To be a programmer without ever learning assembly language is like being a professional race car driver without understanding how your carburetor works.
Do you really think the driver needs to know how the car works? What a strange analogy.
I like Google fine, but if we want to retain freedom to use our computers to do what we like, we need a system for searching that doesn't rely on a single source. The algorithms are for the most part public; someone needs to make a peer-to-peer search engine.
De-oxygenated hemoglobin isn't magnetic, but oxygenated hemoglobic is paramagnetic. That's why fMRI works. fMRI is a clever technique using the same MRI technology used for imaging, but tuned to see changes in blood oxygen concentration. It's used to estimate brain activity, and also to detect poor circulation in the heart.
One of the major troubles with removing a tumour is that a small amount will remain, and the cancer will regrow. Early surgery had less of this trouble, because their hygiene wasn't so great, so during the surgery the patient was basically guaranteed to get an infection, and this would lead to fever. Since cancer cells are more sensitive to higher temperatures, the remaining few cancer cells would die of the fever, leaving the patient healthy.
So your comment is somewhat true, but not for the reason you thought.
Incidentally, I don't recommend getting a fever if you just had surgery for cancer.
Earlier, considering that Google didn't have any obvious reason to continue to be non-evil, I was considering a way around this. A distributed peer-to-peer search engine might be workable; each machine would spend a certain amount of time spidering, but not too much.
Results for each word would be distributed, so any search would get responses quickly, from machines nearby in the network. Responses from more distant machines would take longer, but at least some response would arrive quickly.
I'm afraid I don't know anything about how to implement this in detail, but other presumably do.
According to the PBS TV show NOW, your Air Force was supposed to check on aircraft off of their flight paths, and that in 2000 they had in fact escorted several planes back to where they belong. (See here.
BREITWEISER: On the morning of September 11th we had four planes drastically off their flight path transponders disconnected and the FAA procedure and protocol to notify NORAD and for NORAD to scramble fighter jets were not followed. And it wasn't like they all happened in the course of an hour. What I think is very frustrating is looking back when I speak to people they say, "Well it happened in such a short span of time."
It did not happen. It happened over the course of two hours. You're telling me over the course of two hours Andrews Air Force Base in the Washington, DC area which houses F-16s which fly cover for Air Force One could not get a plane up in the air to cover the Pentagon?
Why hasn't this hit your mass media? It's incomprehensible.
Re:Nice to see the technology is catching up...
on
Robot Sales Are Exploding
·
· Score: 2, Interesting
But robots are getting pretty good at recognizing
objects, so there is hope that while mowing the
lawn they won't mutilate your pets.
Perhaps they won't mutilate your pet, but it won't be because they recognise them. Vision systems are expensive, and robotic lawnmowers don't have them. They basically have a wire delimiting the perimiter, and the wander inside. I estimated that a huge speed improvement could be had by knowing where in the map the robot is, and always trying to go someplace new (see a few things), but even that wouldn't be cheap.
Building a "sophisticated map of geometry" is impossible with current technology, and certainly isn't the way humans work. Don't you think it would be done if it was easy?
On one co-op work term one of the other students was building a mobile robot for a factory; it would bring parts from one area to another, driving using vision. It's possible, but the thing cost roughly $20k, and we were losing tons of money anyway (it was a research project too).
Unless these spots are particularly difficult to identify, someone need only write a filter to detect them and fill in the offending space, possibly with the average of the previous and next frame.
Not-so stupid little robots
on
AI Going Nowhere?
·
· Score: 2, Interesting
I personally built and programmed one of these "stupid little robots"; it's a wheelchair programmed to navigate in an office environment, using vision to determine where in the office it is. Nobody asserts that it can "reason". It navigates using a collection of local effects, in much the same manner that simple creatures operate. Watch the film "Baraka" for some rather amusing examples. At one point the film shows a bunch of caterpillars, each following the scent trail of the next --- unfortunately someone flipped the first one around, so it follows the last, and the whole colony just moves around and around until they die of starvation.
I think you would be surprised how easily remarkably complex behaviours can be achieved by a collection of very simple responses. Try fiddling around with Rossum's Playhouse, and read Brooks' book Cambrian Intelligence.
This reminds me most strongly of a short story called The Machine Stops. Written in 1909, it anticipates display screens, robotic doctors, video conferencing, etc. You might find it interesting.
It seems quite clear that the shuttle is gone, and the astronauts
dead. This is very unfortunate for all involved.
What will be the consequences to the US space programme? After the
Challenger exploded, US space flight decreased considerably and very
gradually recovered. Will the same happen now? The greatest hope is
that competition with China, a nation now promising a man on the moon
within the decade, will spur the US to continue working on its various
space projects. How likely do we think this is? I'd like to hear some
ideas of the future.
Um, "idée fixe" is very easy to translate; it means "dogmatic". That's why it's funny, it's not really a pun unless you understand both English and French.
An idea I've had for some time involves using METAFONT, the font generating software for TeX. We need a program that takes a fairly high resolution bitmap of a few characters of a desired (fairly standard, serif or sans-serif) font, then fiddles with the parameters and generates a "computer modern" type font that looks like the font we're copying. Then one could use METAFONT to Postscript or True Type conversion software, if one is foolish enough not to use LaTeX.:)
It wouldn't work for cloning bizarre fonts like Comics Cartoon, but it should automatically clone most normal fonts.
You are, of course, free to use whatever computer system you like. But by
choosing non-free software such as that produced by Microsoft, you are giving
up some very important rights, such as the right to use your computer the way
you want instead of the way Microsoft wants you to.
It's true that the "average user" cares little about this. But they probably
should.
One problem with computer solutions like this is that it
is easy to believe that the result is the objective best
result, because the computer says so. But the computer
was programmed by a human; a human chose the cost
function for the various keys. By changing the cost
function (ie how long it supposedly takes to type keys
in a given location) the keyboard layout result will
be changed.
To get a layout truly optimal for humans, it would
be necessary to train humans in each new layout,
then measure their speed. Obviously this would take
too long. But it must not be forgotten that unless this
were done, the objective solution provided by the
computer is dependent on the subjective choice of
cost function.
When humans create a robot in the fashion of Rod Brook, they are training a system analogous to our own peripheral nervous system. Why force the machine to learn to walk when we can tell it how to walk from our own experience (knowledge of physics, etc).
Brooks opposes explicit models of the world. His robots don't understand physics, and their designers didn't consider physics in the design. (I know, because I have built a number of them.)
Easy way for manufacturers to fix this --- just use a low-pass filter (ie a resistor in series and a capacitor to ground). This will remove high frequencies, essentially smoothing out the signal.
They claim that 81W are waiting to be harvested from a sleeping human. This is incorrect, due to Carnot's law (a thermodynamic law). Basically if we have a heat source at Th (the body) and a heat sink at Tl (the environment) the maximum possible efficiency is
1- Tl/Th
All temperatures must be in Kelvin (or Rankine). So for a human at 37C = 310K, with an environment at room temperature 20C = 293K, the best efficiency is
1 - 293/310 = 5.5%
If they can get 3% efficiency with current materials, they're already doing extremely well. At this efficiency a sleeping human, putting of 81W of heat, can only provide
81W * 5.5% = 4.4W
of usable energy. It's true that 4.4W can power a fair bit of energy-efficient technology, but they're starting with a lot less available energy than they claimed in the article.
The greatest problem with having an accreditated program is that is requires the university to follow guidelines set by an external organization. This can be good, but when it prevents innovation and enforces scads of really stupid stuff it becomes most frustrating.
In general there is no reason to assume that my co-workers are my friends. I want to "party" with my friends, not a bunch of other geeks I see too much of anyway. There's no reason co-workers can't be friends, if they happen to be the kind of people I like, but I hardly want to feel obliged to spend more time with them.
Slashdot Mirror
To be a programmer without ever learning assembly language is like being a professional race car driver without understanding how your carburetor works.
Do you really think the driver needs to know how the car works? What a strange analogy.
I like Google fine, but if we want to retain freedom to use our computers to do what we like, we need a system for searching that doesn't rely on a single source. The algorithms are for the most part public; someone needs to make a peer-to-peer search engine.
I don't have the expertise; do you?
bush$ rm -rf /bin/laden
Well, that's sort of right.
De-oxygenated hemoglobin isn't magnetic, but oxygenated hemoglobic is paramagnetic. That's why fMRI works. fMRI is a clever technique using the same MRI technology used for imaging, but tuned to see changes in blood oxygen concentration. It's used to estimate brain activity, and also to detect poor circulation in the heart.
One of the major troubles with removing a tumour is that a small amount will remain, and the cancer will regrow. Early surgery had less of this trouble, because their hygiene wasn't so great, so during the surgery the patient was basically guaranteed to get an infection, and this would lead to fever. Since cancer cells are more sensitive to higher temperatures, the remaining few cancer cells would die of the fever, leaving the patient healthy.
So your comment is somewhat true, but not for the reason you thought.
Incidentally, I don't recommend getting a fever if you just had surgery for cancer.
Earlier, considering that Google didn't have any obvious reason to continue to be non-evil, I was considering a way around this. A distributed peer-to-peer search engine might be workable; each machine would spend a certain amount of time spidering, but not too much.
Results for each word would be distributed, so any search would get responses quickly, from machines nearby in the network. Responses from more distant machines would take longer, but at least some response would arrive quickly.
I'm afraid I don't know anything about how to implement this in detail, but other presumably do.
Your media are broken.
According to the PBS TV show NOW, your Air Force was supposed to check on aircraft off of their flight paths, and that in 2000 they had in fact escorted several planes back to where they belong. (See here.
BREITWEISER: On the morning of September 11th we had four planes drastically off their flight path transponders disconnected and the FAA procedure and protocol to notify NORAD and for NORAD to scramble fighter jets were not followed. And it wasn't like they all happened in the course of an hour. What I think is very frustrating is looking back when I speak to people they say, "Well it happened in such a short span of time."
It did not happen. It happened over the course of two hours. You're telling me over the course of two hours Andrews Air Force Base in the Washington, DC area which houses F-16s which fly cover for Air Force One could not get a plane up in the air to cover the Pentagon?
Why hasn't this hit your mass media? It's incomprehensible.
But robots are getting pretty good at recognizing objects, so there is hope that while mowing the lawn they won't mutilate your pets.
Perhaps they won't mutilate your pet, but it won't be because they recognise them. Vision systems are expensive, and robotic lawnmowers don't have them. They basically have a wire delimiting the perimiter, and the wander inside. I estimated that a huge speed improvement could be had by knowing where in the map the robot is, and always trying to go someplace new (see a few things), but even that wouldn't be cheap.
Building a "sophisticated map of geometry" is impossible with current technology, and certainly isn't the way humans work. Don't you think it would be done if it was easy?
On one co-op work term one of the other students was building a mobile robot for a factory; it would bring parts from one area to another, driving using vision. It's possible, but the thing cost roughly $20k, and we were losing tons of money anyway (it was a research project too).
Unless these spots are particularly difficult to identify, someone need only write a filter to detect them and fill in the offending space, possibly with the average of the previous and next frame.
I personally built and programmed one of these "stupid little robots"; it's a wheelchair programmed to navigate in an office environment, using vision to determine where in the office it is. Nobody asserts that it can "reason". It navigates using a collection of local effects, in much the same manner that simple creatures operate. Watch the film "Baraka" for some rather amusing examples. At one point the film shows a bunch of caterpillars, each following the scent trail of the next --- unfortunately someone flipped the first one around, so it follows the last, and the whole colony just moves around and around until they die of starvation.
I think you would be surprised how easily remarkably complex behaviours can be achieved by a collection of very simple responses. Try fiddling around with Rossum's Playhouse, and read Brooks' book Cambrian Intelligence.
This reminds me most strongly of a short story called The Machine Stops. Written in 1909, it anticipates display screens, robotic doctors, video conferencing, etc. You might find it interesting.
It seems quite clear that the shuttle is gone, and the astronauts dead. This is very unfortunate for all involved.
What will be the consequences to the US space programme? After the Challenger exploded, US space flight decreased considerably and very gradually recovered. Will the same happen now? The greatest hope is that competition with China, a nation now promising a man on the moon within the decade, will spur the US to continue working on its various space projects. How likely do we think this is? I'd like to hear some ideas of the future.
Um, "idée fixe" is very easy to translate; it means "dogmatic". That's why it's funny, it's not really a pun unless you understand both English and French.
Applying the Fresnel lense hack discussed some time ago could make the screens merge more successfully.
Can be found in the memo here.
An idea I've had for some time involves using METAFONT, the font generating software for TeX. We need a program that takes a fairly high resolution bitmap of a few characters of a desired (fairly standard, serif or sans-serif) font, then fiddles with the parameters and generates a "computer modern" type font that looks like the font we're copying. Then one could use METAFONT to Postscript or True Type conversion software, if one is foolish enough not to use LaTeX. :)
It wouldn't work for cloning bizarre fonts like Comics Cartoon, but it should automatically clone most normal fonts.
You are, of course, free to use whatever computer system you like. But by choosing non-free software such as that produced by Microsoft, you are giving up some very important rights, such as the right to use your computer the way you want instead of the way Microsoft wants you to.
It's true that the "average user" cares little about this. But they probably should.
One problem with computer solutions like this is that it is easy to believe that the result is the objective best result, because the computer says so. But the computer was programmed by a human; a human chose the cost function for the various keys. By changing the cost function (ie how long it supposedly takes to type keys in a given location) the keyboard layout result will be changed.
To get a layout truly optimal for humans, it would be necessary to train humans in each new layout, then measure their speed. Obviously this would take too long. But it must not be forgotten that unless this were done, the objective solution provided by the computer is dependent on the subjective choice of cost function.
When humans create a robot in the fashion of Rod Brook, they are training a system analogous to our own peripheral nervous system. Why force the machine to learn to walk when we can tell it how to walk from our own experience (knowledge of physics, etc).
Brooks opposes explicit models of the world. His robots don't understand physics, and their designers didn't consider physics in the design. (I know, because I have built a number of them.)
Easy way for manufacturers to fix this --- just use a low-pass filter (ie a resistor in series and a capacitor to ground). This will remove high frequencies, essentially smoothing out the signal.
R
in >--\/\/\---+---> out
|
=C
|
GND
In case this still needs to be emphasized, the Sun Enterprise 5500 (a decently fast machine :P ) has CPUs at only 464 MHz.
I have a Palm III; it's about 3.5 years old, and I don't see any reason to replace it anytime soon.
They claim that 81W are waiting to be harvested from a sleeping human. This is incorrect, due to Carnot's law (a thermodynamic law). Basically if we have a heat source at Th (the body) and a heat sink at Tl (the environment) the maximum possible efficiency is
1- Tl/Th
All temperatures must be in Kelvin (or Rankine). So for a human at 37C = 310K, with an environment at room temperature 20C = 293K, the best efficiency is
1 - 293/310 = 5.5%
If they can get 3% efficiency with current materials, they're already doing extremely well. At this efficiency a sleeping human, putting of 81W of heat, can only provide
81W * 5.5% = 4.4W
of usable energy. It's true that 4.4W can power a fair bit of energy-efficient technology, but they're starting with a lot less available energy than they claimed in the article.
The greatest problem with having an accreditated program is that is requires the university to follow guidelines set by an external organization. This can be good, but when it prevents innovation and enforces scads of really stupid stuff it becomes most frustrating.
In general there is no reason to assume that my co-workers are my friends. I want to "party" with my friends, not a bunch of other geeks I see too much of anyway. There's no reason co-workers can't be friends, if they happen to be the kind of people I like, but I hardly want to feel obliged to spend more time with them.