The world as you know it changes every three months.
It's a reflection of the fact that each human's understanding of the universe depends on back-testing his current understanding with his understanding of the history, and that it will be invalidated by events that could not have been predicted that add up to a gross revision of the model fairly regularly.
And human brains are uniquely designed to recognize and compare these patterns in gross.
Human societies are as malleable as they are varied. (Because that's how they got to be so varied, see?)
You might think you're creating a predictive model, but it only works to predict within those facets of society for which reality has not yet invalidated the model.
A similar problem exists in using back-testing to tune models to predict the stock markets. It's succinctly summed up by the old brokerage saw:
Past results are no guarantee of future performance.
Which is to say, all "technical trading" is as good as voodoo.
The climate may be more tractable, as it hasn't as yet involved control by something as truly random as a human. But the Global Warming argument indicates that the more paranoid among us at least are finding evidence that weakly correlates human activity with climatic change.
But I still don't think the people doing this particular modelling are using a fine-grained enough model, and are likely rushing to steal cycles from projects that are producing viable results.
P.S. It's not just Nextel. This happens with Moto and Samsung and Nokia and Ericsson phones, in my personal experience. It's the phone bursting back after receiving a connection request. If you're in a crummy cell, it's worse, because the phone has to stay at full xmit power all the time.
A DSP is just a CPU with one or twelve little two-step and array-math hacks in it. Any CPU that's 2X faster in FLOPs can do the same thing with ordinary arithmetic code.
There are lots of new CPUs that are faster than lots of 5-year-old DSPs.
--Blair "But then Microsoft puts the code in a directory somewhere under C:\Windows and kills the market."
That sounds a little better. I did go to their website, and saw that they were going to use one of their four models, but I didn't dig farther to see that the journalists (as per usual) didn't understand what they were copying into their notebooks.
But what the researchers should be doing first is back-testing by using the first 25 years as calibration and the second 25 as a check on the extrapolation. Then doing it the other way around. Or maybe the distributed software does that, and all the permutations in-between.
At any rate, where it should fall on its ass is in the prediction of weather that actually makes a difference: hurricanes and tornadoes, which have crucial features that won't be well modeled, if at all, by the large differential boxes they selected. It will also run afoul of interference from random volcanic eruptions on a Pinatubo-Mount St. Helens ashfall scale, which happen on a decade or so time scale, the timing and location of which would be critical to the rest of the test run.
So I'm going to stick with my attitude that this is a tragic waste of CPU cycles that might actually go towards developing a drug that might actually save a life.
--Blair
P.S. SETI is likewise a waste; if we do hear a beep in the darkness, our only logical reaction will be to band together 6 billion of us as one to build the biggest, nastiest zero-time-of-flight weapon we can create, then hunker down in the sweaty dark to wait to fire it. Anyone coming that far is going to be wanting to make a buck off of it, taking chunks of the planet or slaves, and they're going to be ready for casual resistance.
They're starting with different initial conditions and hoping that some subset results in 50 years of weather?
Shouldn't they use the last 50 years of weather as initial conditions and vary parameters of the model instead?
What they're doing is like flipping an imaginary coin 500 times hoping to match the first 250 flips of a real coin to predict the the last 250 flips (albeit in a system with non-independent trials). But then they're taking those 500 flips and matching the first 250 to weather reports (might as well be coin flips) and then imagining the next 250 flips will approximate the future weather reports. What they need to do is fix the initial conditions and modify the model (coin flips vs. rolls of the die vs. LCRNG, etc.) to find a model that approximates the dynamics of the system.
Am I making sense here? How are these bozos not just going to apply their effective innumeracy to waste a few trillion CPU hours that could otherwise have been used to do protein folding or cancer-killing molecule matching?
>I've been waiting for Last Dangerous Visions for over 25 years, but Harlan has never released it, for reasons that he has never explained.
Harlan has explained it, and Priest's essay explained Harlan's explanation, but of course Harlan's explanation didn't really explain it in the first place.
It's been, christ, a decade? since I read Last Deadloss Visions, but the gist of Harlan's excuse was: Harlan procrastinated for a decade or so (and either I forget why or he never explained it), then decided the material wasn't Dangerous any more, and doesn't want to publish it.
It's an explanation that lacks hydroponic containment when you realize that most of us didn't read the first two books until well after they'd ceased being subversive.
--Blair "Note sardonic minusculation of popular religious icon's name. Just me, trying to get into the book."
I'm not going to be surprised if I get all of 2002 off.
I ended a lucrative contract in December, and was planning to hang out until February, but the offers coming in were lame.
Management doesn't understand that it can't improve its business' efficiency by interfering, it can only brake it.
If the job-creation projections are right, I should be getting offers at an attractive rate just about the time my sabbatical fund is down to the 6-month get-a-job-dumbass cushion.
But if it's wrong and I start looking askance at my retirement accounts, hey, framing hammers are cheap, and day-labor that speaks English is the first in the back of the truck.
--Blair "Income averaging, the Welfare of the '00s."
I bought a Compaq notebook in '96, and hated the way it tried to lord over Win95.
I bought Compaq desktops in '97 for work, and couldn't get common hardware to work under WinNT.
I haven't given Compaq a nickel since. The only good thing about that experience is it sent me to the white-box market, which I've discovered is really cool and easy, because they don't mistake corporate inveigling for enhancement.
I think the author of the article was just taking someone's word for that. Any credible scientist who would pronounce any sort of opinion on this device without once having looked inside the CPU case is not a credible scientist.
Our legal system is exactly as screwed up as people think if people realize it often consists of as one overworked, possibly incompetent judge being presented "facts" by two lawyers of varying degrees of competence.
The amount of randomness that adds to the system is anathema to justice.
Slashdot Mirror
Gilroy?
Is the garlic okay?
How about the motorcycles?
--Blair
"The only other thing they make there is trouble for politicians wanting to create a freeway to the east..."
Law can't stop crime. Law can only direct authority to apprehend and punish criminals.
Just thought I'd mention that.
--Blair
Have you ever heard this phrase?
The world as you know it changes every three months.
It's a reflection of the fact that each human's understanding of the universe depends on back-testing his current understanding with his understanding of the history, and that it will be invalidated by events that could not have been predicted that add up to a gross revision of the model fairly regularly.
And human brains are uniquely designed to recognize and compare these patterns in gross.
Human societies are as malleable as they are varied. (Because that's how they got to be so varied, see?)
You might think you're creating a predictive model, but it only works to predict within those facets of society for which reality has not yet invalidated the model.
A similar problem exists in using back-testing to tune models to predict the stock markets. It's succinctly summed up by the old brokerage saw:
Past results are no guarantee of future performance.
Which is to say, all "technical trading" is as good as voodoo.
The climate may be more tractable, as it hasn't as yet involved control by something as truly random as a human. But the Global Warming argument indicates that the more paranoid among us at least are finding evidence that weakly correlates human activity with climatic change.
But I still don't think the people doing this particular modelling are using a fine-grained enough model, and are likely rushing to steal cycles from projects that are producing viable results.
--Blair
Anyone with the density handy want to cobble up the Schwarzchild radius of one of these puppies and see if it fits inside?
In case you need it,
r = 2 G m/c^2.
c = 2.998e+08 m/s
G = 6.672e-11 N m^2/kg^2
--Blair
This is an interesting answer to arguments that online music sharing is nothing but theft.
Didn't you say that about the post announcing Vanilla Coke?
--Blair
We should have leased the rights to a TLD, and charged a royalty of 50%.
--Blair
chit-chit-chit-chit-chit
Telephone!
BLEBLEBLEBLEBLEBLEEEEEP!
Hello?
How to be your own Radar O'Reilly.
--Blair
P.S. It's not just Nextel. This happens with Moto and Samsung and Nokia and Ericsson phones, in my personal experience. It's the phone bursting back after receiving a connection request. If you're in a crummy cell, it's worse, because the phone has to stay at full xmit power all the time.
A DSP is just a CPU with one or twelve little two-step and array-math hacks in it. Any CPU that's 2X faster in FLOPs can do the same thing with ordinary arithmetic code.
There are lots of new CPUs that are faster than lots of 5-year-old DSPs.
--Blair
"But then Microsoft puts the code in a directory somewhere under C:\Windows and kills the market."
That sounds a little better. I did go to their website, and saw that they were going to use one of their four models, but I didn't dig farther to see that the journalists (as per usual) didn't understand what they were copying into their notebooks.
But what the researchers should be doing first is back-testing by using the first 25 years as calibration and the second 25 as a check on the extrapolation. Then doing it the other way around. Or maybe the distributed software does that, and all the permutations in-between.
At any rate, where it should fall on its ass is in the prediction of weather that actually makes a difference: hurricanes and tornadoes, which have crucial features that won't be well modeled, if at all, by the large differential boxes they selected. It will also run afoul of interference from random volcanic eruptions on a Pinatubo-Mount St. Helens ashfall scale, which happen on a decade or so time scale, the timing and location of which would be critical to the rest of the test run.
So I'm going to stick with my attitude that this is a tragic waste of CPU cycles that might actually go towards developing a drug that might actually save a life.
--Blair
P.S. SETI is likewise a waste; if we do hear a beep in the darkness, our only logical reaction will be to band together 6 billion of us as one to build the biggest, nastiest zero-time-of-flight weapon we can create, then hunker down in the sweaty dark to wait to fire it. Anyone coming that far is going to be wanting to make a buck off of it, taking chunks of the planet or slaves, and they're going to be ready for casual resistance.
Two words:
DEC Rainbow.
Hewpaq only got the parts of DEC that Intel hadn't already bought.
--Blair
They're starting with different initial conditions and hoping that some subset results in 50 years of weather?
Shouldn't they use the last 50 years of weather as initial conditions and vary parameters of the model instead?
What they're doing is like flipping an imaginary coin 500 times hoping to match the first 250 flips of a real coin to predict the the last 250 flips (albeit in a system with non-independent trials). But then they're taking those 500 flips and matching the first 250 to weather reports (might as well be coin flips) and then imagining the next 250 flips will approximate the future weather reports. What they need to do is fix the initial conditions and modify the model (coin flips vs. rolls of the die vs. LCRNG, etc.) to find a model that approximates the dynamics of the system.
Am I making sense here? How are these bozos not just going to apply their effective innumeracy to waste a few trillion CPU hours that could otherwise have been used to do protein folding or cancer-killing molecule matching?
--Blair
>I've been waiting for Last Dangerous Visions for over 25 years, but Harlan has never released it, for reasons that he has never explained.
Harlan has explained it, and Priest's essay explained Harlan's explanation, but of course Harlan's explanation didn't really explain it in the first place.
It's been, christ, a decade? since I read Last Deadloss Visions, but the gist of Harlan's excuse was: Harlan procrastinated for a decade or so (and either I forget why or he never explained it), then decided the material wasn't Dangerous any more, and doesn't want to publish it.
It's an explanation that lacks hydroponic containment when you realize that most of us didn't read the first two books until well after they'd ceased being subversive.
--Blair
"Note sardonic minusculation of popular religious icon's name. Just me, trying to get into the book."
I'm not going to be surprised if I get all of 2002 off.
I ended a lucrative contract in December, and was planning to hang out until February, but the offers coming in were lame.
Management doesn't understand that it can't improve its business' efficiency by interfering, it can only brake it.
If the job-creation projections are right, I should be getting offers at an attractive rate just about the time my sabbatical fund is down to the 6-month get-a-job-dumbass cushion.
But if it's wrong and I start looking askance at my retirement accounts, hey, framing hammers are cheap, and day-labor that speaks English is the first in the back of the truck.
--Blair
"Income averaging, the Welfare of the '00s."
I bought a Compaq notebook in '96, and hated the way it tried to lord over Win95.
I bought Compaq desktops in '97 for work, and couldn't get common hardware to work under WinNT.
I haven't given Compaq a nickel since. The only good thing about that experience is it sent me to the white-box market, which I've discovered is really cool and easy, because they don't mistake corporate inveigling for enhancement.
--Blair
What's so funny?
I'm guessing three years, tops.
--Blair
I think the author of the article was just taking someone's word for that. Any credible scientist who would pronounce any sort of opinion on this device without once having looked inside the CPU case is not a credible scientist.
--Blair
Two decades of Hollywood horseshit is being buried under righteous indignation.
Go Stan.
--Blair
"'Nuff sed."
Watch the forecredits on those competitors' games. Most of them will credit Id Software somewhere.
--Blair
Our legal system is exactly as screwed up as people think if people realize it often consists of as one overworked, possibly incompetent judge being presented "facts" by two lawyers of varying degrees of competence.
The amount of randomness that adds to the system is anathema to justice.
--Blair
It does account for time.
Look at the last graph.
Big decline in frequency of wars in the late 1600's and early 1700's. Which ended--oh look--during the American Revolution.
Ever since then, the world has decided it can solve its problems profitably through violence.
--Blair
Science is prosaic. It's not there to be beautiful.
If you twist it to make it beautiful, you're denigrating its value as science.
Every experiment that falsifies its hypothesis is exactly as beautiful as it needs to be.
--Blair
then where do I send my check to support the plaintiffs?
Economic reality: the price of stem cells drops so much you still end up paying $5K net for lipo.
--Blair
"I'll have the Genome McMuffin and a medium RNA rings."
Most companies are massively up from their low on Sept. 21.
Apple has underperformed its prophesies financially and technologically since the mid-'80s.
--Blair