One element of this article sounds just like Stephen Jay Gould's evolutionary theory of Punctuated Equilibrium.
IANAET (evolutionary theorist) so take these comments with care.
One element of the theory of punctuated evolution says that new species arise not by direct competition against their parent species, but by finding an isolated and protected niche where they can develop.
Say a new species of horse is to develop.
A subpopulation becomes isolated and has a chance to develop new characteristics and to become reproductively distinct (no longer interbreeds with the parent species).
Then, when the geographical isolation ends, the new and parent species come into contact and competition. The new species spreads rapidly, having had a chance to strengthen in isolation.
This theory is designed in part to explain gaps in the fossil record. The small, original populaiton of the new species leaves few fossils - we only see them after explosive growth - so some intermediate forms are lost.
That sounds like the article's model of innovation succeeding by finding a niche market before improving the product to compete head on head in the general marketplace.
Wonder what other analogies exist with evolutionary theory and this article.
This article does a terrible job of representing Ben Shneiderman's objections
to Hal-like interfaces. The article focuses only the verbal memory issue.
But, Ben has been objecting to Hal-like interfaces for 20 years. He is a well-known
computer science critic of artifical intelligence and intelligent agents. He specifically
objects to making voice more practical by making the computer more intelligent. Just think
of normal human conversation and the number of errors and corrections and hmms and ahs
and redundancy elements. Verbal communication is rife with the potential for error, and
add that to powerful, independent, software agents and you could have chaos. Ben doesn't
want computers to be "clever secretaries" that take "initiative" on our behalf.
All the posts I've seen on this thread focus on the content of this one Washington Post
article and are not written with knowledge of his extensive writings on the possibility
and ethics of intelligent computer agents.
Here is one way of putting his fundamental objection. Would you rather have:
a) A surgeon working on your heart with his/her own hands?
or
b) A surgeon using a microphone to direct a robotic knife with only auditory feedback?
Even if the latter were programmed to "know" certain surgical tasks, isn't there
a sense of a weak, indirection interaction with the device through voice? Wouldn't
you feel better if the surgeon were more directly, visually connected to the system?
Ben's point has always been that direct, visual interfaction with objects (what he calls direct
manipulation) will be safer and more efficient than a voice system, and that making the voice system
smarter is not a solution to the problem.
Ben has debated this point in many forums. This spring he debated this at the AAAS
(American Association for the Advancement of Science) conference with Jim Hendler,
an AI prof at the University of Maryland. When I was a graduate student there years ago you could
count on Ben to tweak the pure AI types as to the ethics and reliability of their work.
A 1997 debate with Pattie Maes of MIT is easy to find online and is in PDF format. Try
CHI97 debate home debate in PDF
Don't take this Washington Post article as anything but what it is, an extremely superficial
summary of a complex position. There's lots more to read on this. In the long run, Ben's
objections may be left in the dust, but for the moment he serves an important role by critically
evaluating AI claims and predictions.
If you do research on the edge of programing you can't always restrict the use of your work to areas you consider ethical.
Say you don't want to contribute new ideas to the US military for autonomous killing machines. So you work on medical imaging instead. You invent a new algorithm for identifying red blood cells.
Ah, but the algorithm can detect tanks. This is a real case from the 1980s - one ethical member of our research group avoided the military funded grants, wanting to stay non-violent. In the technical report library one day our technical officer from DARPA was hunting up reports and took some written by our ethical member. Loved the reports.
It is a foundational principle of computer science that you can't draw absolute lines in evaluating algorthms (Rice's theorem and all that.) This algorithm, ethical; this algorithm, unethical. Not a decidable proposition.
That makes research in computer science a muddy ethical venue. You can't pretend your work is pure, like G.H. Hardy in "Apology of a Mathematician" who claimed his mathematical research never killed anyone; CP Snow noted that Hardy trained the british scientists who fought WWII.
You can get Matlab for Linux - I run copies on RedHat - so the implication of the post that Matlab for Mac OS X would finally bring Matlab to Unix is a little strong.
I have not bothered to purchase the current Mac Matlab version because it is stabilized at version 5 and has not been updated. I have found no reason to pay many $$ to Matlab
for an obsolete version. But, a letter this spring from Mathworks indicated that my individual license will be converted to looser wording. Under it the license holder can install Matlab on multiple CPUS, under mutiple OSes, as long as the license holder is using them serially. This enables me to get the current Mac Matlab at no cost to supplement the Linux version. Now, if they update it to Mac OS X, I'd be very happy.
Another lesson from the end of the 19th century is the story of the new media of photography for the common man. Handheld cameras, fast emusions, paper roll film and photographic labs were all new then.
A collaborator of Edison, George Eastman of Eastman Kodak, behaved like our own Bill Gates. Eastman tried to corner the patents on the new technology of mass production photographic equipment - lots of good stories about him stiff arming competitors and trying to become a monopolist.
Gives you an opportunity to see what happens to technology monopolists after a hundred years. Got Fuji?
As a NASA research who uses COTS (common off the shelf systems/software), I would argue to a Florida senator that this act would threaten the space agency and other agencies important to the state. The use of off the shelf hardware and software is important to federal agencies because COTS is mass produced and thereby cheaper than equipment produced specially for the agencies.
The bill as I read it does not distinguish between ultimate uses of the technology. I do not see an exception for the manufacture of computers and/or components for space, military, IRS or any scientific use. Not only are consumers sitting at home assumed to be criminals, but so are IRS managers purchasing systems for branch offices.
This means that government agencies and the researchers they support with tax dollars (including the aforementioned NASA with considerable presence in Florida) will be saddled with two bad choices:
1. Continue to buy off the shelf systems that conform to the act, thereby paying a Hollywood tax in extra complexity, resulting in higher costs (ie, higher taxes) and less reliability. (Think of the space shuttle running on software with DRM.)
2. Switch over to special systems without DRM produced especially for these agencies, resulting in higher purchasing costs and possibly higher management costs (who is going to make sure that the supercomputers at NOAA aren't cracking DVDs?) And, this produces a manufacturing base for non-comforming systems that now has to be policed.
The military angle might the hardest hitting.
The Democratic senators supporting this act would see to be leaving themselves open to charges they are soft on defense - that this bill would negatively impact our military by complicating military technology. I'd ask Hollings this:
Which industry do you support, the one that gives us American Pie or the one that won the battles in Afganistan?
Follow up arguments on how the act would effect research can be found in a letter at the ACM web site:
Write letters to the eidtor of your local newspaper, write your old instructor at college, write your mom.
About 1976 in Missouri a ragtag group of us (and deed major credit to others) defeated an Army Corp of Engineers dam that had been
planned for 30 years for the Merrimac River
south of St. Louis. Our representatives all
supported it - big money, inside influence,
and all that was on the other side.
But, they did not focus on public opinion. We
did. We organized letters to the St. Louis and
all the local papers on a regular basis - we talked to the editors of the newspapers to get
positive editorials - we honed and refined our
arguments - we conducted opinion surveys -
we did a thousand little grass roots activities.
We got to know reporters, held protests, created organizations and issued press releases.
In this case, think of:
a) Finding a local newspaper and write to them this week. Then get a friend to write a letter the next weeek. Then repeat. The editors decide what to publish partly on volume.
b) After a while, organize a group and ask to see the edtior about the newspaper talking a stand in an editorial. See what it takes to get them to commit.
c) Figure out who does the local tech beat at the paper (or tv station) and figure what it takes to get them to write a story. Talk to them if you can. Organize, hold protests, issue press releases, picket a movie theater - get press.
d) Find allies. Who else cares? Movie theater owners? Perhaps a secure digital pipe means the theaters will lose out. Having chamber of commerce types on your side helps.
e) Produce graphics. T-shirts "I hate Hollings - ask me why!", Web banners, posters, anything to make it easy for someone else to express this opinion.
f) And, instead of letters, show up in force at your representatives town meetings. Ask questions at these forums. Make them give their opinion in public.
It takes hard, continuous work to fight a battle like this, and you have to match the opponent on all the battle fields, not just Congress.
I've downloaded from Epson the OS X drivers for the 780 and the 750 printers. No problem.
Seems to be the same driver.
But, the Apple official OS X site did not list the 750 as a supported printer, and I think neither did the Epson official list. I only found the 750 driver by looking for it under the long list of 750 drivers.
Slashdot Mirror
IANAET (evolutionary theorist) so take these comments with care.
One element of the theory of punctuated evolution says that new species arise not by direct competition against their parent species, but by finding an isolated and protected niche where they can develop.
Say a new species of horse is to develop. A subpopulation becomes isolated and has a chance to develop new characteristics and to become reproductively distinct (no longer interbreeds with the parent species).
Then, when the geographical isolation ends, the new and parent species come into contact and competition. The new species spreads rapidly, having had a chance to strengthen in isolation.
This theory is designed in part to explain gaps in the fossil record. The small, original populaiton of the new species leaves few fossils - we only see them after explosive growth - so some intermediate forms are lost.
That sounds like the article's model of innovation succeeding by finding a niche market before improving the product to compete head on head in the general marketplace.
Wonder what other analogies exist with evolutionary theory and this article.
This article does a terrible job of representing Ben Shneiderman's objections to Hal-like interfaces. The article focuses only the verbal memory issue.
But, Ben has been objecting to Hal-like interfaces for 20 years. He is a well-known computer science critic of artifical intelligence and intelligent agents. He specifically objects to making voice more practical by making the computer more intelligent. Just think of normal human conversation and the number of errors and corrections and hmms and ahs and redundancy elements. Verbal communication is rife with the potential for error, and add that to powerful, independent, software agents and you could have chaos. Ben doesn't want computers to be "clever secretaries" that take "initiative" on our behalf.
All the posts I've seen on this thread focus on the content of this one Washington Post article and are not written with knowledge of his extensive writings on the possibility and ethics of intelligent computer agents.
Here is one way of putting his fundamental objection. Would you rather have:
a) A surgeon working on your heart with his/her own hands?
or
b) A surgeon using a microphone to direct a robotic knife with only auditory feedback?
Even if the latter were programmed to "know" certain surgical tasks, isn't there a sense of a weak, indirection interaction with the device through voice? Wouldn't you feel better if the surgeon were more directly, visually connected to the system?
Ben's point has always been that direct, visual interfaction with objects (what he calls direct manipulation) will be safer and more efficient than a voice system, and that making the voice system smarter is not a solution to the problem.
Ben has debated this point in many forums. This spring he debated this at the AAAS (American Association for the Advancement of Science) conference with Jim Hendler, an AI prof at the University of Maryland. When I was a graduate student there years ago you could count on Ben to tweak the pure AI types as to the ethics and reliability of their work. A 1997 debate with Pattie Maes of MIT is easy to find online and is in PDF format. Try
CHI97 debate home
debate in PDF
Don't take this Washington Post article as anything but what it is, an extremely superficial summary of a complex position. There's lots more to read on this. In the long run, Ben's objections may be left in the dust, but for the moment he serves an important role by critically evaluating AI claims and predictions.
Say you don't want to contribute new ideas to the US military for autonomous killing machines. So you work on medical imaging instead. You invent a new algorithm for identifying red blood cells.
Ah, but the algorithm can detect tanks. This is a real case from the 1980s - one ethical member of our research group avoided the military funded grants, wanting to stay non-violent. In the technical report library one day our technical officer from DARPA was hunting up reports and took some written by our ethical member. Loved the reports.
It is a foundational principle of computer science that you can't draw absolute lines in evaluating algorthms (Rice's theorem and all that.) This algorithm, ethical; this algorithm, unethical. Not a decidable proposition.
That makes research in computer science a muddy ethical venue. You can't pretend your work is pure, like G.H. Hardy in "Apology of a Mathematician" who claimed his mathematical research never killed anyone; CP Snow noted that Hardy trained the british scientists who fought WWII.
You can get Matlab for Linux - I run copies on RedHat - so the implication of the post that Matlab for Mac OS X would finally bring Matlab to Unix is a little strong.
I have not bothered to purchase the current Mac Matlab version because it is stabilized at version 5 and has not been updated. I have found no reason to pay many $$ to Matlab for an obsolete version. But, a letter this spring from Mathworks indicated that my individual license will be converted to looser wording. Under it the license holder can install Matlab on multiple CPUS, under mutiple OSes, as long as the license holder is using them serially. This enables me to get the current Mac Matlab at no cost to supplement the Linux version. Now, if they update it to Mac OS X, I'd be very happy.
A collaborator of Edison, George Eastman of Eastman Kodak, behaved like our own Bill Gates. Eastman tried to corner the patents on the new technology of mass production photographic equipment - lots of good stories about him stiff arming competitors and trying to become a monopolist.
Gives you an opportunity to see what happens to technology monopolists after a hundred years. Got Fuji?
The bill as I read it does not distinguish between ultimate uses of the technology. I do not see an exception for the manufacture of computers and/or components for space, military, IRS or any scientific use. Not only are consumers sitting at home assumed to be criminals, but so are IRS managers purchasing systems for branch offices.
This means that government agencies and the researchers they support with tax dollars (including the aforementioned NASA with considerable presence in Florida) will be saddled with two bad choices:
1. Continue to buy off the shelf systems that conform to the act, thereby paying a Hollywood tax in extra complexity, resulting in higher costs (ie, higher taxes) and less reliability. (Think of the space shuttle running on software with DRM.)
2. Switch over to special systems without DRM produced especially for these agencies, resulting in higher purchasing costs and possibly higher management costs (who is going to make sure that the supercomputers at NOAA aren't cracking DVDs?) And, this produces a manufacturing base for non-comforming systems that now has to be policed.
The military angle might the hardest hitting. The Democratic senators supporting this act would see to be leaving themselves open to charges they are soft on defense - that this bill would negatively impact our military by complicating military technology. I'd ask Hollings this:
Which industry do you support, the one that gives us American Pie or the one that won the battles in Afganistan?
Follow up arguments on how the act would effect research can be found in a letter at the ACM web site:
http://www.acm.org/usacm/SSSCA-letter.html
Write letters to the eidtor of your local newspaper, write your old instructor at college, write your mom.
About 1976 in Missouri a ragtag group of us (and deed major credit to others) defeated an Army Corp of Engineers dam that had been planned for 30 years for the Merrimac River south of St. Louis. Our representatives all supported it - big money, inside influence, and all that was on the other side.
But, they did not focus on public opinion. We did. We organized letters to the St. Louis and all the local papers on a regular basis - we talked to the editors of the newspapers to get positive editorials - we honed and refined our arguments - we conducted opinion surveys - we did a thousand little grass roots activities. We got to know reporters, held protests, created organizations and issued press releases.
In this case, think of:
a) Finding a local newspaper and write to them this week. Then get a friend to write a letter the next weeek. Then repeat. The editors decide what to publish partly on volume.
b) After a while, organize a group and ask to see the edtior about the newspaper talking a stand in an editorial. See what it takes to get them to commit.
c) Figure out who does the local tech beat at the paper (or tv station) and figure what it takes to get them to write a story. Talk to them if you can. Organize, hold protests, issue press releases, picket a movie theater - get press.
d) Find allies. Who else cares? Movie theater owners? Perhaps a secure digital pipe means the theaters will lose out. Having chamber of commerce types on your side helps.
e) Produce graphics. T-shirts "I hate Hollings - ask me why!", Web banners, posters, anything to make it easy for someone else to express this opinion.
f) And, instead of letters, show up in force at your representatives town meetings. Ask questions at these forums. Make them give their opinion in public.
It takes hard, continuous work to fight a battle like this, and you have to match the opponent on all the battle fields, not just Congress.
But, the Apple official OS X site did not list the 750 as a supported printer, and I think neither did the Epson official list. I only found the 750 driver by looking for it under the long list of 750 drivers.