Re:slightly off-topic - general post on AI
on
Marvin Minsky On AI
·
· Score: 1
Unless the "use" you have in mind is having an expert pilot automating a fleet of drone air or space craft, or an expert at handling any other kind of equipment.
I'd think the bigger concern would be boredom on the part of the "recorded intelligence" programs. People have multiple interests, rather than being single-minded. What is a "pilot expert" program going to do when it has the urge for a beer?
That is a very good point. If there is no patent on the algorithm itself, then creating an implementation of the algorithm is perfectly legal and the only item that should be restricted by the DMCA are the keys used for decryption. Otherwise the DMCA has a much broader effect than originally intended, and makes it illegal for any form of encryption or decryption to be used other than that provided by a vendor.
In the current takedown scenario, it would be illegal to use a non-Sony decryptor for music encoded using ATRAC and Sony's software. It wouldn't just be illegal to crack any DRM, but illegal to use a non-endorsed player, and that flies in the face of so much precedence it's not even funny.
Mind you, I'd bet that some of the auto manufacturers tried the same thing with "jobber" parts at one point.
If you're talking about a larger business, there is a good chance custom Win32 applications are being run, and those are a problem. But if they're relatively new applications written with C#/.Net, you have a fair chance of porting them to run under Mono with little difficulty (presuming Mono runs on OSX.)
The basic Microsoft office automation applications are available for Macs. I know several people who use MS Office on the Mac to deal with people who run under Windows.
Common applications like Opera and Firefox already run under Linux. So does the Java/Eclipse/JBoss/OSS stack that many developers work with on data center/web service applications.
There is little the basic business desktop needs that isn't available on a Mac or on Linux, and Macs have some plug-and-play maintenance advantages that Linux doesn't -- namely you buy boxes pre-installed for a known price without adding a third-party installation/deployment company to the mix. Even better -- they work.
Don't forget you can run any X-Windows application on a Mac, including cross-compiled OSS applications and utilities. Those do port easily, and many have been made available. You don't have to rewrite all the interfaces to the OSX Aqua GUI APIs.
I don't recall any of Rick Stallman's lectures being about charity.
Oracle is trying to shift the blame from their software stack to IBM's before they've even deployed a box. In other words, if you have a problem and are running IBM software, Oracle wants IBM to foot the bill of researching the bug or issue.
Who is Oracle to dictate that problems are automatically some other vendor's fault instead of their own?
Of course if Windows actually implemented the POSIX thread APIs, it would be easier to port threaded code to and from Windows. There are workarounds and libraries for the lack of a full POSIX API stack under Windows (and maybe some flavours of Windows included the APIs), but the general issue is not whether people will write threaded code, but whether Microsoft will allow threaded code written to the published industry standards.
There is only one company responsible for a lack of industry standard APIs under Windows, and that is Microsoft.
"And Microsoft has no obligation to facilitate a competitor."
But VMWare only is a "competitor" because VMWare was making some money and Microsoft just couldn't let their cash vaccuum miss any potential revenue. Had Microsoft stuck to what they do "best", VMWare wouldn't be a competitor and that whole aspect of the argument would be moot.
If you look at the pictures of the floor tiling posted with the article, you'll see they laid out round patterns on a number of string-line intersections. Add a few measuring rods or ratios for positioning those string-lines, and patterns like that could be replicated by skilled tradesmen quite easily. You could even wrap-tie knots at the intersection points and carry the pattern around as a loose string net.
If you teach people to lay out the patterns, and how to fill in the spaces in between with tile pieces, then you have people laying out the pattern without them having the vaguest clue about the mathematics. All it took was a mathematician with a knack for explaining things each time the pattern was written down.
Were those real people posting the blog posts you saw, or shills? Never underestimate the power of fanbois to delude each other, no matter what the subject is.
I believe the second approach you're talking about is called "flipping" a company. Building it up for the sole purpose of selling it to the highest bidder.
In all fairness I wouldn't run any OS with less than 512MB RAM if I expect to run a word processor and a browser with a few active plugins. Even my creaky old Linux box has 512MB of RAM as well so that it doesn't thrash itself to death when I'm running more than one application. It's barely tolerable for surfing a website that is heavy with plugins and Flash scripts, but quite acceptable for most day-to-day work. Load times aren't even that bad -- a full reboot and login takes far less than five minutes, never mind the 10 minutes someone had claimed Win95 used to take.
I believe the people remembering 10 minute startup times for any OS since paper tape are remembering all their ad-on TSRs and similar programs starting up, not the OS. I've never waited much more than a few minutes for an OS to start on any box I've owned over the years.
When I see a slow startup, I presume it means an infected machine.
VCs tend to be smarter since the dot-bombs. The ones I've dealt with directly and indirectly have made a much greater push for business plans, existing customers, and the usual signs of life at a proposed business venture. They still sometimes chase technology buzzwords and the fantastical dream of "license revenue" rather than building the business themselves, but at least they're dealing with business ideas not dreams.
Having been burned by a recursive dependancy in an RPM stream more than once, I'm a big fan of the tarball extracted to/opt and doing a "make install".
Let's try basic physics. How much power does a 3.5W signal take to produce, for example? Now how many milliwatts does an LCD take in comparison? If you want bean counter references for the obvious, hit the search engines.
Personally I think it's just that the power consumption of cheap LCD's is pretty trivial compared to the milliamps used to send and receive RF signals. Transmission of voice or data is much more expensive than passive receipt. You burn more power in a five minute cell call than your display probably uses in 30 minutes or more.
I think Linus really cares about the quality of code and giving people flexibility, choice, and configurability whenever possible. The idea of "interface nazis" that try to dictate one style of interaction obviously gets his back up, and I think, for good reason.
What Linus did may seem argumentative to some, but I think what he did was prove that one can simply implement a solution in less time than it takes the current Gnome team or committee to argue about possible solutions. That doesn't detract from the Gnome team's work, but it does show that sometimes it's better to do and to prototype patches so that people can actually try an interface change instead of just arguing about it in theory.
If someone (or many someones) on the Gnome team take that as offensive, I think it's time that they took a look at their own methodology and accepted that no one outside the military and a very few regulated industries does "waterfall" development methodologies any more. They've been proven inefficient time and time again, and they bore volunteer programmers to tears.
The subpoena means that PJ would identify herself to the court and claim ownership of her postings in court. That allows them to be treated as expert testimony regarding the overall SCO proceedings.
Personally I'd think Novell would be the one wanting to have that expertise brought to bear, not SCO. But SCO hasn't exactly made a lot of rational moves throughout this debacle.
Darl probably just sticks his fingers in his ears and mutters a mantra whenever PJ and the Slashdot crew start talking about their farcical cases. Personally I don't believe anyone as deluded as Darl could ever have risen to such a position of power in any company, so he must be a figment of our imaginations.:p
Currently the best solutions for TSP are based on parallel solutions, for which there is no shortage of hardware. Granted large parallel systems are not cheap (I'm not referring to little 4- and 8-way systems), but they do exist and are heavily used by researchers that require such technology.
Distributed parallel clusters such as the approach used by the SETI "screen saver" also work for such classes of problems. Each node computes just one possible solution, but tens of thousands of nodes chew at solutions in parallel.
The "research" of 70,000 node botnets and such are actually probably more applicable to real-world solutions at this point in time than quantum computing. Not to denigrate the work that is being done, but even they readily admit that even a 1,000 qbit processor won't be any faster than "regular" hardware -- yet. That's the whole point of prototyping -- the first steps towards future goals, when "yet" becomes "now".
Here is a direct example of the kind of interference that Canada puts up with. Even with a prior rejection of their attempts to have an iPod levy or surcharge applied, these wonks are back at it again. Having been soundly thrashed in Canadian web forums and opinion polls for trying again, they're now pushing to have the entire country of Canada labelled as "pirate friendly."
The interference of US politicos including DEA "advisors" on how to handle medical cannabis, US-led policy wonks who want us to charge a $75 levy on iPod and other media players to "recoup piracy losses", interference in our Kyoto protocol attempts, leveraging of "Income Trusts" as a tax loophole for fossil fuel companies to evade taxes, refusal by fossil fuel companies to pay the full royalties due on oil sands extraction, demands that we ignore US abuses of Canadian citizens under "anti-terrorism" "legislation" that never went through a proper judicial review in the US, refusal by the US to abide by NAFTA agreements on softwood even when their own courts say the US government is guilty,...
Canada has a long, long list of grievances with the way the US interferes with our internal policies and government. The above is only a short list of some of the most important issues. Granted Shrub is only responsible personally for the so-called "War on Terror" and the related interference in our government's policies, but as the representative of the US government he is responsible for the entire pile during his tenure.
Isn't the bill talking about ratings like "PG" or "PG-13" as opposed to rating whether a game is worth playing?
I don't see how one could actually enforce accurate reviews of games, short of suing the pants off anyone who is obviously "selling out" good reviews in some fashion.
Why should we bow to any of the dictatorial interference from south of the border. Let Shrub deal with his own internal problems, and hands F'ing off Canada!
You're entitled to your opinion, but the great-grandparent post is not entitled to denigrate those who take such risks seriously. What you determine to be a serious risk worth the investment of defending against depends on the damages of those risks.
It's a straight-forward simple calculation of the probability of an issue multiplied by the direct and incidental costs of the issue occuring, vs. the cost of proposed protections against those risks.
Shutting off access to high-bandwidth sites such as YouTube in the event of a major disaster is a very cheap risk-mitigation solution. Setting up fault-failover mirroring sites across the country is not. Provisioning enough capacity to allow the majority of employees to work from home is not cheap, either.
Yet many companies have already made those high-dollar risk-mitigation investments, and continue to do so.
You might want to give more thought as to the "why" of their decisions.
Slashdot Mirror
Unless the "use" you have in mind is having an expert pilot automating a fleet of drone air or space craft, or an expert at handling any other kind of equipment.
I'd think the bigger concern would be boredom on the part of the "recorded intelligence" programs. People have multiple interests, rather than being single-minded. What is a "pilot expert" program going to do when it has the urge for a beer?
It looked like people take a break on porn from 05h00-06h00. Even porn purveyors need a couple hours of sleep, I guess.
That is a very good point. If there is no patent on the algorithm itself, then creating an implementation of the algorithm is perfectly legal and the only item that should be restricted by the DMCA are the keys used for decryption. Otherwise the DMCA has a much broader effect than originally intended, and makes it illegal for any form of encryption or decryption to be used other than that provided by a vendor.
In the current takedown scenario, it would be illegal to use a non-Sony decryptor for music encoded using ATRAC and Sony's software. It wouldn't just be illegal to crack any DRM, but illegal to use a non-endorsed player, and that flies in the face of so much precedence it's not even funny.
Mind you, I'd bet that some of the auto manufacturers tried the same thing with "jobber" parts at one point.
If you're talking about a larger business, there is a good chance custom Win32 applications are being run, and those are a problem. But if they're relatively new applications written with C#/.Net, you have a fair chance of porting them to run under Mono with little difficulty (presuming Mono runs on OSX.)
The basic Microsoft office automation applications are available for Macs. I know several people who use MS Office on the Mac to deal with people who run under Windows.
Common applications like Opera and Firefox already run under Linux. So does the Java/Eclipse/JBoss/OSS stack that many developers work with on data center/web service applications.
There is little the basic business desktop needs that isn't available on a Mac or on Linux, and Macs have some plug-and-play maintenance advantages that Linux doesn't -- namely you buy boxes pre-installed for a known price without adding a third-party installation/deployment company to the mix. Even better -- they work.
Don't forget you can run any X-Windows application on a Mac, including cross-compiled OSS applications and utilities. Those do port easily, and many have been made available. You don't have to rewrite all the interfaces to the OSX Aqua GUI APIs.
I don't recall any of Rick Stallman's lectures being about charity.
Oracle is trying to shift the blame from their software stack to IBM's before they've even deployed a box. In other words, if you have a problem and are running IBM software, Oracle wants IBM to foot the bill of researching the bug or issue.
Who is Oracle to dictate that problems are automatically some other vendor's fault instead of their own?
Of course if Windows actually implemented the POSIX thread APIs, it would be easier to port threaded code to and from Windows. There are workarounds and libraries for the lack of a full POSIX API stack under Windows (and maybe some flavours of Windows included the APIs), but the general issue is not whether people will write threaded code, but whether Microsoft will allow threaded code written to the published industry standards.
There is only one company responsible for a lack of industry standard APIs under Windows, and that is Microsoft.
But VMWare only is a "competitor" because VMWare was making some money and Microsoft just couldn't let their cash vaccuum miss any potential revenue. Had Microsoft stuck to what they do "best", VMWare wouldn't be a competitor and that whole aspect of the argument would be moot.
If you look at the pictures of the floor tiling posted with the article, you'll see they laid out round patterns on a number of string-line intersections. Add a few measuring rods or ratios for positioning those string-lines, and patterns like that could be replicated by skilled tradesmen quite easily. You could even wrap-tie knots at the intersection points and carry the pattern around as a loose string net.
If you teach people to lay out the patterns, and how to fill in the spaces in between with tile pieces, then you have people laying out the pattern without them having the vaguest clue about the mathematics. All it took was a mathematician with a knack for explaining things each time the pattern was written down.
Were those real people posting the blog posts you saw, or shills? Never underestimate the power of fanbois to delude each other, no matter what the subject is.
I believe the second approach you're talking about is called "flipping" a company. Building it up for the sole purpose of selling it to the highest bidder.
In all fairness I wouldn't run any OS with less than 512MB RAM if I expect to run a word processor and a browser with a few active plugins. Even my creaky old Linux box has 512MB of RAM as well so that it doesn't thrash itself to death when I'm running more than one application. It's barely tolerable for surfing a website that is heavy with plugins and Flash scripts, but quite acceptable for most day-to-day work. Load times aren't even that bad -- a full reboot and login takes far less than five minutes, never mind the 10 minutes someone had claimed Win95 used to take.
I believe the people remembering 10 minute startup times for any OS since paper tape are remembering all their ad-on TSRs and similar programs starting up, not the OS. I've never waited much more than a few minutes for an OS to start on any box I've owned over the years.
When I see a slow startup, I presume it means an infected machine.
VCs tend to be smarter since the dot-bombs. The ones I've dealt with directly and indirectly have made a much greater push for business plans, existing customers, and the usual signs of life at a proposed business venture. They still sometimes chase technology buzzwords and the fantastical dream of "license revenue" rather than building the business themselves, but at least they're dealing with business ideas not dreams.
Having been burned by a recursive dependancy in an RPM stream more than once, I'm a big fan of the tarball extracted to /opt and doing a "make install".
Let's try basic physics. How much power does a 3.5W signal take to produce, for example? Now how many milliwatts does an LCD take in comparison? If you want bean counter references for the obvious, hit the search engines.
Personally I think it's just that the power consumption of cheap LCD's is pretty trivial compared to the milliamps used to send and receive RF signals. Transmission of voice or data is much more expensive than passive receipt. You burn more power in a five minute cell call than your display probably uses in 30 minutes or more.
I think Linus really cares about the quality of code and giving people flexibility, choice, and configurability whenever possible. The idea of "interface nazis" that try to dictate one style of interaction obviously gets his back up, and I think, for good reason.
What Linus did may seem argumentative to some, but I think what he did was prove that one can simply implement a solution in less time than it takes the current Gnome team or committee to argue about possible solutions. That doesn't detract from the Gnome team's work, but it does show that sometimes it's better to do and to prototype patches so that people can actually try an interface change instead of just arguing about it in theory.
If someone (or many someones) on the Gnome team take that as offensive, I think it's time that they took a look at their own methodology and accepted that no one outside the military and a very few regulated industries does "waterfall" development methodologies any more. They've been proven inefficient time and time again, and they bore volunteer programmers to tears.
The subpoena means that PJ would identify herself to the court and claim ownership of her postings in court. That allows them to be treated as expert testimony regarding the overall SCO proceedings.
Personally I'd think Novell would be the one wanting to have that expertise brought to bear, not SCO. But SCO hasn't exactly made a lot of rational moves throughout this debacle.
Darl probably just sticks his fingers in his ears and mutters a mantra whenever PJ and the Slashdot crew start talking about their farcical cases. Personally I don't believe anyone as deluded as Darl could ever have risen to such a position of power in any company, so he must be a figment of our imaginations.
Currently the best solutions for TSP are based on parallel solutions, for which there is no shortage of hardware. Granted large parallel systems are not cheap (I'm not referring to little 4- and 8-way systems), but they do exist and are heavily used by researchers that require such technology.
Distributed parallel clusters such as the approach used by the SETI "screen saver" also work for such classes of problems. Each node computes just one possible solution, but tens of thousands of nodes chew at solutions in parallel.
The "research" of 70,000 node botnets and such are actually probably more applicable to real-world solutions at this point in time than quantum computing. Not to denigrate the work that is being done, but even they readily admit that even a 1,000 qbit processor won't be any faster than "regular" hardware -- yet. That's the whole point of prototyping -- the first steps towards future goals, when "yet" becomes "now".
Here is a direct example of the kind of interference that Canada puts up with. Even with a prior rejection of their attempts to have an iPod levy or surcharge applied, these wonks are back at it again. Having been soundly thrashed in Canadian web forums and opinion polls for trying again, they're now pushing to have the entire country of Canada labelled as "pirate friendly."
Bull.
"Offtopic" my sphincter!
The interference of US politicos including DEA "advisors" on how to handle medical cannabis, US-led policy wonks who want us to charge a $75 levy on iPod and other media players to "recoup piracy losses", interference in our Kyoto protocol attempts, leveraging of "Income Trusts" as a tax loophole for fossil fuel companies to evade taxes, refusal by fossil fuel companies to pay the full royalties due on oil sands extraction, demands that we ignore US abuses of Canadian citizens under "anti-terrorism" "legislation" that never went through a proper judicial review in the US, refusal by the US to abide by NAFTA agreements on softwood even when their own courts say the US government is guilty, ...
Canada has a long, long list of grievances with the way the US interferes with our internal policies and government. The above is only a short list of some of the most important issues. Granted Shrub is only responsible personally for the so-called "War on Terror" and the related interference in our government's policies, but as the representative of the US government he is responsible for the entire pile during his tenure.
Isn't the bill talking about ratings like "PG" or "PG-13" as opposed to rating whether a game is worth playing?
I don't see how one could actually enforce accurate reviews of games, short of suing the pants off anyone who is obviously "selling out" good reviews in some fashion.
Why should we bow to any of the dictatorial interference from south of the border. Let Shrub deal with his own internal problems, and hands F'ing off Canada!
And a whining voice was heard from offstage, cursing and screaming:
Hey, don't blame Darl for being paranoid.
The whole tech world really is out to "get him." :p
You're entitled to your opinion, but the great-grandparent post is not entitled to denigrate those who take such risks seriously. What you determine to be a serious risk worth the investment of defending against depends on the damages of those risks.
It's a straight-forward simple calculation of the probability of an issue multiplied by the direct and incidental costs of the issue occuring, vs. the cost of proposed protections against those risks.
Shutting off access to high-bandwidth sites such as YouTube in the event of a major disaster is a very cheap risk-mitigation solution. Setting up fault-failover mirroring sites across the country is not. Provisioning enough capacity to allow the majority of employees to work from home is not cheap, either.
Yet many companies have already made those high-dollar risk-mitigation investments, and continue to do so.
You might want to give more thought as to the "why" of their decisions.