> I think it is safe to conclude that most scientists working on AI projects would try to replicate human sentience either intentionally or unintentially.
Any AI researcher who claimed to be trying to replicate human sentience at this date would be laughed at.
> Is the digital creationists, who'll tell us that Computer Science is an atheist lie
Their leading lights have been saying for years that genetic algorithms don't "really" work. William Dembski insists that we program the desired answers into them, though his inclusion of neural networks under the rubric "genetic algorithms" gives us a hint regarding how much he knows about the subject matter.
> There was an article in a 1975 edition of Newsweek where scientists were sure of a global cooling. Now it's a global warming?
Yes, ~30 years ago scientists had just discovered the astronomical drivers of periodic climate change and extrapolated them to show that we should start sliding into another ice age RSN. That was before we had a good grip on what anthropogenic effects were in play.
These days at least some researchers are saying that the anthropogenic effects have actually staved off the onset of the predicted ice age, but unfortunately overcompensated for it, with the result that we're now getting a warm-up when the planet would have been getting a cool-down without us.
> I think it's foolish to think that us humans can have such an impact of the climate.
Is it? The oceans certainly aren't so big that we can't make fish dangerous to eat due to mercury pollution.
> I'd hedge my bets that volcanic eruptions and other natural occurrences play a far more significant role that cars and buses.
I don't know whether that's true or not, but you must understand that what we generate is in addition to whatever nature is doing. The world is either warming up faster or cooling down slower than it would be without us.
> So what source do you cite that says Christians think it's okay to ignore global warming because of the Second Coming? I'd like to see that article.
> The article has some interesting plots; look it up if you get a chance.
Here are two of the plots showing the unexpected reversals in trends for CO2 and CH4. (The one on the left expects methane to more or less track solar radiation, and certainly not to make a u-turn like it does.)
Unfortunately, I can't find an on-line version of the summary plot for temperature vs. expected temperature over the past 10K years, so you may still find it worth reading the magazine article.
Not everyone agrees with Ruddiman, of course. Here is a discussion thread at RealClimate, revealing a range of views about his proposals.
>>> The laws of chance dictate that the generators should churn out equal numbers of ones and zeros - which would be represented by a nearly flat line on the graph.
>> No, the laws of chance do not say any such thing. In fact, the laws of chance say exactly the opposite. If you have two choices chosen at random over a series (a 1 and a 0; or heads and tails on a coin), there is a high probability that one of the choices will be chosen a significantly higher number of times than the other. Over time, the percentage disparity will decrease to near zero, but the total numerical disparity is likely to increase.
> Except that statistics does show that over enough time the series will converge into equal numbers. It may take a million times, or ten million, but eventually you'll end up with almost exactly equal number of ones and zeros.
Presumably an infinite run would cross the balance point an infinite number of times, but that's not the point. Consider this subtlety:
Suppose you have a "fair" bit generator and you plan to let it run for a month. What's the expected value for the number of excess heads at the end of the month? Yes, zero. But suppose you run it for a week and notice that it now has, say, 100 excess heads. What is the expected value of excess heads between now and the end of the month?
And what is now the expected value for the number of excess heads over the course of the entire experiment? [The answer is printed up side down, below.]
Man descended from monkeys? Please. I think
you need to delve into the sources of the material you so unquestioningly swallowed before spouting trash about the history of man as a species.
Amazing... smart enough to do HTML markup, yet still rejects evolution.
> The Earth's 'normal' temperature isn't what we are used to anyway. Our civilisation has developed entirely in the aftermath of an ice age, and the Earth is still warming up after that.
I don't know what the concensus of scientists is on that, but I've read several articles lately that say we would already be freezing up again, if not for anthropogenic global warming. The problem is, we're warming things up too much, so in additon to neutralizing the onset of an ice age we are actually warming things up compared to what we had in the Neolithic.
By chance there's an article by William F. Ruddiman in the March Scientific American (arrived yesterday). His position is that you can model the long term fluctuations of temperature, CO2, and CH4 on the basis of several astronomical cycles, but something has gone awry in the past 8,000 years. Apparently early agriculture and the associated deforestation started driving the CO2 up about 8,000 years ago, and the invention of wet rice farming started driving CH4 up about 5,000 years ago. Each had been declining on the curve predicted by the astronomical cycles up until then, but suddenly started increasing when the should have kept on decreasing. (The article has some interesting plots; look it up if you get a chance.)
The astronomical cycles also predict that reglaciation should have started about 5,000 years ago, but instead the temperature remained essentially flat from then until the start of the Industrial Revolution. (The global warming increased as agriculture spread, fortuitously keeping temperature flat when it should have been dropping - until the Industrial Revolution kicked in.)
Thus at the start of Industrial Revolution things were already warmer than we had any right to expect, and then we started really driving it up from there. Regarding the present delta between actual temperature and expected temperature, Ruddiman attributes about half of it to historical agriculture and half to the Industrial Revolution, though like most other scientists he expects the I.R. component to keep going up (until we run out of cheap fossil fuels).
In a side bar he makes an interesting suggestion that the major cooling periods of the past 2000 years have followed plagues and depopulation of the Americas, both resulting in farmland reverting to forest (a CO2 sink). Frankly his graph for this effect doesn't look as convincing to me as the ones supporting his main thesis, but perhaps we'll be hearing more for or agains the idea in the future.
> an accurate forcast for two days out in Oklahoma [...] then I will believe them.
Have you considered that it might be easier to predict long-term trends in the global average temperature than to predict the detailed short-term fluctuations in a single location?
> In other words, they look at the data after something has happened searching for a "spike" that will almost certainly be there.
To give an illustration of one aspect of the problem you mention:
At the EGG Story page, scroll down and look at the plots labeled "Cumulative Deviation (Random Walk)", "New Years, 1998", "Pope in Holy Land". In these plots the smooth curve represents the 95% confidence bound on how far the deviation can be expected to go by pure chance. (I'm assuming their calculations are correct.)
Notice that in all cases the curve and the data plot both start at t=0, y=0, which I will call the "zero point" for the plot. Now consider the effect of the specific choice of t=0. Look at the first plot mentioned above, the "Cumulative Deviation (Random Walk)" plot, and notice that the data drops down to y=0 just a bit before t=300. Suppose you scrolled the data leftward until your zero point was at that point just before t=300; I pick this point because it has the same y as the original zero point, so nothing changes on the y axis: the boundary curve doesn't change at all, but the data is shifted leftward.
Hey! This random walk now has a sudden upward trend at t=0 (formerly t=~300), and the deviation rides above the boundary line for about 100 time steps. But wait - there's more! We can do the same think if we pick the plot's original t=700 or so, though with a slightly less impressive jump above the boundary line. Or we can get a really nice peak if we move our zero point to t=600 or so, and re-zero the data on the y axis so that the new zero point has x=0, y=0.
I can create three "significant" indicators in their example 10-minute random walk simply by cherry-picking the starting point.
How many do you think I could create if I had a free rein to pick anwhere in the previous four days of data?
The "significance" of the result critically depends on where you put your t=0 in the data stream. So go back and look at the other two plots, for the papal visit and the New Year's celebration. What if you used t=3 for your zero point on the New Year's analysis?
Re the papal visit, you might think the Pope's schedule pins down the time of interest so that we don't have any option on where to place the zero point. Well, be that as it may, whoever generated the plot did cherrypick the zero point. The schedule linked right above shows that it was actually a seven day trip; they didn't count the day the Pope left Italy and started his visit to Jordan. But what would the plot look like if they had started 24 hours earlier? (Not a rhetorical question: we don't have the data.) What's the "right" time to pick for this plot's zero point? When the Pope left Italy? When he arrived in Jordan? When he arrived in Israel? When the media coverage ramped up? Should it start a a particular how of the day? What time zone?
This is data cherrypicking of the crassest sort. The 75 scientists should be ashamed of themselves.
> Although many would consider the project's aims to be little more than fools' gold, it has still attracted a roster of 75 respected scientists from 41 different nations.
To put that into perspective, creationists claim to have 100 scientists in the fold.
> Just because something sounds crazy, doesn't mean it is.
Yeah, we know: they laughed at $SCIENTIST... but they also laughed at $CLOWN.
> Now watch microsoft drop that price from 200$ to 10$....
I don't know about that. I called a press conference and announced that I was going to play Doom instead of Age of Empires, and I didn't get one red cent out of the cheap bastards.
> For example, if they didn't properly sign out the data and disks that they were borrowing, then they would be responsible for a mistake like this even if they didn't lose anything.
Failure to sign out for non-existing disks? I suppose we're all guilty of that.
Slashdot Mirror
> I think it is safe to conclude that most scientists working on AI projects would try to replicate human sentience either intentionally or unintentially.
Any AI researcher who claimed to be trying to replicate human sentience at this date would be laughed at.
> Is the digital creationists, who'll tell us that Computer Science is an atheist lie
Their leading lights have been saying for years that genetic algorithms don't "really" work. William Dembski insists that we program the desired answers into them, though his inclusion of neural networks under the rubric "genetic algorithms" gives us a hint regarding how much he knows about the subject matter.
> There was an article in a 1975 edition of Newsweek where scientists were sure of a global cooling. Now it's a global warming?
Yes, ~30 years ago scientists had just discovered the astronomical drivers of periodic climate change and extrapolated them to show that we should start sliding into another ice age RSN. That was before we had a good grip on what anthropogenic effects were in play.
These days at least some researchers are saying that the anthropogenic effects have actually staved off the onset of the predicted ice age, but unfortunately overcompensated for it, with the result that we're now getting a warm-up when the planet would have been getting a cool-down without us.
I posted a summary of a very recent article on that earlier today.
> I think it's foolish to think that us humans can have such an impact of the climate.
Is it? The oceans certainly aren't so big that we can't make fish dangerous to eat due to mercury pollution.
> I'd hedge my bets that volcanic eruptions and other natural occurrences play a far more significant role that cars and buses.
I don't know whether that's true or not, but you must understand that what we generate is in addition to whatever nature is doing. The world is either warming up faster or cooling down slower than it would be without us.
> So what source do you cite that says Christians think it's okay to ignore global warming because of the Second Coming? I'd like to see that article.
Reagan era Secretary of the Interior James Watt.
> The article has some interesting plots; look it up if you get a chance.
Here are two of the plots showing the unexpected reversals in trends for CO2 and CH4. (The one on the left expects methane to more or less track solar radiation, and certainly not to make a u-turn like it does.)
Unfortunately, I can't find an on-line version of the summary plot for temperature vs. expected temperature over the past 10K years, so you may still find it worth reading the magazine article.
Not everyone agrees with Ruddiman, of course. Here is a discussion thread at RealClimate, revealing a range of views about his proposals.
>>> The laws of chance dictate that the generators should churn out equal numbers of ones and zeros - which would be represented by a nearly flat line on the graph.
>> No, the laws of chance do not say any such thing. In fact, the laws of chance say exactly the opposite. If you have two choices chosen at random over a series (a 1 and a 0; or heads and tails on a coin), there is a high probability that one of the choices will be chosen a significantly higher number of times than the other. Over time, the percentage disparity will decrease to near zero, but the total numerical disparity is likely to increase.
> Except that statistics does show that over enough time the series will converge into equal numbers. It may take a million times, or ten million, but eventually you'll end up with almost exactly equal number of ones and zeros.
Presumably an infinite run would cross the balance point an infinite number of times, but that's not the point. Consider this subtlety:
Suppose you have a "fair" bit generator and you plan to let it run for a month. What's the expected value for the number of excess heads at the end of the month? Yes, zero. But suppose you run it for a week and notice that it now has, say, 100 excess heads. What is the expected value of excess heads between now and the end of the month?
And what is now the expected value for the number of excess heads over the course of the entire experiment? [The answer is printed up side down, below.]
Spoiler follows ===============
i 00l 5,Ll 'll34 40
> The Earth's 'normal' temperature isn't what we are used to anyway. Our civilisation has developed entirely in the aftermath of an ice age, and the Earth is still warming up after that.
I don't know what the concensus of scientists is on that, but I've read several articles lately that say we would already be freezing up again, if not for anthropogenic global warming. The problem is, we're warming things up too much, so in additon to neutralizing the onset of an ice age we are actually warming things up compared to what we had in the Neolithic.
By chance there's an article by William F. Ruddiman in the March Scientific American (arrived yesterday). His position is that you can model the long term fluctuations of temperature, CO2, and CH4 on the basis of several astronomical cycles, but something has gone awry in the past 8,000 years. Apparently early agriculture and the associated deforestation started driving the CO2 up about 8,000 years ago, and the invention of wet rice farming started driving CH4 up about 5,000 years ago. Each had been declining on the curve predicted by the astronomical cycles up until then, but suddenly started increasing when the should have kept on decreasing. (The article has some interesting plots; look it up if you get a chance.)
The astronomical cycles also predict that reglaciation should have started about 5,000 years ago, but instead the temperature remained essentially flat from then until the start of the Industrial Revolution. (The global warming increased as agriculture spread, fortuitously keeping temperature flat when it should have been dropping - until the Industrial Revolution kicked in.)
Thus at the start of Industrial Revolution things were already warmer than we had any right to expect, and then we started really driving it up from there. Regarding the present delta between actual temperature and expected temperature, Ruddiman attributes about half of it to historical agriculture and half to the Industrial Revolution, though like most other scientists he expects the I.R. component to keep going up (until we run out of cheap fossil fuels).
In a side bar he makes an interesting suggestion that the major cooling periods of the past 2000 years have followed plagues and depopulation of the Americas, both resulting in farmland reverting to forest (a CO2 sink). Frankly his graph for this effect doesn't look as convincing to me as the ones supporting his main thesis, but perhaps we'll be hearing more for or agains the idea in the future.
> an accurate forcast for two days out in Oklahoma [...] then I will believe them.
Have you considered that it might be easier to predict long-term trends in the global average temperature than to predict the detailed short-term fluctuations in a single location?
It will only be the hottest year on record for a year or so.
The "significance" of the result critically depends on where you put your t=0 in the data stream. So go back and look at the other two plots, for the papal visit and the New Year's celebration. What if you used t=3 for your zero point on the New Year's analysis?> In other words, they look at the data after something has happened searching for a "spike" that will almost certainly be there.
To give an illustration of one aspect of the problem you mention:
At the EGG Story page, scroll down and look at the plots labeled "Cumulative Deviation (Random Walk)", "New Years, 1998", "Pope in Holy Land". In these plots the smooth curve represents the 95% confidence bound on how far the deviation can be expected to go by pure chance. (I'm assuming their calculations are correct.)
Notice that in all cases the curve and the data plot both start at t=0, y=0, which I will call the "zero point" for the plot. Now consider the effect of the specific choice of t=0. Look at the first plot mentioned above, the "Cumulative Deviation (Random Walk)" plot, and notice that the data drops down to y=0 just a bit before t=300. Suppose you scrolled the data leftward until your zero point was at that point just before t=300; I pick this point because it has the same y as the original zero point, so nothing changes on the y axis: the boundary curve doesn't change at all, but the data is shifted leftward.
Hey! This random walk now has a sudden upward trend at t=0 (formerly t=~300), and the deviation rides above the boundary line for about 100 time steps. But wait - there's more! We can do the same think if we pick the plot's original t=700 or so, though with a slightly less impressive jump above the boundary line. Or we can get a really nice peak if we move our zero point to t=600 or so, and re-zero the data on the y axis so that the new zero point has x=0, y=0.
Re the papal visit, you might think the Pope's schedule pins down the time of interest so that we don't have any option on where to place the zero point. Well, be that as it may, whoever generated the plot did cherrypick the zero point. The schedule linked right above shows that it was actually a seven day trip; they didn't count the day the Pope left Italy and started his visit to Jordan. But what would the plot look like if they had started 24 hours earlier? (Not a rhetorical question: we don't have the data.) What's the "right" time to pick for this plot's zero point? When the Pope left Italy? When he arrived in Jordan? When he arrived in Israel? When the media coverage ramped up? Should it start a a particular how of the day? What time zone?
This is data cherrypicking of the crassest sort. The 75 scientists should be ashamed of themselves.
> Although many would consider the project's aims to be little more than fools' gold, it has still attracted a roster of 75 respected scientists from 41 different nations.
To put that into perspective, creationists claim to have 100 scientists in the fold.
> Just because something sounds crazy, doesn't mean it is.
Yeah, we know: they laughed at $SCIENTIST... but they also laughed at $CLOWN.
> Enterprise is by far the worst Star Trek series of all time, why save it?
They want to postpone the start of the next series, which will be even worse.
Is this story an ad for an ad?
> Windows XP, Microsoft Office, and all the rest, these programs actually work.
Get an account, Bill.
> Now watch microsoft drop that price from 200$ to 10$....
I don't know about that. I called a press conference and announced that I was going to play Doom instead of Age of Empires, and I didn't get one red cent out of the cheap bastards.
Just keep two copies of the GPL on your system.
> For example, if they didn't properly sign out the data and disks that they were borrowing, then they would be responsible for a mistake like this even if they didn't lose anything.
Failure to sign out for non-existing disks? I suppose we're all guilty of that.
Think of it as paranoia in action.
> Never read Snow Crash, but the proper pluralization of cherub is cherubim. (::seraph:seraphim::nephil:nephilim, etc.)
Ah, see and die. Check out the Wikipedia article on harmful sensation motif.
> The country is divided by redneck inbred fucktards that do our manual labor, and intellectual blowhards who bullshit constantly.
I don't know whether to be offended by that remark, or offended by that remark. I guess it depends on which group I belong in...
etc.> he probably had what Johnny Cash called "ring of fire"
Actually, I think Johnny was singing about eating p*ssy:
> Honestly, do you people think that *everyone* in the State Department are drooling idiots??
Uh... yes.
> no, wait, I run linux. Wonder what's on TV tonight?
Switch to Gentoo and you'll have something to do tonight.
> > "A vulnerability is not a vulnerability till somebody discovers it..."
> Huh?
Sir Lancelot: "I hate to go into battle with this big f*ing hole in my chainmail, but fortunately my tabard will hide it."