Re:Artificial intelligence and intellectual proper
on
Marvin Minsky On AI
·
· Score: 1
I was going to make some comments on several parts of your post, but I think it's enough to contest your starting assumption:
Any AI would have the self same survival imperatives that we do.
Prove that, or at least give some reasons why it should be true. You're anthropormorphizing the AI, which is a very common mistake (especially in science fiction movies). You can read about this here:
Maybe you're right, but in that case, you also think that just starting the dominos (which clearly is done by an external force) is enough to make it non-turing complete?
So you don't believe brain emulation is possible? Because if it is, all the problems you said will go away.
Re:slightly off-topic - general post on AI
on
Marvin Minsky On AI
·
· Score: 1
You're assuming that everyone will want to live as what we today call a "human". I'm almost sure that if it's possible, some people will want to transfer themselves to entirely robotic brains and bodies which are easier to repair and upgrade than our biological bodies.
Re:Artificial intelligence and intellectual proper
on
Marvin Minsky On AI
·
· Score: 3, Insightful
Why would someone program a true AI which has no built-in goals?
I also don't understand how MS can be blamed for anything to do with the Playstation 3, considering it has only recently launched (not in europe yet though!).
I guess some of the blame can go to Microsoft and some evil, because they advertised Windows 95 to be far more then what it was, Windows Vista is one step closer to offering what Microsoft promised us in Windows 95.
Do you have any links about what Microsoft promised for Windows 95? I'd be really interested to read that:)
That's precisely why I said that improving the algorithms is more important than optimizing an implementation. No need to preach to the choir here, we all know that complexity is very important; that's not what I was attacking about your post.
If it really was due to crossing the date line, then yes, it is ridiculous. That's one of the things which should be tested very accurately. How many more date/time bugs do we need until we start taking that seriously?
Of course, this is all speculation, unless it has been proved that the problem really was that one...
Oops, my previous reply had a misspelling in the blockquote tag!
Linear improvements are completely uninteresting. Take algorithm X in Perl, then a version in optimized assembly, and there you go, a 10X improvement without having done anything groundbreaking.
That's the kind of thinking which leads to the current kind of bloated and slow programs. Of course a better algorithm is always the first step, but programmers should always try to optimize their bottleneck algorithms. Would you like it if your software was 10 times slower? What about 100 times slower?
In my opinion, it is ridiculous to say that linear improvements are irrelevant. If we believe the current predictions about CPUs (number of cores increasing as opposed to GHz), this is even more true, since the old excuse of "the hardware will get better in 6 months" will stop being true.
Linear improvements are completely uninteresting. Take algorithm X in Perl, then a version in optimized assembly, and there you go, a 10X improvement without having done anything groundbreaking. That's the kind of thinking which leads to the current kind of bloated and slow programs. Of course a better algorithm is always the first step, but programmers should always try to optimize their bottleneck algorithms. Would you like it if your software was 10 times slower? What about 100 times slower?
In my opinion, it is ridiculous to say that linear improvements are irrelevant. If we believe the current predictions about CPUs (number of cores increasing as opposed to GHz), this is even more true, since the old excuse of "the hardware will get better in 6 months" will stop being true.
"Dynamics" is not the name of a product. It is the name under which all of MS's business applications are known. For example Dynamics NAV is former Navision, Dynamics AX is former Axapta (both of those were Danish companies bought by Microsoft, and they're still developed in Denmark). Dynamics GP is former great plains (which you have mentioned), and there are some other products, you can read about it in wikipedia!
Err, that sentence started with "Besides", it doesn't mean that I agreed with everything the article said. I'd actually like to see some proof that most people are disabling UAC. If it's true, I'll stand corrected...
With my first sentence, I meant that the GP post was wrong in saying that UAC is the only security feature in Vista.
That's just part of the security, if you had RTFA you might have learned a few things. Besides, most people probably don't even care/know how to disable UAC, so I doubt that will be a big problem...
For one thing, IE7, at least on Vista, is no longer such a dangerous web browser. It may still be the buggiest, the most easily exploited, and the most often exploited browser in internet history, and probably will be forever, but it has become safer to use, despite its many shortcomings.
It's funny the way he uses "IE7" when he's apparently talking about a mixture of IE and IE7... As far as I know, IE7 doesn't have many security bugs known until now, and especially not on Vista due to protected mode... Three letters - F, U, D.
Of course, it only works if everyone stays out of the admin account as much as possible, and if everyone with an admin password knows better than to install a questionable program with admin privileges. And there's the catch: "Windows needs your permission to install this cleverly-disguised Trojan nifty program. Click Yes to get rooted continue."
So you see that, here again, MS's security strategy involves shifting responsibility to the user.
So how exactly could this be better? By preventing the user to install/run any applications downloaded from the internet? It seems that the author of the article never heard about Security vs Usability tradeoffs (however he mentions them somewhere else, which makes it even worse...).
And once UAC is disabled, all of its security enhancements are lost.
Oh my god, how surprising...
The old problems never go away: too many networking services enabled by default
Some references would be nice, as well as proving that DEP and address space randomization won't be enough to counter the threat present due to those services...
Data hygiene is still an absolute disaster on Windows. In fact, it's worse than it ever was in some ways, and that's very bad indeed. Browser traces still in the registry, heavy and complicated indexing to improve search, new locations where data is being stored. It all adds up to a privacy nightmare. Keeping a Vista box "clean" is going to be impossible for all but the most knowledgeable and fastidious users.
That might be true, however it has nothing to do with the remark that Vista won't be enough to make the internet have less malware, etc.
If it turns out to be true that future improvements on CPUs will be more on the number of cores than the raw GHz speed, maybe we'll see this situation getting better. I say this because not all applications are easily parallelizable, so developers who want to differentiate themselves from the competition in terms of performance will really have to optimize their algorithms...
Now the question is: will the market and developers give more value to more features and bloat, or performance?
Slashdot Mirror
Prove that, or at least give some reasons why it should be true. You're anthropormorphizing the AI, which is a very common mistake (especially in science fiction movies). You can read about this here:
http://www.singinst.org/ourresearch/publications/
Maybe you're right, but in that case, you also think that just starting the dominos (which clearly is done by an external force) is enough to make it non-turing complete?
You can download it here.
So you don't believe brain emulation is possible? Because if it is, all the problems you said will go away.
You're assuming that everyone will want to live as what we today call a "human". I'm almost sure that if it's possible, some people will want to transfer themselves to entirely robotic brains and bodies which are easier to repair and upgrade than our biological bodies.
Why would someone program a true AI which has no built-in goals?
You can get them up again, but of course you need energy for that. You also need energy for running an electronic computer.
Wait, you don't even know if the C language was used, yet you are sure that the problem would go away with another language? Oh my god...
Pedantic. Because Overgeneralizing doesn't start with a P.
I also don't understand how MS can be blamed for anything to do with the Playstation 3, considering it has only recently launched (not in europe yet though!).
Do you have any links about what Microsoft promised for Windows 95? I'd be really interested to read that
That's precisely why I said that improving the algorithms is more important than optimizing an implementation. No need to preach to the choir here, we all know that complexity is very important; that's not what I was attacking about your post.
Come on, how many more date/time bugs do they need before they start testing those properly?
Assuming it was a date/time problem of course...
That would be just as lame as a time zone bug, or perhaps even more so...
If it really was due to crossing the date line, then yes, it is ridiculous. That's one of the things which should be tested very accurately. How many more date/time bugs do we need until we start taking that seriously?
Of course, this is all speculation, unless it has been proved that the problem really was that one...
That's the kind of thinking which leads to the current kind of bloated and slow programs. Of course a better algorithm is always the first step, but programmers should always try to optimize their bottleneck algorithms. Would you like it if your software was 10 times slower? What about 100 times slower?
In my opinion, it is ridiculous to say that linear improvements are irrelevant. If we believe the current predictions about CPUs (number of cores increasing as opposed to GHz), this is even more true, since the old excuse of "the hardware will get better in 6 months" will stop being true.
Linear improvements are completely uninteresting. Take algorithm X in Perl, then a version in optimized assembly, and there you go, a 10X improvement without having done anything groundbreaking.
That's the kind of thinking which leads to the current kind of bloated and slow programs. Of course a better algorithm is always the first step, but programmers should always try to optimize their bottleneck algorithms. Would you like it if your software was 10 times slower? What about 100 times slower?
In my opinion, it is ridiculous to say that linear improvements are irrelevant. If we believe the current predictions about CPUs (number of cores increasing as opposed to GHz), this is even more true, since the old excuse of "the hardware will get better in 6 months" will stop being true.
"Dynamics" is not the name of a product. It is the name under which all of MS's business applications are known. For example Dynamics NAV is former Navision, Dynamics AX is former Axapta (both of those were Danish companies bought by Microsoft, and they're still developed in Denmark). Dynamics GP is former great plains (which you have mentioned), and there are some other products, you can read about it in wikipedia!
Err, that sentence started with "Besides", it doesn't mean that I agreed with everything the article said. I'd actually like to see some proof that most people are disabling UAC. If it's true, I'll stand corrected...
With my first sentence, I meant that the GP post was wrong in saying that UAC is the only security feature in Vista.
Do you know how to read a question?
I think it would be nice if it came from the Creationism Class
That's just part of the security, if you had RTFA you might have learned a few things. Besides, most people probably don't even care/know how to disable UAC, so I doubt that will be a big problem...
It's funny the way he uses "IE7" when he's apparently talking about a mixture of IE and IE7... As far as I know, IE7 doesn't have many security bugs known until now, and especially not on Vista due to protected mode... Three letters - F, U, D.
So how exactly could this be better? By preventing the user to install/run any applications downloaded from the internet? It seems that the author of the article never heard about Security vs Usability tradeoffs (however he mentions them somewhere else, which makes it even worse...).
Oh my god, how surprising...
Some references would be nice, as well as proving that DEP and address space randomization won't be enough to counter the threat present due to those services...
That might be true, however it has nothing to do with the remark that Vista won't be enough to make the internet have less malware, etc.
If it turns out to be true that future improvements on CPUs will be more on the number of cores than the raw GHz speed, maybe we'll see this situation getting better. I say this because not all applications are easily parallelizable, so developers who want to differentiate themselves from the competition in terms of performance will really have to optimize their algorithms...
Now the question is: will the market and developers give more value to more features and bloat, or performance?
I don't think he assumed anything, he specifically said "Just because aliens might exist doesn't mean they'd want to interact with us"...
:)
Anyway, we both get the point