I think you missed the point of the question. The question is not about how to scale experience up/out. Scaling is fairly well understood. The question is how do you get a computer to experience anything in the first place.
Learning without forgetting is possible if, for example, you reconstruct the network, preserving the old one (and this can be optimized so the entire network doesn't have to be duplicated.)
But I'm curious why you think a mind is necessarily a neural network. Are you saying there is no other possible way to construct a mind? As far as I can tell, there are lots of other designs, many of them far superior to neural networks, especially for such basic things as representing knowledge.
They look on the surface like.Net based solutions, but the.net components are thin wrappers around COM. If anything, Microsoft is moving back toward native and away from.net (although they seem admittedly schizophrenic about it.) But that doesn't matter, because one thing we can be sure of is that backward compatibility with COM is not going away. Actually.NET is just the new COM anyway (In fact it started out being called COM 3.0... then they dropped the name).
As for adding ActiveX support to Webkit, well, your idea of trivial is different from mine. But lets say they did it. As someone pointed out elsewhere in this thread, there are hundreds of interfaces involved. Implementing them in a way that was backward compatible with existing COM would just tie their fork of Webkit to Windows, for what exactly?
[...] an engine that does not serve a competitive purpose anymore
Trident literally makes Microsoft NO money [...]
Both false.
Internet explorer does many things in the Windows/Office universe that no other browser does. Those things make Microsoft money by driving sales of Windows and Office and many other pieces of the Microsoft ecosystem (e.g. Sharepoint, SQL Server, etc.).
If all your desktops are Windows with Office and IE, you can develop intranet applications that use Office and can make direct calls to Win32. Yes this totally ties your application to Windows and Office, but many businesses are fine with that, even like it that way.
Active-X may be a security disaster on the internet, but in a locked down corporate intranet environment, you can easily do powerful things, like have a web page that embeds a live excel spreadsheet (the real excel, not a bloated, slow, feature-deprived javascript 'spreadsheet') displaying editable data from a database or web service. Click a link to open in excel, still editable, still connected to the server. You can do that kind of thing with very very little code, but only if you can assume you have Windows and Office on the client, and it only works in IE.
That is one of the competitive purposes of IE, and one of the ways they make money from it.
If Microsoft gets rid of the "Win32 cruft dating back to the 80s and 90s", then there will be no reason for anyone to choose Windows over any other operating system.
There is some truth to this, but my feeling is that as long as Microsoft's own desktop software is Windows only*, that will be enough to keep the business desktop on Windows. But on top of that, you can count on ISVs producing RT versions of their software, but many still don't have much incentive to port them to anything else. Businesses want to standardize the desktop, even if it causes them some pain. That standard will continue to be Windows, because it is currently the standard, if for no other reason.
* Office, Outlook, Visual Studio, IE, and many others that most people never heard of but are common in business. The Mac version of Office lacks features used in business environments.
I've been using C for so long that I think I've lost objectivity. C is the first language I learned (other than line numbered basic.) In my mind, C is the language all other languages are judged against.
But if there's any truth to this (when did the TIOBE index become the official word?) it makes me wonder if it's not C itself that is making a comeback, but good old fashioned procedural style programming.
All these fancy new languages with their polymorphism, encapsulation, templates and functional features have lost their sparkle. Programmers are rediscovering that there isn't anything you can't do (even runtime polymorphism) with just functions, structs, arrays and pointers. It can be easier to understand, and although it may be more typing, it has the virtue that you know exactly what the compiler is going to do with it.
Personally, I think the whole issue has been overblown, but...
The fact that they are doing the right thing now doesn't matter to those who are worried that they could change their minds at any time. Asking for a public commitment to continue to do what they are doing and plan on doing has value for those people. Nothing douche-y about that at all.
You're right. There's no reason why Linux can't work perfectly with (and benefit from) secure boot.
There's also no reason why doing so should require less frequent kernel releases. It doesn't require anything from Torvalds. Whoever builds your kernel just needs to sign it, and the end user just needs a way to say whether or not they trust that signer (e.g. a way to add Canonical's keys to your firmware.)
(Compiling the kernel from sources would require support from the entire toolchain, but that's another issue. Thatwould require something from the kernel maintainers: they would have to sign the 'official' sources. But it would take much more than that...)
You have the right idea, but you're mistaken about the details.
A rootkit doesn't install a modified version of ps, it modifies the system calls that ps uses. That way the rootkit is able to hide its processes from any program that enumerates processes. (There's much more to it of course.)
That also makes it easier to defend against. There's no need to prevent the user from running whatever userland code they want. All you need to do is ensure that the kernel you are running is the one you THINK you're running (as you put it.) Once you've verified the kernel, then you can trust the kernel to verify userland software (if desired).
In fact, you don't have have to protect the entire kernel, just a small portion of it that is responsible for loading and verifying the rest.
That's exactly what Secure Boot does, and it is an idea that is long overdue.
In 20 years, there will still be general-purpose computers, but they'll be extremely expensive.
While I admire your extreme cynicism, you haven't been paying attention to hardware trends. General purpose computers will be expensive relative to the special purpose ones, which is to say they will be dirt cheap (and obscenely powerful by today's standards).
Until they make it illegal, someone will always be willing to manufacture general-purpose-do-what-you-want machines.
Those numbers aren't that hard to get, and they are pretty good estimates, not "guesses".
But you don't even need to know the manufacture price or markup to know Apple is making a fucking blizzard of cash on these devices.
Here are the first two paragraphs from Apple's own press release regarding their most recent quarterly results. Pretty much speaks for itself:
CUPERTINO, California—October 25, 2012—Apple today announced financial results for its fiscal 2012 fourth quarter ended September 29, 2012. The Company posted quarterly revenue of $36.0 billion and quarterly net profit of $8.2 billion, or $8.67 per diluted share. These results compare to revenue of $28.3 billion and net profit of $6.6 billion, or $7.05 per diluted share, in the year-ago quarter. Gross margin was 40.0 percent compared to 40.3 percent in the year-ago quarter. International sales accounted for 60 percent of the quarter’s revenue.
The Company sold 26.9 million iPhones in the quarter, representing 58 percent unit growth over the year-ago quarter. Apple sold 14.0 million iPads during the quarter, a 26 percent unit increase over the year-ago quarter. The Company sold 4.9 million Macs during the quarter, a 1 percent unit increase over the year-ago quarter. Apple sold 5.3 million iPods, a 19 percent unit decline from the year-ago quarter.
Wow, I feel for you. If QA is not testing against the same build the developers are using, they're doing it horribly wrong. Or did you mean QA is doing their own build for their testing tools? That I can understand.
You're a better programmer for assuming it's not a compiler bug and trying harder to figure out what you did wrong.
I've been programming professionally for over 20 years, mostly in C/C++ (MSVC, GCC, and recently CLang (and others back in the olden days)). I've seen maybe two serious compiler bugs in the past 10 years. They used to be common.
On the other hand, I can't count how many times I've seen coders insist there must be a compiler bug when after investigation, the compiler had done exactly what it should according to the standard (or according to the compiler vendor's documentation when the compiler intentionally deviated from the standard).
By "serious", I mean the compiler itself doesn't crash, issues no warnings or errors, but generates incorrect code. Maybe I've just been lucky. (Or maybe QA just never found them;-)
Oh, and btw, yes I realize you were joking (and I found it funny.)
This kind of defeatist moral nihilism wouldn't be so annoying if it was expressed with a little intellectual humility.
What, has scientific evidence of their moral wrongness been unearthed?
Maybe. You seem to think (with no uncertainty) that it's a foregone conclusion that such evidence is impossible. It isn't.
Mathematical proofs are for math. Science is about weight of evidence, not proof.
You seem to think the only two logical possibilities are moral nihilism or morality from religion. They aren't.
No, of course not. They know no such thing.
Again, so arrogant, and yet so ignorant.
This stuff has been debated for thousands of years, right up to the present day, in philosophy and science (yes, science.) But never mind that -- AC on Slashdot has it all figured out.
Educate yourself, or STFU. Here's a good place to start:
Moral relativism, spoken with such conviction! Of course this is a popular point of view, especially among geeks, but rarely do you see anyone willing to go so far as to state that "there is no such thing as an immoral act" out loud (even if they believe that.)
I congratulate you for that, even though you did it anonymously.
Normally I would state something like the following as an opinion, but your matter-of-fact confidence inspired me, so I'll do the same:
Moral relativism is utterly false. Moral questions do have objectively right or wrong answers, founded on empirical evidence. (Or to be a little more precise, the answers to these questions are no less objectively right or wrong than are the answers to 'ordinary science' questions.) The answers don't depend on culture or upbringing any more than answers to physics questions do.
The reason why we codify law is because we don't agree on the answers to moral questions. But we don't agree on the answers to scientific questions either. That doesn't mean the answers don't exist.
Slashdot Mirror
I think you missed the point of the question. The question is not about how to scale experience up/out. Scaling is fairly well understood. The question is how do you get a computer to experience anything in the first place.
Learning without forgetting is possible if, for example, you reconstruct the network, preserving the old one (and this can be optimized so the entire network doesn't have to be duplicated.)
But I'm curious why you think a mind is necessarily a neural network. Are you saying there is no other possible way to construct a mind? As far as I can tell, there are lots of other designs, many of them far superior to neural networks, especially for such basic things as representing knowledge.
They look on the surface like .Net based solutions, but the .net components are thin wrappers around COM. If anything, Microsoft is moving back toward native and away from .net (although they seem admittedly schizophrenic about it.) But that doesn't matter, because one thing we can be sure of is that backward compatibility with COM is not going away. Actually .NET is just the new COM anyway (In fact it started out being called COM 3.0... then they dropped the name).
As for adding ActiveX support to Webkit, well, your idea of trivial is different from mine. But lets say they did it. As someone pointed out elsewhere in this thread, there are hundreds of interfaces involved. Implementing them in a way that was backward compatible with existing COM would just tie their fork of Webkit to Windows, for what exactly?
[...] an engine that does not serve a competitive purpose anymore
Trident literally makes Microsoft NO money [...]
Both false.
Internet explorer does many things in the Windows/Office universe that no other browser does. Those things make Microsoft money by driving sales of Windows and Office and many other pieces of the Microsoft ecosystem (e.g. Sharepoint, SQL Server, etc.).
If all your desktops are Windows with Office and IE, you can develop intranet applications that use Office and can make direct calls to Win32. Yes this totally ties your application to Windows and Office, but many businesses are fine with that, even like it that way.
Active-X may be a security disaster on the internet, but in a locked down corporate intranet environment, you can easily do powerful things, like have a web page that embeds a live excel spreadsheet (the real excel, not a bloated, slow, feature-deprived javascript 'spreadsheet') displaying editable data from a database or web service. Click a link to open in excel, still editable, still connected to the server. You can do that kind of thing with very very little code, but only if you can assume you have Windows and Office on the client, and it only works in IE.
That is one of the competitive purposes of IE, and one of the ways they make money from it.
If it works in Chrome on Windows, it will work on Safari on the iPhone, without needing to test if it actually works on the iPhone.
What color is the sky in your world?
Where I work, if you see a programmer at 8:00 AM, it's because he's been there all night.
Where I work, if you see a programmer in the office at 8am, he's been there all night.
Indeed. I solved it in 2-3 minutes, spent 15 minutes refactoring it, then just put it back the way it was.
If Microsoft gets rid of the "Win32 cruft dating back to the 80s and 90s", then there will be no reason for anyone to choose Windows over any other operating system.
There is some truth to this, but my feeling is that as long as Microsoft's own desktop software is Windows only*, that will be enough to keep the business desktop on Windows. But on top of that, you can count on ISVs producing RT versions of their software, but many still don't have much incentive to port them to anything else. Businesses want to standardize the desktop, even if it causes them some pain. That standard will continue to be Windows, because it is currently the standard, if for no other reason.
* Office, Outlook, Visual Studio, IE, and many others that most people never heard of but are common in business. The Mac version of Office lacks features used in business environments.
Maybe you're like me.
I've been using C for so long that I think I've lost objectivity. C is the first language I learned (other than line numbered basic.) In my mind, C is the language all other languages are judged against.
But if there's any truth to this (when did the TIOBE index become the official word?) it makes me wonder if it's not C itself that is making a comeback, but good old fashioned procedural style programming.
All these fancy new languages with their polymorphism, encapsulation, templates and functional features have lost their sparkle. Programmers are rediscovering that there isn't anything you can't do (even runtime polymorphism) with just functions, structs, arrays and pointers. It can be easier to understand, and although it may be more typing, it has the virtue that you know exactly what the compiler is going to do with it.
I'm not in favor of forced sterilization, but at least the person would have other reasons to go on living.
But I must be missing something here, because shouldn't the question be:
Is it worth it to cure addiction if you utterly destroy everything that makes life worth living?
How could any rational person think this is a good idea?
I don't see why the MPAA/RIAA can't push to make it illegal to bypass arbitrary lock down schemes
They can. (Haven't they done so already?)
The question is will you continue to be able to get hardware that doesn't have those restrictions, at a reasonable price?
I think you will.
Personally, I think the whole issue has been overblown, but...
The fact that they are doing the right thing now doesn't matter to those who are worried that they could change their minds at any time. Asking for a public commitment to continue to do what they are doing and plan on doing has value for those people. Nothing douche-y about that at all.
You feeling alright? :)
You're right. There's no reason why Linux can't work perfectly with (and benefit from) secure boot.
There's also no reason why doing so should require less frequent kernel releases. It doesn't require anything from Torvalds. Whoever builds your kernel just needs to sign it, and the end user just needs a way to say whether or not they trust that signer (e.g. a way to add Canonical's keys to your firmware.)
(Compiling the kernel from sources would require support from the entire toolchain, but that's another issue. That would require something from the kernel maintainers: they would have to sign the 'official' sources. But it would take much more than that...)
You have the right idea, but you're mistaken about the details.
A rootkit doesn't install a modified version of ps, it modifies the system calls that ps uses. That way the rootkit is able to hide its processes from any program that enumerates processes. (There's much more to it of course.)
That also makes it easier to defend against. There's no need to prevent the user from running whatever userland code they want. All you need to do is ensure that the kernel you are running is the one you THINK you're running (as you put it.) Once you've verified the kernel, then you can trust the kernel to verify userland software (if desired).
In fact, you don't have have to protect the entire kernel, just a small portion of it that is responsible for loading and verifying the rest.
That's exactly what Secure Boot does, and it is an idea that is long overdue.
The ONLY issue is who controls the keys.
In 20 years, there will still be general-purpose computers, but they'll be extremely expensive.
While I admire your extreme cynicism, you haven't been paying attention to hardware trends. General purpose computers will be expensive relative to the special purpose ones, which is to say they will be dirt cheap (and obscenely powerful by today's standards) .
Until they make it illegal, someone will always be willing to manufacture general-purpose-do-what-you-want machines.
Those numbers aren't that hard to get, and they are pretty good estimates, not "guesses".
But you don't even need to know the manufacture price or markup to know Apple is making a fucking blizzard of cash on these devices.
Here are the first two paragraphs from Apple's own press release regarding their most recent quarterly results. Pretty much speaks for itself:
CUPERTINO, California—October 25, 2012—Apple today announced financial results for its fiscal 2012 fourth quarter ended September 29, 2012. The Company posted quarterly revenue of $36.0 billion and quarterly net profit of $8.2 billion, or $8.67 per diluted share. These results compare to revenue of $28.3 billion and net profit of $6.6 billion, or $7.05 per diluted share, in the year-ago quarter. Gross margin was 40.0 percent compared to 40.3 percent in the year-ago quarter. International sales accounted for 60 percent of the quarter’s revenue.
The Company sold 26.9 million iPhones in the quarter, representing 58 percent unit growth over the year-ago quarter. Apple sold 14.0 million iPads during the quarter, a 26 percent unit increase over the year-ago quarter. The Company sold 4.9 million Macs during the quarter, a 1 percent unit increase over the year-ago quarter. Apple sold 5.3 million iPods, a 19 percent unit decline from the year-ago quarter.
Negative. The Mac and *nix versions of IE were completely separate code bases, developed by completely separate teams.
Negative Kelvin does have meaning. I've read that article 6 times and I still don't know what it means, but apparently it means something.
So true. It's a dysfunctional love/hate relationship I have with the C++ standard. And just like most abusive relationships, I refuse to leave her. :)
I wish D would gain some momentum.
Wow, I feel for you. If QA is not testing against the same build the developers are using, they're doing it horribly wrong. Or did you mean QA is doing their own build for their testing tools? That I can understand.
You're a better programmer for assuming it's not a compiler bug and trying harder to figure out what you did wrong.
;-)
I've been programming professionally for over 20 years, mostly in C/C++ (MSVC, GCC, and recently CLang (and others back in the olden days)). I've seen maybe two serious compiler bugs in the past 10 years. They used to be common.
On the other hand, I can't count how many times I've seen coders insist there must be a compiler bug when after investigation, the compiler had done exactly what it should according to the standard (or according to the compiler vendor's documentation when the compiler intentionally deviated from the standard).
By "serious", I mean the compiler itself doesn't crash, issues no warnings or errors, but generates incorrect code. Maybe I've just been lucky. (Or maybe QA just never found them
Oh, and btw, yes I realize you were joking (and I found it funny.)
They didnt. they define "Open Source". Caps have a purpose, you know.
There are a handful of case-sensitive words in English. "Open Source" isn't one of them.
What, has scientific evidence of their moral wrongness been unearthed?
Maybe. You seem to think (with no uncertainty) that it's a foregone conclusion that such evidence is impossible. It isn't.
Mathematical proofs are for math. Science is about weight of evidence, not proof.
You seem to think the only two logical possibilities are moral nihilism or morality from religion. They aren't.
No, of course not. They know no such thing.
Again, so arrogant, and yet so ignorant.
This stuff has been debated for thousands of years, right up to the present day, in philosophy and science (yes, science.) But never mind that -- AC on Slashdot has it all figured out.
Educate yourself, or STFU. Here's a good place to start:
Science of Morality
Sam Harris: Science can answer moral questions?
Moral relativism, spoken with such conviction! Of course this is a popular point of view, especially among geeks, but rarely do you see anyone willing to go so far as to state that "there is no such thing as an immoral act" out loud (even if they believe that.)
I congratulate you for that, even though you did it anonymously.
Normally I would state something like the following as an opinion, but your matter-of-fact confidence inspired me, so I'll do the same:
Moral relativism is utterly false. Moral questions do have objectively right or wrong answers, founded on empirical evidence. (Or to be a little more precise, the answers to these questions are no less objectively right or wrong than are the answers to 'ordinary science' questions.) The answers don't depend on culture or upbringing any more than answers to physics questions do.
The reason why we codify law is because we don't agree on the answers to moral questions. But we don't agree on the answers to scientific questions either. That doesn't mean the answers don't exist.