That was the point. The so-called back surgery with stem cells could be a scam too, pretty much right up to the point where they get their work through peer review.
Thanks, scp i use already, but didn't know about the windows version, and dyndns is new to me. OTOH, I am one of those rare people with a static ip at the moment, so not as needed. Streamlining the internet does unfortunately still leave you with a long wait unless you have an awfully fast connection.
You have 400 megabyte of data. You want to take it with you to work on (or maybe listen to) at another computer. You can:
Flash drive: Copy to flash drive at 10megabytes/sec. Call that a minute with overhead. Requires the destination computer have USB.
Internet: Email it through google mail, using googlefs at the speed of your internet connection. Typically, most people today are living with 5 megabit per second or less. Call that 15 minutes, more if you can't max out your connection, or are living with a slower connection. Requires destination computer have (fast!) internet service. 15 minutes or more likely to extract your data at the other end. This is all assuming there is no overhead for google mail. If you have static ip, maybe you are hosting this data directly, still requires a typical 15 minute one way trip, but how many people have a static ip for their home machine?
Portable hard drive: Copy to portable hard drive at 20 megabytes/sec. Call that 30 seconds, but costs more than the flash option.
I'll take either of the carry it with me options over the internet most days. Even more so on days when my data set that needs to travel is 30 gigabyte.
What they got out of xbox was a change in market and mindshare. Prior to XBOX, the console market had 3 players: Sony, Nintendo, Sega, in that order.
Post XBOX we have
Sony, Microsoft, Nintendo, in that order.
In addition, they've set the expectation that in the next generation, it will either be Sony or Microsoft in the lead, and if Sony is leading, Microsoft will at least be close, and Microsoft will be first to market.
At $4B, I think Microsoft got itself a bargain to be in this position already.
Sadly, knowing an awful lot of people in the world, I can't honestly imagine a single one who would look at each browser and make a sound logical choice.
I think everyone i've ever met can safely be put into one of these categories:
1) IE comes with windows, and they don't care enough to ever switch 2) had some severe negative experience with IE, but didn't know enough about computers to fix, switched 3) hate microsoft in general, switched
Not one person I can think of made any sort of careful rational comparison of available browsers. That's just not normal lazy human behavior. I know your post was supposed to be sarcastic, but you're wrong if you think there is going to be any significant number of people who do a browser comparison and pick the best one.
I can't see where. It in no way specifies in what ways I may make the copy. Producing the copy by p2p looks perfectly legitimate to me, and I've certainly never heard of a court case of any kind conflicting with that.
Note that in the context of the article, we'd have to question whether we'd be better off with one of these than with the space shuttle.
1) couldn't have been done for less than the space shuttle, which provided cheap reliable frequent trips to a lower orbit. 2) surely not as useful as the space shuttle 3,4) only really useful to astronomy, and they got a good enough chunk of the budget as it was 5) couldn't have been done for the price of the shuttle
So it seems likely that none of these would have been preferable to what we did in the past, what about the future?
1) no getting around the fuel costs without 5, why not wait for 5? 2) not that useful without 5. 3) could come out of the new moon program, so in some sense we are working on this. certainly not significantly harder than having a heavy lifting moon program. 4) we have other telescope programs in the pipeline, the complaints about telescopes come from a small group of vocal astonomers who always want more telescope time in their own tiny frequency range to explore one extreme edge case or another. advances in ground based telescope design have made nearly all of these people cheaper to satisfy on the ground, and the people who have made the best cases for new telescopes in space have new telescopes in space coming. 5) should wait a while yet, as many of the nano manufacturing technologies needed are still developing, and will continue to develop without involvement from nasa. the materials people have enough motivation for this, we don't need to take a chunk of nasa's limited money away for this right now.
The reason many people want laptops is not to be tied to work at any time, but to be freed from the need to be in the office. Many computer workers can work just as effectively from the location of their choice with a laptop. From another state, another country, to the park down the street. Think of a laptop as a massive upgrade in the quality of your workspace, better than the corner office on the top floor.
Assuming you've learned the basics: variables, printing, and logic then you are most likely stuck on either program organization (in which case you might want to try to read a book on object oriented programming methodology), or on the complexities of learning an external interface for a specific task (ie learning how to program graphics, given the example you gave in another post of wanting to program quake), in which case, you'd want to read some book on that specific subject.
Graphics in particular is one of the most difficult areas of programming, so if that is where you are getting stuck, don't be too surprised. Even 2d graphics requires a fair amount of math (mainly algebra, geometry), and 3d is just a nightmare unless your math skills (particularly linear algebra, geometry) are really sharp. If you want to learn graphics, you may want to learn/review those math subjects before diving into a graphics programming book.
I've developed 2d and 3d games, so if you're really interested, feel free to write back, i'd be happy to try to answer your questions.
Under section 117, you or someone you authorize may make a copy of an original computer program if:
* the new copy is being made for archival (i.e., backup) purposes only;
* you are the legal owner of the copy; and
* any copy made for archival purposes is either destroyed, or transferred with the original copy, once the original copy is sold, given away, or otherwise transferred.
So as long as the copy is being made for archival purposes, the downloader is in the clear.
To clear up my issue with NERO: the AI will be at the level to make RTS gaming interesting when the player no longer has to be involved. When the AI can effectively choose obstacles and enemies to train its own units, we'll be where we need to be.
Having written genetic algorithm driven neural net based AI in the past, I think I have a reasonable idea what is going on in the internals of NERO. I could be a little wrong, but I'm probably not far off. I'm unsurprised that it can learn some adequate squad based tactics level techniques. But there are two problems here:
1) adequate will not compete with a decent human player, nor even with good scripting. 2) the squad level is not where games have trouble with their ai today (at the squad level, rapidity of control lends an enormous advantage to the ai).
Where AI needs to improve is at the strategic level.
None of what i'm posting is meant to denegrate the NERO work. It's great work, definitely innovative, very interesting. It just doesn't address the AI problem that most people perceive in games today at all.
I do actually think that something in the realm of genetically trained neural net based AI is the way of the future... but the CPU needed to really do this at a sufficiently complicated level just doesn't exist right now.
If you know those guys... ask them what it would take (iterations, cpu, memory) for their system to train a 16 million input net, if i'm willing to grant that an average node has connections to only roughly the square-root of the number of nodes (4 thousand). That's what I believe will be able to effectively play starcraft.
Yeah, I've seen NERO before. I think the key problems with it in my view are:
a) player spends lots of time training the AI. This is essentially equivalent to what goes on in game development houses already, where the developers spend lots of time training the AI. But in games, it is not typically fun, nor should it be expected that the player will train the AI.
b) the player trains the AI. When AI will really take off is when the AI will train itself. And that will take AI with close to the brainpower of the training player. Granted, it will be specialized for training in the context of the specific game, so it won't take nearly as much computing power as a general purpose brain, but it will still take a lot more CPU than we have available today.
c) the AI is very narrowly trainable. There are a very small number of inputs and outputs to the NERO intelligence. For a more realistic game to be played against a human, this will have to be scaled way up, and unfortunately this technique tends to scale worse than linearly in the number of inputs and outputs.
Re:keeping pc gaming alive
on
Ask Sid Meier
·
· Score: 1
I agree, it isn't about to dry up and disappear... it has been doing that for quite some time. PC market share has been sliding one or two percent per year for the last decade. There is still plenty of money to be made there, but at some point the experience will become more like the mac, where the PC will only get second rate remakes of top console titles.
Lawsuits do not illegality make. Note again that the person sued was the uploader, not the downloader. I don't deny that the person serving the content is taking a potentially civilly liable action, I'm claiming the downloader is in the clear.
Having done a lot of work in game AI, i'll give you one answer, we'll see if Sid agrees.
Game AI needs a lot more computing power to be really interesting. There are fairly straightforward reasons for this:
On the stupid end of AI, you have a flat scripted AI. Any scripted AI becomes entirely predictable when the player becomes sufficiently experienced.
In the middle ground of AI, you have a tree scripted AI which goes down various branches in response to game conditions. This AI fares better against the player, but again suffers once the player learns the tree, or how to force the AI down a specific branch.
Further up from here, you have an AI which manages and combines multiple branches based on game conditions, and randomly picks some branches to prevent itself from being forced into any one branch condition. This AI still suffers when the human player learns all of its tricks, and so no possible branch is effective.
So this is basically where AI is stuck at today. Dynamic management of multiple pre-scripted strategies.
The next level is where AI gets interesting: dynamic strategy development. Here we're orders of magnitude off in terms of the processor power needed to really do this effectively.
And that's the basic problem. The obvious next rung in the ladder of AI is just way more computationally complex than existing methods.
That's not the definition of civil disobedience at all. It can be all about taking something denied to you by law. For example:
A black person riding on a whites only bus, or drinking from a whites only drinking fountain, or eating at a whites only restaurant.
Socrates teaching students that he had been forbidden to teach.
Jesus holding religious ceremonies in violation of the law.
Free speech advocates holding rallies without a permit.
Re:keeping pc gaming alive
on
Ask Sid Meier
·
· Score: 1
Well, that would really get at the heart of the question: if consoles are closing in on PC capabilities, but are a more consistent (easier) and profitable environment for developers, what if anything will keep PC gaming going?
keeping pc gaming alive
on
Ask Sid Meier
·
· Score: 5, Interesting
What factors do you think help keep PC gaming alive when competing with consoles, and do you foresee that PC gaming will continue to survive when confronted with the next generation of consoles? Or from the reverse perspective, what prevents consoles from finally killing off PC gaming?
Well duh, they can't put two cameras back to back and take a 360 degree photo, that would be against the laws of physics. And man you do not want to be pulled over by god for a physics violation, that is a bad scene.
Downloading may or may not be infringement is the problem. If you own the media, you have the right to make a backup. Downloading a digital copy is a very effective way to make a backup.
Slashdot Mirror
That was the point. The so-called back surgery with stem cells could be a scam too, pretty much right up to the point where they get their work through peer review.
Thanks, scp i use already, but didn't know about the windows version, and dyndns is new to me. OTOH, I am one of those rare people with a static ip at the moment, so not as needed. Streamlining the internet does unfortunately still leave you with a long wait unless you have an awfully fast connection.
I think the main reason is this:
You have 400 megabyte of data. You want to take it with you to work on (or maybe listen to) at another computer. You can:
Flash drive: Copy to flash drive at 10megabytes/sec. Call that a minute with overhead. Requires the destination computer have USB.
Internet: Email it through google mail, using googlefs at the speed of your internet connection. Typically, most people today are living with 5 megabit per second or less. Call that 15 minutes, more if you can't max out your connection, or are living with a slower connection. Requires destination computer have (fast!) internet service. 15 minutes or more likely to extract your data at the other end. This is all assuming there is no overhead for google mail. If you have static ip, maybe you are hosting this data directly, still requires a typical 15 minute one way trip, but how many people have a static ip for their home machine?
Portable hard drive: Copy to portable hard drive at 20 megabytes/sec. Call that 30 seconds, but costs more than the flash option.
I'll take either of the carry it with me options over the internet most days. Even more so on days when my data set that needs to travel is 30 gigabyte.
What they got out of xbox was a change in market and mindshare. Prior to XBOX, the console market had 3 players: Sony, Nintendo, Sega, in that order.
Post XBOX we have
Sony, Microsoft, Nintendo, in that order.
In addition, they've set the expectation that in the next generation, it will either be Sony or Microsoft in the lead, and if Sony is leading, Microsoft will at least be close, and Microsoft will be first to market.
At $4B, I think Microsoft got itself a bargain to be in this position already.
Sadly, knowing an awful lot of people in the world, I can't honestly imagine a single one who would look at each browser and make a sound logical choice.
I think everyone i've ever met can safely be put into one of these categories:
1) IE comes with windows, and they don't care enough to ever switch
2) had some severe negative experience with IE, but didn't know enough about computers to fix, switched
3) hate microsoft in general, switched
Not one person I can think of made any sort of careful rational comparison of available browsers. That's just not normal lazy human behavior. I know your post was supposed to be sarcastic, but you're wrong if you think there is going to be any significant number of people who do a browser comparison and pick the best one.
Actually, you do. That's what prevents faith healing from getting a lot of scholarly respect.
I can't see where. It in no way specifies in what ways I may make the copy. Producing the copy by p2p looks perfectly legitimate to me, and I've certainly never heard of a court case of any kind conflicting with that.
A little polite respect, and the threat of expensive fines.
And what would you do differently if you could go back and reverse that decision?
http://www.sciencemaster.com/jump/space/shuttle.ph p
u ttlefacts-toc.html
t s.htm
e r.htm
a ts.html
http://www.nasa.gov/centers/kennedy/news/facts/sh
http://www-pao.ksc.nasa.gov/kscpao/factoids/shfac
http://www-pao.ksc.nasa.gov/kscpao/factoids/orbit
http://www.cbsnews.com/network/news/space/spacest
http://www.astronautix.com/lvs/shuttle.htm
You're welcome.
Note that in the context of the article, we'd have to question whether we'd be better off with one of these than with the space shuttle.
1) couldn't have been done for less than the space shuttle, which provided cheap reliable frequent trips to a lower orbit.
2) surely not as useful as the space shuttle
3,4) only really useful to astronomy, and they got a good enough chunk of the budget as it was
5) couldn't have been done for the price of the shuttle
So it seems likely that none of these would have been preferable to what we did in the past, what about the future?
1) no getting around the fuel costs without 5, why not wait for 5?
2) not that useful without 5.
3) could come out of the new moon program, so in some sense we are working on this. certainly not significantly harder than having a heavy lifting moon program.
4) we have other telescope programs in the pipeline, the complaints about telescopes come from a small group of vocal astonomers who always want more telescope time in their own tiny frequency range to explore one extreme edge case or another. advances in ground based telescope design have made nearly all of these people cheaper to satisfy on the ground, and the people who have made the best cases for new telescopes in space have new telescopes in space coming.
5) should wait a while yet, as many of the nano manufacturing technologies needed are still developing, and will continue to develop without involvement from nasa. the materials people have enough motivation for this, we don't need to take a chunk of nasa's limited money away for this right now.
The reason many people want laptops is not to be tied to work at any time, but to be freed from the need to be in the office. Many computer workers can work just as effectively from the location of their choice with a laptop. From another state, another country, to the park down the street. Think of a laptop as a massive upgrade in the quality of your workspace, better than the corner office on the top floor.
Assuming you've learned the basics: variables, printing, and logic then you are most likely stuck on either program organization (in which case you might want to try to read a book on object oriented programming methodology), or on the complexities of learning an external interface for a specific task (ie learning how to program graphics, given the example you gave in another post of wanting to program quake), in which case, you'd want to read some book on that specific subject.
Graphics in particular is one of the most difficult areas of programming, so if that is where you are getting stuck, don't be too surprised. Even 2d graphics requires a fair amount of math (mainly algebra, geometry), and 3d is just a nightmare unless your math skills (particularly linear algebra, geometry) are really sharp. If you want to learn graphics, you may want to learn/review those math subjects before diving into a graphics programming book.
I've developed 2d and 3d games, so if you're really interested, feel free to write back, i'd be happy to try to answer your questions.
Not at all ... making a single backup copy of a work you own is a usage authorized by the copyright laws.
l
http://www.copyright.gov/help/faq/faq-digital.htm
Under section 117, you or someone you authorize may make a copy of an original computer program if:
* the new copy is being made for archival (i.e., backup) purposes only;
* you are the legal owner of the copy; and
* any copy made for archival purposes is either destroyed, or transferred with the original copy, once the original copy is sold, given away, or otherwise transferred.
So as long as the copy is being made for archival purposes, the downloader is in the clear.
To clear up my issue with NERO: the AI will be at the level to make RTS gaming interesting when the player no longer has to be involved. When the AI can effectively choose obstacles and enemies to train its own units, we'll be where we need to be.
... but the CPU needed to really do this at a sufficiently complicated level just doesn't exist right now.
... ask them what it would take (iterations, cpu, memory) for their system to train a 16 million input net, if i'm willing to grant that an average node has connections to only roughly the square-root of the number of nodes (4 thousand). That's what I believe will be able to effectively play starcraft.
Having written genetic algorithm driven neural net based AI in the past, I think I have a reasonable idea what is going on in the internals of NERO. I could be a little wrong, but I'm probably not far off. I'm unsurprised that it can learn some adequate squad based tactics level techniques. But there are two problems here:
1) adequate will not compete with a decent human player, nor even with good scripting.
2) the squad level is not where games have trouble with their ai today (at the squad level, rapidity of control lends an enormous advantage to the ai).
Where AI needs to improve is at the strategic level.
None of what i'm posting is meant to denegrate the NERO work. It's great work, definitely innovative, very interesting. It just doesn't address the AI problem that most people perceive in games today at all.
I do actually think that something in the realm of genetically trained neural net based AI is the way of the future
If you know those guys
Yeah, I've seen NERO before. I think the key problems with it in my view are:
a) player spends lots of time training the AI. This is essentially equivalent to what goes on in game development houses already, where the developers spend lots of time training the AI. But in games, it is not typically fun, nor should it be expected that the player will train the AI.
b) the player trains the AI. When AI will really take off is when the AI will train itself. And that will take AI with close to the brainpower of the training player. Granted, it will be specialized for training in the context of the specific game, so it won't take nearly as much computing power as a general purpose brain, but it will still take a lot more CPU than we have available today.
c) the AI is very narrowly trainable. There are a very small number of inputs and outputs to the NERO intelligence. For a more realistic game to be played against a human, this will have to be scaled way up, and unfortunately this technique tends to scale worse than linearly in the number of inputs and outputs.
I agree, it isn't about to dry up and disappear ... it has been doing that for quite some time. PC market share has been sliding one or two percent per year for the last decade. There is still plenty of money to be made there, but at some point the experience will become more like the mac, where the PC will only get second rate remakes of top console titles.
Lawsuits do not illegality make. Note again that the person sued was the uploader, not the downloader. I don't deny that the person serving the content is taking a potentially civilly liable action, I'm claiming the downloader is in the clear.
squares are a lot easier to render with rectangular textures, and lend themselves more naturally to array based representations.
Having done a lot of work in game AI, i'll give you one answer, we'll see if Sid agrees.
Game AI needs a lot more computing power to be really interesting. There are fairly straightforward reasons for this:
On the stupid end of AI, you have a flat scripted AI. Any scripted AI becomes entirely predictable when the player becomes sufficiently experienced.
In the middle ground of AI, you have a tree scripted AI which goes down various branches in response to game conditions. This AI fares better against the player, but again suffers once the player learns the tree, or how to force the AI down a specific branch.
Further up from here, you have an AI which manages and combines multiple branches based on game conditions, and randomly picks some branches to prevent itself from being forced into any one branch condition. This AI still suffers when the human player learns all of its tricks, and so no possible branch is effective.
So this is basically where AI is stuck at today. Dynamic management of multiple pre-scripted strategies.
The next level is where AI gets interesting: dynamic strategy development. Here we're orders of magnitude off in terms of the processor power needed to really do this effectively.
And that's the basic problem. The obvious next rung in the ladder of AI is just way more computationally complex than existing methods.
That's not the definition of civil disobedience at all. It can be all about taking something denied to you by law. For example:
A black person riding on a whites only bus, or drinking from a whites only drinking fountain, or eating at a whites only restaurant.
Socrates teaching students that he had been forbidden to teach.
Jesus holding religious ceremonies in violation of the law.
Free speech advocates holding rallies without a permit.
Well, that would really get at the heart of the question: if consoles are closing in on PC capabilities, but are a more consistent (easier) and profitable environment for developers, what if anything will keep PC gaming going?
What factors do you think help keep PC gaming alive when competing with consoles, and do you foresee that PC gaming will continue to survive when confronted with the next generation of consoles? Or from the reverse perspective, what prevents consoles from finally killing off PC gaming?
Well duh, they can't put two cameras back to back and take a 360 degree photo, that would be against the laws of physics. And man you do not want to be pulled over by god for a physics violation, that is a bad scene.
Downloading may or may not be infringement is the problem. If you own the media, you have the right to make a backup. Downloading a digital copy is a very effective way to make a backup.