PDP-8 here, I was lucky. Dad got an 8S in a stunning surplus deal "it fell off the truck" - then later a "straight 8" - and we sneered and jeered at those later 8 bit chip junk machines. Predicting is hard...we were wrong, but the experience did serve well when I later applied for a job at DEC and they were stunned to find I already had plenty of experience with their stuff (as a then stereo repairman).
Detonation vs deflagration. Meerling is simply wrong. A strong case hasn't been required since deflagrating black powder was replaced by real detonating explosives. For nitroglycerin, TNT, PETN, HMX, RDX and so on - a baggie will do. Yes, I've worked with high explosives. Numerous examples exist of a simple block of C4 or other explosive being used with just a blasting cap. Plenty of brisance to involve the whole mess before it breaks up when you use HE with detonation rates of a few km/second.
You are lucky. I have around 400 orders this year and it happens on about every 3rd one. I paid for prime 'free shipping' and defaulting to the most expensive alternative for just one item - usually the least pressing - on an order really stinks. They probably do lose some on me as I'm out in the boonies...most things come USPS when it's "two day" which is more often "3 day" - and I know this and don't care that much. The extra 50 bucks? I care. I'm already paying for prime.
I built a LAN of things for homestead automation, it's on my website in part (most people can find me if they care). I don't need voice commands to do what I want done, so...
And I sure don't need traffic on the internet for that stuff. In the boonies, it's slow, expensive, and unreliable when things like water, heat, electricity (I'm off grid) and so on are at stake.
Surprising. While Amazon can reward careful shopping and jumping on "loss-leaders of opportunity", even prime members without one click ordering will get hosed if they don't pay attention. Often an "all prime" order I make will default one item to "same day or one day" shipping, adding >$50 to the price of a $5 thing. No sweat if you pay attention and click it back before submitting...do you have that chance with one click or Alexa?
Further, dash buttons can be a real joke. I might order something that's right now a loss leader or at least a bargain compared to driving to the store for it - but when I need it again, the item on the dash button has doubled in price from the original loss leader. All this is common in brick and mortar, which is why you pay attention. .
The evil is that some of these newer forms of transaction remove the ability to pay attention if you were so inclined, and if you weren't, fleece you on a whole new level.
In short - we'd make up an enemy if required to keep the status quo in power.
I think HL Mencken had something to say about it:
"Civilization, in fact, grows more and more maudlin and hysterical; especially under democracy it tends to degenerate into a mere combat of crazes; the whole aim of practical politics is to keep the populace alarmed (and hence clamorous to be led to safety) by menacing it with an endless series of hobgoblins, most of them imaginary."
Yep, that describes our (and most other) governments. Russia's done the same using us as the boogyman, North Korea...maybe it's why we don't like looking in the mirror much. It's the oldest trick in the book - hey, look, a bear is even more attention-getting than hey, look, a squirrel.
Yes, feed-forward closed solutions to recursive systems (n-body gravitational, fractals, weather, most real world systems!)
Design me an IIR filter from a frequency response to whatever accuracy I spec. That's two right off the bat.
Solve systems of nonlinear equations (see above).
Huge practical and obvious applications for these, no truly meaningful work - the entire world is a system where the next input is the last output - recursive, fractal - did we really have to wait for Mandelbrot to even begin to look at the shape of the problem at all? And all I see is bean counting and button sorting, not some decent analytical approach to solving systems like that. Else we'd have among other useful things, weather prediction, simulations of group behavior of plasmas (eg we'd have insight into how to do fusion that meant something) - it's a long, long list.
There is or isn't a law - is a BS argument. Laws are not truth or justice or wisdom, they're just laws. The article is kind of a demonstration of that. People come here because they think it'll be a better life than it was where they originated. They should consider why that might be - maybe our culture works better in some sense, and they should learn it and our language in order to be a contributing member. And sure, we should also learn from the better parts of their culture, but ditch the horrible flaws that are the reason these people left their original home in the first place. I don't see what's so hard to understand about that.
Probably shouldn't bother feeding an obvious troll, even one with a low ID.
You think criteria of being alive helps with fraud? The dead never vote early and often? There's no medicate fraud? It's impossible or even difficult to forge more than one identity? Do you really even think that ease of enforcement, if not imaginary as this would be, is gonna help our corrupt system work?
I'm afraid you don't know a lot about humans and what they get up to. What it would take to stop that is a police state so severe the only corruption would be within the police state itself...um...oh wait.
Come up with a printing press that prints value instead of mere currency and we can talk. Till then you don't even need real shortage of anything to create crime over it, perception is plenty, and greed is always with us.
After all, why do some people not stop after a billion or few? Gimme a break, ignoring human nature won't make it go away.
Just wanting something badly doesn't make it realistic.
Anytime there is some perceived gain from fraud, there will be fraud. There have been zero exceptions to that. Argument fail. If your wish comes true, who feeds the 80% less bureaucrats and entertains them to keep them out of trouble? Note, I'm not on their side, I think they are held down by the system they worship and kept out of my hair thereby. I don't want them sharing my pool.
I always think of the vast bureaucracy managing things as they are as a jobs program for people who'd be more trouble otherwise. It gives them a sense of purpose and mostly keeps them out of the hair of the real contributors. ?
You're joking that a UBI replaces everything, right? Who keeps track of everyone to make sure people get theirs, and only theirs? You think fraud goes away like magic? You think we can just cut those admin costs...and now all those guys need money too for doing nothing. (Checks for free? I'll be here all week.) Man, I want some of what you smoke. .
We could wish they'd fund more than one approach, to be sure. There's no one "wrong place" and I suspect, more than one useful place to find clues.
I have a friend at CERN and we've discussed their detector filtering algos - they need to do some reduction in data rates. The thing is, at some point in this process, it's easy to eliminate anything you weren't looking for anyway...and they know it. Technical limitations as much as anything, but also cognitive. .
"Expect the unexpected" is harder to do than most people think - if for no other reason that there's quite a lot of it.
I was just going off about math taking credit that isn't always deserved. Sometimes it is. Just not all of it.
Reading comprehension issue, my friend?
I didn't say violent crime was down. I said perceived annoyance is UP, and that people suck at evaluating risk.
The OP might have wrongly thought it was all about crime, but that's not what _I_ said.
You could hope for that. After all, a lot of problems in signal processing are solved more easily by flipping to a different domain (say, frequency instead of time), solving it there, and then converting back.
In fact, something similar is what Ed Witten was on about in joining the various (sub) string theories, so that a problem insoluble in one of them could be solved in another, then converted back.
Too bad mathematics has, after all this time, no feed-forward closed-form solution whatever to the N body gravitational problem.
Much less one with charge (Coulomb) and strong force.
Or for that matter, any recursive/fractal problem - weather is a simple example. You just have to perturb endlessly.
We'd have fusion if they did.
Now, trying to simulate a 10^20 something body problem as a particle in grid is heat death of the universe difficult.
Math isn't the queen of all the science we need. It was just lucky for awhile.
Burning mod points here but...
The NN might indeed be unbiased. So what? All practical systems do some preprocessing to cut the data rate to something reasonable, not just raw pixels at random orientations - you've all see the pics of faces with polygons drawn over them, right?
Guess where that comes from, and how it was tuned? Decisions about how to data reduce the input - create bias.
What helps tell say, white faces apart and white from black (to vastly oversimplify, not trying to exclude any race etc) - might stink at telling black faces from one another. "They all look alike to me!"
And this is where bias creeps in, alongside numerous easy to make statistical errors; See, for example Timothy Master's Practical Neural Network Recipes in C++ for extensive discussion of this and what I'm saying is beyond obvious.
?
It's very easy even to bias training samples by accident. One classic failure was due to data preparation unconsciously cropping pictures in one class just a little differently than the other - and the network worked great for that, and stank for all else.
Kendall's mistake is a common one from those who haven't actually done this or written the code from the ground up. Easy libraries will do that to ya. Yes, within limits, the net is unbiased. Now go make a perfect training set - goodluckwiththat. May Bayes work for you! Except when it doesn't.
Agree, BUT. The threshold for "annoying" is way down. Things people used to ignore are now hate speech and other "triggers". And people suck at calculating risk anyway, as the war on terror shows, the fear of flying shows...people are afraid of what's least likely to harm them! Or, even less flattering, are unwilling to face the fact we failed to keep our government honest by laziness. !
And increasingly being the centers of their own universe, being annoyed is now a "violent crime" and in this day of no personal ownership or responsibility for oneself, "someone else's fault" too. !
And the likelihood of this being downvoted into oblivion won't make this observation less true or less based on 50+ years of observations.
Moon has more solar flux, a smaller gravity well to boost out of, is closer and faster to communicate with both electromagnetically and physically, and we need a way station outside of earth. Mars...is kinda "out there" and attractive as an adventure, but a rescue for some sort of "oops" that threatens a colony is out of the question for mars, hard enough for the moon. !
Slashdot Mirror
Appalachia - $82/mo for 4 megabits down and one up. And almost the best I can buy - the best is 6/1.5 and $30/mo more.
PDP-8 here, I was lucky. Dad got an 8S in a stunning surplus deal "it fell off the truck" - then later a "straight 8" - and we sneered and jeered at those later 8 bit chip junk machines. Predicting is hard...we were wrong, but the experience did serve well when I later applied for a job at DEC and they were stunned to find I already had plenty of experience with their stuff (as a then stereo repairman).
Detonation vs deflagration. Meerling is simply wrong. A strong case hasn't been required since deflagrating black powder was replaced by real detonating explosives. For nitroglycerin, TNT, PETN, HMX, RDX and so on - a baggie will do. Yes, I've worked with high explosives. Numerous examples exist of a simple block of C4 or other explosive being used with just a blasting cap. Plenty of brisance to involve the whole mess before it breaks up when you use HE with detonation rates of a few km/second.
You are lucky. I have around 400 orders this year and it happens on about every 3rd one. I paid for prime 'free shipping' and defaulting to the most expensive alternative for just one item - usually the least pressing - on an order really stinks. They probably do lose some on me as I'm out in the boonies...most things come USPS when it's "two day" which is more often "3 day" - and I know this and don't care that much. The extra 50 bucks? I care. I'm already paying for prime. I built a LAN of things for homestead automation, it's on my website in part (most people can find me if they care). I don't need voice commands to do what I want done, so... And I sure don't need traffic on the internet for that stuff. In the boonies, it's slow, expensive, and unreliable when things like water, heat, electricity (I'm off grid) and so on are at stake.
Further, dash buttons can be a real joke. I might order something that's right now a loss leader or at least a bargain compared to driving to the store for it - but when I need it again, the item on the dash button has doubled in price from the original loss leader. All this is common in brick and mortar, which is why you pay attention.
.
The evil is that some of these newer forms of transaction remove the ability to pay attention if you were so inclined, and if you weren't, fleece you on a whole new level.
In short - we'd make up an enemy if required to keep the status quo in power.
I think HL Mencken had something to say about it:
"Civilization, in fact, grows more and more maudlin and hysterical; especially under democracy it tends to degenerate into a mere combat of crazes; the whole aim of practical politics is to keep the populace alarmed (and hence clamorous to be led to safety) by menacing it with an endless series of hobgoblins, most of them imaginary."
Yep, that describes our (and most other) governments. Russia's done the same using us as the boogyman, North Korea...maybe it's why we don't like looking in the mirror much. It's the oldest trick in the book - hey, look, a bear is even more attention-getting than hey, look, a squirrel.
Yes, feed-forward closed solutions to recursive systems (n-body gravitational, fractals, weather, most real world systems!)
Design me an IIR filter from a frequency response to whatever accuracy I spec. That's two right off the bat.
Solve systems of nonlinear equations (see above).
Huge practical and obvious applications for these, no truly meaningful work - the entire world is a system where the next input is the last output - recursive, fractal - did we really have to wait for Mandelbrot to even begin to look at the shape of the problem at all?
And all I see is bean counting and button sorting, not some decent analytical approach to solving systems like that. Else we'd have among other useful things, weather prediction, simulations of group behavior of plasmas (eg we'd have insight into how to do fusion that meant something) - it's a long, long list.
Damn, out of mod points. ^^^^^^^^^
Like bringing your lunch from home isn't a viable alternative?
There is or isn't a law - is a BS argument. Laws are not truth or justice or wisdom, they're just laws.
The article is kind of a demonstration of that.
People come here because they think it'll be a better life than it was where they originated. They should consider why that might be - maybe our culture works better in some sense, and they should learn it and our language in order to be a contributing member. And sure, we should also learn from the better parts of their culture, but ditch the horrible flaws that are the reason these people left their original home in the first place. I don't see what's so hard to understand about that.
What, you mean liberals and progressives are anything but? I need to lie down...
Probably shouldn't bother feeding an obvious troll, even one with a low ID.
You think criteria of being alive helps with fraud? The dead never vote early and often? There's no medicate fraud? It's impossible or even difficult to forge more than one identity? Do you really even think that ease of enforcement, if not imaginary as this would be, is gonna help our corrupt system work?
I'm afraid you don't know a lot about humans and what they get up to. What it would take to stop that is a police state so severe the only corruption would be within the police state itself...um...oh wait.
Come up with a printing press that prints value instead of mere currency and we can talk. Till then you don't even need real shortage of anything to create crime over it, perception is plenty, and greed is always with us.
After all, why do some people not stop after a billion or few? Gimme a break, ignoring human nature won't make it go away.
Just wanting something badly doesn't make it realistic.
Anytime there is some perceived gain from fraud, there will be fraud. There have been zero exceptions to that. Argument fail. If your wish comes true, who feeds the 80% less bureaucrats and entertains them to keep them out of trouble? Note, I'm not on their side, I think they are held down by the system they worship and kept out of my hair thereby. I don't want them sharing my pool.
?
You're joking that a UBI replaces everything, right? Who keeps track of everyone to make sure people get theirs, and only theirs? You think fraud goes away like magic? You think we can just cut those admin costs...and now all those guys need money too for doing nothing. (Checks for free? I'll be here all week.) Man, I want some of what you smoke.
.
At least I can spell dystopian.
.
"Expect the unexpected" is harder to do than most people think - if for no other reason that there's quite a lot of it.
I was just going off about math taking credit that isn't always deserved. Sometimes it is. Just not all of it.
Does it being accepted make it not-wrong? Can we vote on other aspects of reality and morality too? Gravity - it's a pain.
Hey, I lived through them. It'll make you stronger!
Reading comprehension issue, my friend? I didn't say violent crime was down. I said perceived annoyance is UP, and that people suck at evaluating risk. The OP might have wrongly thought it was all about crime, but that's not what _I_ said.
You could hope for that. After all, a lot of problems in signal processing are solved more easily by flipping to a different domain (say, frequency instead of time), solving it there, and then converting back.
In fact, something similar is what Ed Witten was on about in joining the various (sub) string theories, so that a problem insoluble in one of them could be solved in another, then converted back.
Too bad mathematics has, after all this time, no feed-forward closed-form solution whatever to the N body gravitational problem.
Much less one with charge (Coulomb) and strong force.
Or for that matter, any recursive/fractal problem - weather is a simple example. You just have to perturb endlessly.
We'd have fusion if they did.
Now, trying to simulate a 10^20 something body problem as a particle in grid is heat death of the universe difficult.
Math isn't the queen of all the science we need. It was just lucky for awhile.
Guess where that comes from, and how it was tuned? Decisions about how to data reduce the input - create bias.
What helps tell say, white faces apart and white from black (to vastly oversimplify, not trying to exclude any race etc) - might stink at telling black faces from one another. "They all look alike to me!"
And this is where bias creeps in, alongside numerous easy to make statistical errors; See, for example Timothy Master's Practical Neural Network Recipes in C++ for extensive discussion of this and what I'm saying is beyond obvious.
?
It's very easy even to bias training samples by accident. One classic failure was due to data preparation unconsciously cropping pictures in one class just a little differently than the other - and the network worked great for that, and stank for all else.
Kendall's mistake is a common one from those who haven't actually done this or written the code from the ground up. Easy libraries will do that to ya. Yes, within limits, the net is unbiased. Now go make a perfect training set - goodluckwiththat. May Bayes work for you! Except when it doesn't.
!
And increasingly being the centers of their own universe, being annoyed is now a "violent crime" and in this day of no personal ownership or responsibility for oneself, "someone else's fault" too.
!
And the likelihood of this being downvoted into oblivion won't make this observation less true or less based on 50+ years of observations.
Intel's attempts at IoT also come to mind. Never design in parts from a company that does this.
!
I say walk before you run.
Those who talk either don't know, or are talking their book. Tesla is the most shorted stock in the market. I'm sure you can figure out the rest...