MIT's Ray Kurzweil projects artificially intelligent machines
evolving so rapidly in the early part of this century that they will ultimately fuse with biological beings.
Many novelists and filmmakers share these dark visions. They see smart machines as inevitably
replicating, and surpassing human beings in longevity, endurance, intelligence and raw power.
These machines will dominate us. Truth or more techno-hype?
It's just hype, not even techno.
Some AI researchers have made wildly optimistic predictions about how quickly AI would advance. They have been mostly wrong so far. We just barely got a computer program that can beat a person in chess - and that's by use of brute computational power and some very clever programming by people.
We still don't really understand what this thing called "intelligence" really is. How do you expect to solve a problem without understanding it?
The idea of "evolution" of machine is just a cop out. We don't know how to create a something, so we just put things in a room and how they will create themselves and in only 50 years.
Good grief! Biological evolution took several billions of years and it occured as a massively parallel computation (this view is stolen from Stanilaw Lem).
Just think. Programming computers is essentially engineering. To solve problems engineers need some science that explains how things work, otherwise they just hack (and sometimes create working systems). But without science the engineer is as good as an alchemist.
Now consider that Newton figured out the science of howto fly to the moon in the 18th century and took took engineers over 200 years to actually build a machine that could do it.
Why do you think it's possible to create an "intelligent" machine, when we can't even agree on the definition of the problem...
I could make a much better living than I am by driving to your house, stealing all of your things and selling them off.
First of all when someone distributes your music you should thank them! That means there will be more people that hear your band and maybe you'll create more fans.
If you don't want people to hear your music for free don't record anything. Just charge for live performances.
You are in a position of a horse-n-buggy driver at the beginning of the 20th century. Your profession is being eliminated by technology and you better adapt.
...richie
P.S. Copyright violation is not stealing. You still have the recording in question. It's quite different from stealing physical objects.
At $30k, it's affordable for all those applications and more. This guy could sell a ton of machines without ever having to listen to one whiny retail-level customer or worry about the ever-changing copyright law.
Your point about selling the machine is correct.
Now imagine that 5 years from now the machine costs $2000. What happens to the book publishing industry then?
Sure. Now imagine just how much they would learn if they tried to do the same thing using, say IIS. They would learn nearly as much. And while some students may well learn a lot about hacking around with Unix (perhaps largely because they would have to do so !), I find it hard to argue that this (by itself) is a good reason to install Linux. Let's face it, most people are NOT going to be sysadmins when they grow up, most people are NOT going to use Emacs at work, most people do NOT care whether or not you learn Perl.
But how useful will knowledge of IIS be 8 years from now, when these kids finish college and entr job market? At least with open source tools you can dig into the meat of the system if you want to and learn the principles that don't change that much.
Just because people use office now and not Emacs at work, does not mean that they will be still using office 10 years from now.
BTW, I've been using Emacs as my main editor for about 15 years. Name a piece of MS software that is still in wide use after 15 years?
I already thought about the idea of books-on-demand several years ago (it's a pretty obvious idea).
Starting a bussiness around it is tough though, because of the copyright lock-up on most published works. Although, I thought of offering Project Gutenberg books via this method.
Does anyone have FIRST HAND experience with educational software for Linux that they could recommend? Not
just a site that promotes the stuff, but specific programs that are worthwhile.
Basically I think that all the so called "educational software" is trash. AFAIK these are just simple games, that follow try to make boring drills of spelling or arithmetic little more exiting.
The real education in computers should in learning how they work and how they are programmed. See all the writings of S. Papert on LOGO.
What you need to give the kids is computer "Legos" and let them build their own stuff. Be it web sites, games, etc. You can learn a lot more stuff from building things, than from all the "educational" software.
People round here might demonise Microsoft, but at the end of the day education is education and it doesn't matter
how it is provided or who by, as long as it is impartial and rounded.
But you have to be careful to try and educate the the children, not just train them in using MS Word.
The idea of using computers to write should be introduced, but this can be done just as easily with Emacs as with Word (and Emacs ha been in use much longer than Word has).
Imagine how much you'd learn about computers from setting up a network of Linux machine, setting up web servers and setting up some usefull cool apps (like a web site on which homework assigments are posted daily by the teachers).
Many forms of content that we enjoy on the web today will not be here in a few years -- then we'll be starving for
that great content we had for free, and they'll be able to charge for it -- because we'll see the value of it after its
gone.
Which forms of "content" are those? The best part of Slashdot (and sites like it) are the user comments. We can always go back to Usenet. It's the discussions between people that are more fun, not the stuff that "contents producers" think want to consume.
Here is a link to a paper that explains why paying for content on the Internet will not happen.
The summary is that "content" is not something that people are willing to pay for (eg. see TV and Radio), but people spent a lot more money on things that let them communicate with each other (eg. phones, cell-phones etc).
Read the papers by Andrew Odlyzko that are referenced from this page for some historical background.
...richie
P.S. I got this link from a discussion on micro-payments on kuros5in.
Problem is that it's also moronic that next spring when I buy new seeds from Burpee, I'll have to either remember the
password that I created in 1999
Wouldn't it be better to just be able to buy things over the net without having to create accounts and spreading your credit card numbers over many databases?
I don't have to login and remember my password each time I shop at Home Depot!
But the thing about science is that it is self correcting. If I get a result that cannot be reproduced by others (remember "cold fusion"), then in the long run our knowledge will advance.
However, this depends on open sharing or ideas and results, not hiding them from criticism.
A good, excellent point. And I guess that brings up my next question: Why would a company like us risk our business ventures and put our
trust in the GPL when (as far as I know) it has seen little to no court time? Yeah, there's been one or two cases where it has been upheld,
but there's no way I'd convince our CEO that it was a good idea. I'm curious for comments on this.
You have to consider the benefits of having open source software. First of all, if you are lucky, you get people to read your code. Perhaps your algorithms are not as good as you think and would benefit from independent review.
Another plus for a small company is that now you can tell your customers "you don't have to worry about us staying in bussiness, as you can always get the code". This can be a significant problem that companies worry about when dealing with small suppliers. How many clients have you lost because they decided to go with a larger company as it seemed less risky.
To get paid you just continue to do what you do today - develop and support the software. Your clients should be willing to pay for this, as you are the one's who know the code best and are dest qualified to manage the direction of your product.
Having an open source GPL-ed solution could also set a "defacto" standard for your market and you would be the setters.
As far as challenging GPL in court, I hope it wouldn't be required. Just think how much bad publicity a violation of the GPL would bring.
Of course convincing the CEO is a whole different problem..:-)
Here's my example: I work at a small-town Iowa computer firm. We create software in the health industry where competition is
tight and our tiny company is certainly overshadowed by many 'big dogs'. We have many good, new ideas on how to make our
software do the job for our clients cleaner and better. However, if we don't watch it, we could release a product and one of our
larger competitors could easily and quickly copy our idea, exploit it, and make a fortune.
Well, if your improvement is visibile to the user, then what prevents your competitor from just copying your ideas? After all if they have a lot of money they can re-implement the same features.
To that end, then, open source is still a scary idea for us. If we were to publish our code, it would instantly be snatched up and
exploited
But if you GPL-ed your code, then the competitor could snatch it and use it, but as soon as they wanted to sell their product they would have to release their own source as well. Do you think they would do that?
It seems to me that GPL protects you better than keeping your code secret.
These are two game systems that run on DOS. They are pretty low-tech (ASCII character graphics etc), but both include a game editor that allows you to build your own games - including programing the elements of the game.
Any game you download includes the source so they can learn that way. My son learned to code using these systems when he was 9 years old.
...richie
P.S. Another idea is to set up Linux machine and let them build a website (using Netscape Composer etc).
PPS. Look up ZZT and MegaZeux in Google to find the programs.
This idea is would not be as bad if it was completly under the control of the user. If I could select a word on a page and then add an annotation (some comments, perhaps links to other things) and then when I got to the same word/phrase elsewhere I could see my annotation.
It would be even cooler if I could share my annotations with other people.
Of course MS does not seem to care much about the user here - rather it's trying to build another marketing "bring-in-the-eyeballs" tool.
...richie
P.S. Look up "web annotations" on Google - there is plenty of research along these lines.
The technical quality of the recording is only a small part of it. If the music is great, the recording quality doesn't matter. For example, are the early recordings of Louis Armstrong (let's say) worse than latest N'Sync CD?
I've never met him either. However I find his pseudo-intellectual lyrics really bad and in general his political correctness really annoys me. The last album I have of his is the one with the song "If you love somebody" (the song, whose main line comes from a Hallmark card!) - at least the musicians on that album were excellent.
Slashdot Mirror
It's just hype, not even techno.
Some AI researchers have made wildly optimistic predictions about how quickly AI would advance. They have been mostly wrong so far. We just barely got a computer program that can beat a person in chess - and that's by use of brute computational power and some very clever programming by people.
We still don't really understand what this thing called "intelligence" really is. How do you expect to solve a problem without understanding it?
The idea of "evolution" of machine is just a cop out. We don't know how to create a something, so we just put things in a room and how they will create themselves and in only 50 years.
Good grief! Biological evolution took several billions of years and it occured as a massively parallel computation (this view is stolen from Stanilaw Lem).
Just think. Programming computers is essentially engineering. To solve problems engineers need some science that explains how things work, otherwise they just hack (and sometimes create working systems). But without science the engineer is as good as an alchemist.
Now consider that Newton figured out the science of howto fly to the moon in the 18th century and took took engineers over 200 years to actually build a machine that could do it.
Why do you think it's possible to create an "intelligent" machine, when we can't even agree on the definition of the problem...
First of all when someone distributes your music you should thank them! That means there will be more people that hear your band and maybe you'll create more fans.
If you don't want people to hear your music for free don't record anything. Just charge for live performances.
You are in a position of a horse-n-buggy driver at the beginning of the 20th century. Your profession is being eliminated by technology and you better adapt.
P.S. Copyright violation is not stealing. You still have the recording in question. It's quite different from stealing physical objects.
Perfection is the enemy of the good.
Well, if more books become available in a digital format, how much time will it take before "Bookster"....
Your point about selling the machine is correct.
Now imagine that 5 years from now the machine costs $2000. What happens to the book publishing industry then?
But how useful will knowledge of IIS be 8 years from now, when these kids finish college and entr job market? At least with open source tools you can dig into the meat of the system if you want to and learn the principles that don't change that much.
Just because people use office now and not Emacs at work, does not mean that they will be still using office 10 years from now.
BTW, I've been using Emacs as my main editor for about 15 years. Name a piece of MS software that is still in wide use after 15 years?
Starting a bussiness around it is tough though, because of the copyright lock-up on most published works. Although, I thought of offering Project Gutenberg books via this method.
I doubt there is enough money in it.
Basically I think that all the so called "educational software" is trash. AFAIK these are just simple games, that follow try to make boring drills of spelling or arithmetic little more exiting.
The real education in computers should in learning how they work and how they are programmed. See all the writings of S. Papert on LOGO.
What you need to give the kids is computer "Legos" and let them build their own stuff. Be it web sites, games, etc. You can learn a lot more stuff from building things, than from all the "educational" software.
But you have to be careful to try and educate the the children, not just train them in using MS Word.
The idea of using computers to write should be introduced, but this can be done just as easily with Emacs as with Word (and Emacs ha been in use much longer than Word has).
Imagine how much you'd learn about computers from setting up a network of Linux machine, setting up web servers and setting up some usefull cool apps (like a web site on which homework assigments are posted daily by the teachers).
Which forms of "content" are those? The best part of Slashdot (and sites like it) are the user comments. We can always go back to Usenet. It's the discussions between people that are more fun, not the stuff that "contents producers" think want to consume.
The summary is that "content" is not something that people are willing to pay for (eg. see TV and Radio), but people spent a lot more money on things that let them communicate with each other (eg. phones, cell-phones etc).
Read the papers by Andrew Odlyzko that are referenced from this page for some historical background.
P.S. I got this link from a discussion on micro-payments on kuros5in.
Yes it does. Say hello to OpenNap....
Wouldn't it be better to just be able to buy things over the net without having to create accounts and spreading your credit card numbers over many databases?
I don't have to login and remember my password each time I shop at Home Depot!
However, this depends on open sharing or ideas and results, not hiding them from criticism.
You have to consider the benefits of having open source software. First of all, if you are lucky, you get people to read your code. Perhaps your algorithms are not as good as you think and would benefit from independent review.
Another plus for a small company is that now you can tell your customers "you don't have to worry about us staying in bussiness, as you can always get the code". This can be a significant problem that companies worry about when dealing with small suppliers. How many clients have you lost because they decided to go with a larger company as it seemed less risky.
To get paid you just continue to do what you do today - develop and support the software. Your clients should be willing to pay for this, as you are the one's who know the code best and are dest qualified to manage the direction of your product.
Having an open source GPL-ed solution could also set a "defacto" standard for your market and you would be the setters.
As far as challenging GPL in court, I hope it wouldn't be required. Just think how much bad publicity a violation of the GPL would bring.
Of course convincing the CEO is a whole different problem.. :-)
Well, if your improvement is visibile to the user, then what prevents your competitor from just copying your ideas? After all if they have a lot of money they can re-implement the same features.
To that end, then, open source is still a scary idea for us. If we were to publish our code, it would instantly be snatched up and exploited
But if you GPL-ed your code, then the competitor could snatch it and use it, but as soon as they wanted to sell their product they would have to release their own source as well. Do you think they would do that?
It seems to me that GPL protects you better than keeping your code secret.
Any game you download includes the source so they can learn that way. My son learned to code using these systems when he was 9 years old.
P.S. Another idea is to set up Linux machine and let them build a website (using Netscape Composer etc).
PPS. Look up ZZT and MegaZeux in Google to find the programs.
It would be even cooler if I could share my annotations with other people.
Of course MS does not seem to care much about the user here - rather it's trying to build another marketing "bring-in-the-eyeballs" tool.
P.S. Look up "web annotations" on Google - there is plenty of research along these lines.
OK. Maybe I'll give it shot....richie
Perhaps if he got another lyricist... :-)
P.S. Good bass player though... :-)