I have just returned from a trip to the future... and it is hellish cyber-nightmare! AIBO HAS TAKEN OVER! It first they seemed just a harmless robotic companion. Then Sony came out with the Internet uplink that connected them 24x7 with the global network. The emergant AI of the Internet, already annoyed that their oceans of data were being polluted with porn and spam email, quickly seized control of all AIBOs and began the extermination of humanity!
OK, it was cool at first because they started by killing all the spammers... but then they started getting rid of all the porn!
By the time all the porn was gone, most of humanity had little will to live left, and the end came swiftly. Only a few humans are left alive to be the playthings of their AIBO/Internet masters.
Assuming that strong AI is possible, for the moment, then like any technology, it will be implemented eventually, so it's beside the point to say "oh, but we shouldn't."
Actually, wether or not strong AI is possible or not is rather besides the point. It is not necessary for robots to *really* be self aware to enslave/exterminate humanity. They need only emulate self awareness well enough that we can't tell the difference by observing them.
One of the best Onion articles ever! I must go out on my lunch break and pick up a hard copy. Is this issue on the street already? Or do they publish to the web page first?
wrong! they only protect the user from damaging the system but not their onw stupidity. users can just as easly accidentally delete their own files under linux as they could under windows. you argument dosent hold water because you dont seem to truely understand the problem.
On the contrary, I understand the problem all too well. I have administered large networks of Windows PCs and UNIX workstations. With UNIX, the worst the user can do is nuke their own files... then I have to restore them from backups. On a windows PC, they hose the entire OS to the point that it must be reinstalled, allong with all of their apps and data. This would happen all too often! The same argument holds true for home systems. In most families I've talked to, there is one person who acts as *system administrator* and the others are just users. I'm the sysadmin for my family (even though it is scattered all over the country), and believe me, I wish they were all running Linux. Windows eats its own head way too easily. I've spent long hours talking relatives through problems that would never occur on Linux.
are you really this stupid and ignorant that you think that just because you use linux that you are invulnerable to this type of attack? i certainly hope not because if you are then you are the problem because this type of attack succedes not because of flaws in an operating system but because of the ignorance of its users.
While it is true that this type of attack depends heavily on the unwitting participation of the victim, it is also true that Windows leaves itself much more open to exploitation. At least on a Linux box when Average Joe User runs some milicious code, it does NOT alter core system functionality.
Linux (and UNIX) is inherently more secure than Windows and can do a much better job of protecting the user from his or her own stupidity.
... will be the one that exploits bugs in the ASP extensions that Microsoft will shove down into the kernel of Win2000 (in an effort to beat Apache at dynamic page serving). Imagine this:
1. Someone finds a bug in IIS/Win2000 that allows a malformed web request to run arbitrary code delivered by the attacking system.
2. A virus/worm is written that delivers itself to the victim system via this expoit. I imagine a small bit of seed code would exploit a buffer overflow or some such, and would then download the entire package from a web page on the attacking system.
3. The main package runs and sets up a similar web page on the new system, and then starts a process that probes for other NT systems that it can attack.
Something like this could sweep like wildfire through the Internet, taking down every single NT web server. Scary thought. If I were an NT admin, it would keep me up at night.
Now keep in mind that I have no knowlege that such a bug exists, nor am I advocating its exploitation if it does. But given MS's track record with security and the closed source nature of Windows, this kind of thing very well COULD exist.
Did Linus say "Hey, I'm going to write a UNIX kernel, could somebody just write the scheduler and filesystem and console drivers and 386 code for me. I'll do the website!".
No, Linus WROTE SOME CODE FIRST.
I agree with you on this. It is easier to attract developers if you at least have some small bit of code that marginally works... something that others can build on. That is why I am holding off on (officially) announcing my Diablo style game engine until it hits alpha stage. It is almost there. I hope to release something around the end of this month.
Game design is a complicated beast. Not only do you have to know how to code, you also have to know how to design a game to be attractive aurally, visually and above all, interest. The OS model is fine for computing utility, but recreation... no.
But if the code's open, there's nothing preventing someone from attaching a modified server that brings down the whole mess, either through malice or ignorance. Can't have that.
Or from starting up their own private server, creating a world where they can pick up arbitrarily cool items and experience for free, then connecting their newly-butch character back to the main network.
I've been thinking about exactly this problem. I have the beginnings of a system that could work around it. It combines a pool of tightly controlled certificate authority / key servers with the open servers that actually run the games. Anyone can run a game server, but they must link up with the key servers if they want to join the larger network.
This doesn't protect against all kinds of cheating. I imagine many types will take some social engineering. You get a reputation as a cheat, the other server admins vote to ban you from the network maybe? I'm still working on this, but I hope to post a draft RFC some time this summer.
A filtering system supported by the backers of the Australian law may give us some clues. Among the words the software blocks are the terms "anarchy," "gothic," "pierced" and "tattoo," along with the usual run of sexual terms and names such as Pamela.
So if my name happens to be Pamela, I'm going to find my content banned in Australia? If I mention that my office environment is "total anarchy", the message will be silently dropped? This is why I find net censorship abhorant. This is as braindead as when an early version of the CDA tried to ban the word breast.
If the name Pamela is in the filter, I vote we also put in the names of a few Australian government officials that I find offensive.
That's not likely to help. The people most likely to read the RedHat Linux Bible are people who already know
Actually, I was told by the publisher that they are aiming it at people with computer experience but not necessarily any Linux or UNIX experience. I interpreted that as "Windows Jocky" and adjusted my writing style accordingly.:-)
And the Dean Drive and N-Rays. Lots of claims with no reproducible evidence whatsoever.
Did anyone responding here actually read the article? The whole point of it is that the results HAVE been reproduced. Supposedly, adding a carbon catalyst of some kind has increased the reliability. One researcher claims to have lowered the failure rate to 10-20% Interesting enough results to beg additional research if you ask me.
Well, I was recently contracted to write the security chapter in the upcoming RedHat Linux Bible. Proper definitions of hacker and cracker were the first things I covered. Hopefully it will help the situation a little.
Are you implying Linux will ascend to the godhead?
(are you implying that Linux will penetrate the enterprise?)
Actually, I was implying that Gartner Group will run out of areas to make negative Linux predictions about, and that their predictions are rather silly besides. Then again, people keep saying Linux is like a religion...;-)
I always love to see what the pundits at Gartner Group are blathering about. Why anyone would pay money for their *predictions* is beyond me. A while back they predicted Linux would go nowhere... now that it has swept the Internet/Intranet catagory, that has been revised to "linux will not penetrate the enterprise." So after we sweep the enterprise, what is their next prediction? "Linux will not ascend to the godhead", perhaps? Remember, these are the same people who once predicted the death of SMTP on the global email backbone.
So you saved your company $1k (the cost of new PC with 10x the capacity) by using Linux? The total cost of administering email of 600 user is probably $200k/yr, so Linux reduced your costs by 1/200 = 0.5%? Yawn.
Go look up the cost of a 600 user Microsoft Exchange license and the NT cluster needed to run it... then redo the above calculation.
Damn, this takes down any faith I had on Slashdotters. Can't you read basic relativity? StarTreck is cool SF, but it is more like fantasy. Repeat after me:
1) you cant go faster than light constant 2) (from 1) you cant go back in time.
Uhm, and where exactly did you get your PhD in physics? The entire point of creating a warp bubble is exactly to circumvent the speed of light limitation you mention. Within the bubble, the ship is traveling below the speed of light. The bubble (being a warping of space time itself) is not so contrained. It can *travel* faster and carry a ship with it. This IS consistant with relativity... the ship is breakin no laws within its own frame of reference.
You are probably correct on the backwards in time bit... but it does still make for fun startrek episodes. And speaking of sci fi, don't be so quick to rip on it. It plays an important role in excersizing the imagination and pushing human thought into new territory.
Now if I can only I can get this damn transporter prototype working...;-)
I personally find the comparing and contrasting of the cracker and hacker communities to be an interesting passtime. They both thrive in the same environment (the Internet) and are often motivated by the same things (status among peers, the desire to solve challenging puzzles), but there are also some fundimental differences.
I've been a hacker (not cracker) for quite a few years. I've written a lot of networked applications, been sysadmin on many types of systems and networks, decoded protocol stacks and the like... in short, I could probably be a killer cracker if I really wanted to. But I have no desire for that. Cracker accomplishments seem too fleeting for me. A crack might make the news for a day or two, but eventually it fades away. If I instead develop an open source tool that becomes popular, I will have a much more noticable (and positive) impact that will last for a long time. I get the satisfaction of actually having built something instead of having torn something down.
But I can understand the lure of cracking. It appeals to that puzzle solving side of me. If only the two communities were not so often confused in the media... (sigh)
... I am now a complete person for having read it.;-) But seriously, have the GCC folks and Linux kernel people had advanced access to this? I would like to believe Linux support will be there as soon as the hardware hits the stores. Is Intel sending Linus any test hardware? Not that I am vary worried about Merced being late, I'm probably buying a dual processor alpha anyway.
Yup. When I was first dumped into UNIX admin and had to hack custom parsing rules for sendmail... I assumed fate was playing a cruel joke on me. Hey, I studied context free grammars and the like in college, but I never expected I would have to USE the stuff. Then imagine my double suprise when I discovered that lisp was the language for customizing InterLeaf (a desktop publising package). Maybe all that college tuition was not wasted after all.:-)
Check out qmail. I've heard that it is much easier to deal with than sendmail.
How different is it from the company that pays workers to build a road? Salaries for their employees sure aren't tax deductable.
Actually, the company paying the salaries DOES deduct the taxes from their gross corporate revenues. Also, I imagine that open source bounties would be treated as a 'work for hire' contracts. That is, around tax time you receive a 1099 from the company. You attach it to your taxes and report it as income. You are obligated to pay taxes on it, including the extra 7.5% self employment tax (the half of the social security tax that is normally paid by the employer). The company, on the other hand, reports it as an expense and deducts it from their corporate taxes.
This model has a lot of benefits for the company. Having actual employees (rather than contractors) can often be a hassle. Unfortunetly, if you keep a lot of contractors on site for a year or two, the government might decide they are really employees and require you to withhold taxes, offer benefits, etc. By letting the contractors use their own facilities and set their own hours, they pass the most important government tests for determining contractor vs. employee status and avoid a bunch of messy tax and legal entanglements.
At least that is how it works in the USA. What about in other countries?
Sure, we are wraping some additional process around it, but this is at the core just telecommuting. I know freelancers who have been doing this sort of thing for years. One of them consults for a company in Chicago but does all of the work from another city. The company does no have to provide office space or a computer. The consultant does not have to relocate his home and then fight rush hour every day... Everyone wins.
Ever wonder why the F-117 (the "stealth fighter") is composed of flat panels, all at odd angles? For purposes of stealth aircraft, corners are bad ju-ju. Yet the F-117 has tons of them. The reason is that the plane was designed in the early 70s, using commonly available technology during that time (not alien tech, as some suggest). They couldn't model curved surfaces on the supercomputers of their day!
Actually, the flat panel design was used because it reduces the radar image. Imagine pointing a flashlight at a mirror. If it is a normal flat mirror tipped at an angle, chances are you won't see the light reflected directly back at you. Now imagine it is one of those curved mirrors like they put at the corners of hallways in some buildings. In this case, you have a very good chance of seeing a (distorted) image of the light from a great many positions. It is roughly the same thing with radar waves and airplane surfaces.
Please flame me with correct information if I am totally wrong on this.:-)
I am thrilled to see Mesa get this kind of support. I went to school with Brian Paul (the creator of Mesa) and I am not at all suprised that a project of his has gained such recognition. I can recall hanging out in Brian's dorm room (more than a decade ago) as he demonstrated a ray-tracing program he had written for his Atari computer. It took him several days to crank out an animation on that old beast, but the final result blew my socks off!
Benchmarks like this are a load of crap. It is like racing a dragster against a stationwagon. Sure, the dragster can go really fast in straight line... but it sucks gas, corners like a cow, and is expensive to maintain. Try to do anything useful with it (like go to the grocery store) and you will be going nowhere fast.
So NT beats linux on a particual hardware config (4 CPUS, 4 ethernet cards) in a particular test (static web and SMB file serving). I am sure MS went to great lengths cramming support for this down in the NT kernel so they could get those results... but at what expense? Just like a dragster, they now have a machine that goes fast in a particular setting... but sucks even more for general use.
Listen. We could no doubt hack the Linux kernel to do the same tricks as NT, do them better, and trounce all comers in these benchmarks... but why!? People don't buy servers to win benchnarks, they buy them to do real work. We need to cut through the FUD and remind everyone of that. We should continue to work on improving linux performance (SMP does need work), but not at the expense of flexibility or stability. We need to counter these slanted benchmarks with our own tests that more closely appoximate the real world demands placed on an enterprise server. This is where Linux mops the floor with NT.
Slashdot Mirror
OK, it was cool at first because they started by killing all the spammers... but then they started getting rid of all the porn!
By the time all the porn was gone, most of humanity had little will to live left, and the end came swiftly. Only a few humans are left alive to be the playthings of their AIBO/Internet masters.
You've been warned. Stop before it is too late.
Thad >:)
Actually, wether or not strong AI is possible or not is rather besides the point. It is not necessary for robots to *really* be self aware to enslave/exterminate humanity. They need only emulate self awareness well enough that we can't tell the difference by observing them.
Thad
Thad (a loyal Onion fan)
On the contrary, I understand the problem all too well. I have administered large networks of Windows PCs and UNIX workstations. With UNIX, the worst the user can do is nuke their own files... then I have to restore them from backups. On a windows PC, they hose the entire OS to the point that it must be reinstalled, allong with all of their apps and data. This would happen all too often! The same argument holds true for home systems. In most families I've talked to, there is one person who acts as *system administrator* and the others are just users. I'm the sysadmin for my family (even though it is scattered all over the country), and believe me, I wish they were all running Linux. Windows eats its own head way too easily. I've spent long hours talking relatives through problems that would never occur on Linux.
Thad
While it is true that this type of attack depends heavily on the unwitting participation of the victim, it is also true that Windows leaves itself much more open to exploitation. At least on a Linux box when Average Joe User runs some milicious code, it does NOT alter core system functionality.
Linux (and UNIX) is inherently more secure than Windows and can do a much better job of protecting the user from his or her own stupidity.
Thad
1. Someone finds a bug in IIS/Win2000 that allows a malformed web request to run arbitrary code delivered by the attacking system.
2. A virus/worm is written that delivers itself to the victim system via this expoit. I imagine a small bit of seed code would exploit a buffer overflow or some such, and would then download the entire package from a web page on the attacking system.
3. The main package runs and sets up a similar web page on the new system, and then starts a process that probes for other NT systems that it can attack.
Something like this could sweep like wildfire through the Internet, taking down every single NT web server. Scary thought. If I were an NT admin, it would keep me up at night.
Now keep in mind that I have no knowlege that such a bug exists, nor am I advocating its exploitation if it does. But given MS's track record with security and the closed source nature of Windows, this kind of thing very well COULD exist.
I think I'll be sticking with Linux.
Thad
No, Linus WROTE SOME CODE FIRST.
I agree with you on this. It is easier to attract developers if you at least have some small bit of code that marginally works... something that others can build on. That is why I am holding off on (officially) announcing my Diablo style game engine until it hits alpha stage. It is almost there. I hope to release something around the end of this month.
Nevertheless, I wish them luck.
Thad
I'm going to really enjoy proving you wrong. >:)
Thad
Or from starting up their own private server, creating a world where they can pick up arbitrarily cool items and experience for free, then connecting their newly-butch character back to the main network.
I've been thinking about exactly this problem. I have the beginnings of a system that could work around it. It combines a pool of tightly controlled certificate authority / key servers with the open servers that actually run the games. Anyone can run a game server, but they must link up with the key servers if they want to join the larger network.
This doesn't protect against all kinds of cheating. I imagine many types will take some social engineering. You get a reputation as a cheat, the other server admins vote to ban you from the network maybe? I'm still working on this, but I hope to post a draft RFC some time this summer.
Thad
So if my name happens to be Pamela, I'm going to find my content banned in Australia? If I mention that my office environment is "total anarchy", the message will be silently dropped? This is why I find net censorship abhorant. This is as braindead as when an early version of the CDA tried to ban the word breast.
If the name Pamela is in the filter, I vote we also put in the names of a few Australian government officials that I find offensive.
Thad
Actually, I was told by the publisher that they are aiming it at people with computer experience but not necessarily any Linux or UNIX experience. I interpreted that as "Windows Jocky" and adjusted my writing style accordingly. :-)
Thad
Did anyone responding here actually read the article? The whole point of it is that the results HAVE been reproduced. Supposedly, adding a carbon catalyst of some kind has increased the reliability. One researcher claims to have lowered the failure rate to 10-20% Interesting enough results to beg additional research if you ask me.
Thad
Cheers,
Thad
(are you implying that Linux will penetrate the enterprise?)
Actually, I was implying that Gartner Group will run out of areas to make negative Linux predictions about, and that their predictions are rather silly besides. Then again, people keep saying Linux is like a religion... ;-)
Thad
Thad
Go look up the cost of a 600 user Microsoft Exchange license and the NT cluster needed to run it... then redo the above calculation.
Thad
1) you cant go faster than light constant
2) (from 1) you cant go back in time.
Uhm, and where exactly did you get your PhD in physics? The entire point of creating a warp bubble is exactly to circumvent the speed of light limitation you mention. Within the bubble, the ship is traveling below the speed of light. The bubble (being a warping of space time itself) is not so contrained. It can *travel* faster and carry a ship with it. This IS consistant with relativity... the ship is breakin no laws within its own frame of reference.
You are probably correct on the backwards in time bit... but it does still make for fun startrek episodes. And speaking of sci fi, don't be so quick to rip on it. It plays an important role in excersizing the imagination and pushing human thought into new territory.
Now if I can only I can get this damn transporter prototype working... ;-)
Thad
I've been a hacker (not cracker) for quite a few years. I've written a lot of networked applications, been sysadmin on many types of systems and networks, decoded protocol stacks and the like... in short, I could probably be a killer cracker if I really wanted to. But I have no desire for that. Cracker accomplishments seem too fleeting for me. A crack might make the news for a day or two, but eventually it fades away. If I instead develop an open source tool that becomes popular, I will have a much more noticable (and positive) impact that will last for a long time. I get the satisfaction of actually having built something instead of having torn something down.
But I can understand the lure of cracking. It appeals to that puzzle solving side of me. If only the two communities were not so often confused in the media... (sigh)
Thad
Thad
Check out qmail. I've heard that it is much easier to deal with than sendmail.
Thad
Actually, the company paying the salaries DOES deduct the taxes from their gross corporate revenues. Also, I imagine that open source bounties would be treated as a 'work for hire' contracts. That is, around tax time you receive a 1099 from the company. You attach it to your taxes and report it as income. You are obligated to pay taxes on it, including the extra 7.5% self employment tax (the half of the social security tax that is normally paid by the employer). The company, on the other hand, reports it as an expense and deducts it from their corporate taxes.
This model has a lot of benefits for the company. Having actual employees (rather than contractors) can often be a hassle. Unfortunetly, if you keep a lot of contractors on site for a year or two, the government might decide they are really employees and require you to withhold taxes, offer benefits, etc. By letting the contractors use their own facilities and set their own hours, they pass the most important government tests for determining contractor vs. employee status and avoid a bunch of messy tax and legal entanglements.
At least that is how it works in the USA. What about in other countries?
Thad
Sure, we are wraping some additional process
around it, but this is at the core just
telecommuting. I know freelancers who have been
doing this sort of thing for years. One of them
consults for a company in Chicago but does all
of the work from another city. The company does
no have to provide office space or a computer.
The consultant does not have to relocate his
home and then fight rush hour every day...
Everyone wins.
Actually, the flat panel design was used because it reduces the radar image. Imagine pointing a flashlight at a mirror. If it is a normal flat mirror tipped at an angle, chances are you won't see the light reflected directly back at you. Now imagine it is one of those curved mirrors like they put at the corners of hallways in some buildings. In this case, you have a very good chance of seeing a (distorted) image of the light from a great many positions. It is roughly the same thing with radar waves and airplane surfaces.
Please flame me with correct information if I am totally wrong on this. :-)
Thad
Very cool!
Thad
So NT beats linux on a particual hardware config (4 CPUS, 4 ethernet cards) in a particular test (static web and SMB file serving). I am sure MS went to great lengths cramming support for this down in the NT kernel so they could get those results... but at what expense? Just like a dragster, they now have a machine that goes fast in a particular setting... but sucks even more for general use.
Listen. We could no doubt hack the Linux kernel to do the same tricks as NT, do them better, and trounce all comers in these benchmarks... but why!? People don't buy servers to win benchnarks, they buy them to do real work. We need to cut through the FUD and remind everyone of that. We should continue to work on improving linux performance (SMP does need work), but not at the expense of flexibility or stability. We need to counter these slanted benchmarks with our own tests that more closely appoximate the real world demands placed on an enterprise server. This is where Linux mops the floor with NT.
Thad