligently made. How can we be sure that what we perceive as intelligence in others isn't just all of their impressions and experiences filtered through some reasonably simple but massively parallel computational process? If that is the case, I think that more powerful computers will be the key that triggers the coming break-troughs in AI research.
(Disclaimer: IANANS (I Am Not A NeuroScientist), so please correct me if you are one and I am way off the mark.)
That's right, teach them early. You can't be trusted (even by your own parents) and it is only natural for the "authorities" to monitor what you're doing.
Seriously though, I'm not saying a 7 year old should be able to keep the parents out. But it's not like a 17 year old isn't going to keep secrets from you and have a private life anyway. And what kind of signals do you think you're sending by requiring full access to their computer?
Kids don't learn from what you tell them, they learn from your actions.
Well, if we are talking about optimization, thoughtful customers maximizing their own situations individually would probably still find it suboptimal to buy from another supplier, if Amazon is the cheapest one. The kind of behaviour you ask for requires cooperation (from a game theoretical point of view, anyway).
I'm not following you here. Wouldn't combining multiple noisy files rather reduce the noise? I mean, that is how you generally improve signal to noise ratio: collect more samples of the same data and combine them.
As far as I know, with nanokernel you usually mean a kernel in which even the scheduler not included. For example, you could have an OS running multiple schedulers at once. (Perhaps that is why you called your suggestion a "Scheduler Nanokernel"...?) See for example the Wikipedia article on nanokernels. Exokernels is a similar concept.
Especially when you consider that the resource footprint of keeping a list of links in memory is utterly insignificant in this context. Talk about premature optimisation...
You are correct. However, reason in combination with a bias, such as Occam's razor, can get you somewhere if you already have assumed some basic moral rules and values.
Taking your example about killing neighbours; suppose there is a specific ethnically different neighbour called Jane Doe that I happen to care about and don't want to see killed. Suppose also that I dislike nepotism. I might then argue that, given that I think it is wrong to kill Jane Doe, it should also be wrong to kill any other ethnically different neighbour. Differentiating between them would introduce additional and unwarranted complexity and result in a less succinct system of values.
Of course, in a strict relativistic sense there is never anything preventing you from having completely different moral rules for each particular situation you are in (e.g. "Killing neighbours is morally right, unless their last name is Doe"). But such a system is quite useless, since it can never tell you how you should act, and it eliminates every possibility of even having discussion about morality.
In fact, the exact same things could be said about science, a field which we generally associate strongly with reason and reality, since its foundation, empiricism, in itself is an arbitrary principle. Reason does not give us science, yet it is usually said that science is grounded in reason.
Nontheless, to claim that either science or morality is straightforward is rather presumptuous, I would say.
I wonder, perhaps it would be possible to create a script generating all those trivial ideas. Obviously, it wouldn't be able to enumerate more complex "inventions", but at least we could clear off all of those "Use known technique A in conjunction with known technique B to accompilsh C". Recursion could be used to cover even more.
Taken to an extreme, we could create a search engine that automatically generates 1000 ideas or so based on the search term. But I suppose that would be pushing it a little.
Or you could do it the ISO standard way and write uint32_t. Unless you're running Visual Studio, of course, which doesn't provide the necessary headers.
And you are certainly allowed to do that. However, unless your software is superb and unique, I doubt it will affect the course of events on a greater scale very much. To me, it is more of a moral standpoint than a pragmatic one, which is OK too of course.
Well, for starters, I'd like to see it in public debate and in the development of hardware standards.
Personally, I think that the impact the GPLv3 will have is primarily to attract attention to the issue.
To meta-paraphrase your paraphrase, that is essentially to say: "If you use my software, and put it on a device which you sell, that device must not be of kind X". A similar requirement would be: "You may not use my software in weapons", or "You may not use it in polluting industries" and so on.
While I find the cause to prevent vendors from locking down consumer hardware with cryptographic keys sound, I'm not yet convinced that the fight really belongs in a software license.
My bank uses cryptographic tokens, where you are required to "sign" any transaction by entering the same data (amounts and account numbers) into the token, which generates a confirmation code. Man-in-the-middle attacks (even if they should control my own computer) then become very difficult, since I have to enter all the numbers manually into the token. That is the only banking security model I have put any trust in so far.
There was actually a case in Sweden last year where the directors Claes Eirksson and Vilgot Sjöman successfully sued Sweden's largest commercial TV station TV4 after it had shown two of their films with interruptions for commercials. In the ruling the court concluded that the interruptions were an infringement of the moral right of the creators, since the station didn't have an express permission to insert them. I imagine a similar argument could be made for web sites.
If I were to hire you, and the first thing you did was to ignore some agreements you had with you former employer, how do you think that would affect my ability to trust you, being your new employer?
But that would mean that we, by chance, happen to be born during the first wave of intelligent life, which seems very improbable if intelligent life will continue to develop throughout the universe. Unless, of course, the first intelligent life forms kill off or somehow prevent any other subsequent intelligent life, or if the universe for some other reason has only a narrow time window in which intelligent life can appear.
Yes. What I actually meant to say was "relativistic anti-matter rocket [...] would be constructed say 80 years after the ark had left". I mean, just building the ark would take decades, so obviously it isn't going to happen soon, if ever.
There are two things that prevent you from doing that, even if you're kind. First, the anti-matter rocket is likely to be much smaller than the ark, so the best you could to would be to stop by and just say hello, basically. Secondly, the idea with the relativistic rocket is to accelerate half the journey and decelerate the other half. If you want to stop in the middle, you'd have to repeat that cycle twice, and since you then would spend less of the distance at near light speed it would take many times longer to reach the destination.
After RTFA I realised that:
a) 22 travel time was in the traveller's reference frame, so it would take longer from the arks perspective, and
b) the summary is incorrect -- 22 years is the time it would take to cross the entire galaxy.
Regardless, the anti-matter rocket would still overtake the ark easily, and arrive at Epsilon Eridani long before the ark. So my point still stands.
Actually, who would even want to go? Sure, arriving at Epsilon Eridani would be interesting, but unless we find a way to extend our life spans dramatically, you're basically condemning yourself, your children, grand children, great grand children etc. to live their entire lives on a space ship.
Also, imagine the irony if alternative two, the relativistic anti-matter rocket, would be constructed in say 80 years from now. When the ark finally arrives, they find that humans have lived there for 600 years already.
Ok, I realise a large portion of Slashdot readers are libertarians or similar, but why exactly was this moderated Flaimbait? It is a controversial suggestion, but not completely unthinkable. Especially since that is sort of how it works with library books in some countries already.
The problem is that with Wikipedia, the community is much of the organisation. While it would survive in some form, it would be severely decimated, losing a large portion of the most frequent and important contributors. The vast majority would perhaps not notice any difference immediately, but Wikipedia would have a tough time not deteriorating significantly over time.
It is curious how everyone seem to think they can just ignore ads, and still corporations find that it repays spending billions buying them.
Also, with ads you still pay for it all, through increased prices the products that are advertised. The only difference is that some of that money is wasted producing the ad (so you actually pay more), and that you have to be bothered watching it. (This is especially accentuated with TV commercials, which is why I have never understood how people can stand ad funded TV stations to such a degree.)
Yes, corporate TV has: entertainment, entertainment, entertainment, entertainment, entertainment, entertainment, entertainment, entertainment, entertainment, entertainment, and entertainment.
Wait... is there a pattern here somewhere?