If CO2 concentration continues to increase (as I'm sure it will) but the warming trend reversed itself, that would falsify the idea of CO2 as an overriding contributor to global warming.
My observation has been that typically, believers in CAGW will take any lack of warming trend as simply a minor setback to be ignored - they come up with some ad hoc explanation that makes the cooling somehow anomalous. With this kind of approach, CAGW can be infinitely justified (30 years of cooling? must've been a volcano. 45 years of cooling? some orbital cycle. 75 years of cooling? A spurious trend caused by solar activity.).
As for myself, it's very likely that the climate is fairly insensitive to most factors, and that natural cycles dominate on nearly every timescale we can imagine. It will be interesting, though, if the solar minimum predictions of cooling till 2050 hold up.
1) You can't assert that simply because CO2 has certain spectral properties, that therefore, as it acts in the atmosphere, it must be an overwhelming positive driver. As for an experiment that shows that some unknown effect balances out any spectral absorption contribution it could have, isn't that *exactly* what this data currently shows? That is, the theoretical model asserted that CO2 would cause one type of behavior, and for some reason, we observed something different.
You still haven't made any statement about what *observation* of temperatures, CO2, weather events, or anything else, that would falsify your hypothesis. The "heads I win, tails you lose" argument is the hallmark of religion, not science.
If you have a theory of God that says he performed this or that specific miracle, then I can falsify it, and you'll have to rewrite your theory
And *that's* exactly the point. How do you falsify a specific miracle? How do you prove that the reason why I found my keys wasn't because of Divine Intervention? What could you possibly observe that could prove that the chain of events that led up to me finding my keys *wasn't* due to a deity?
I'm asking you for any observation that will falsify the miracle of Human emissions of CO2 causing Catastrophic Global Warming. Instead of offering up any type of observation that would prove you wrong, you've instead asked me to either believe that the spectral properties of a single molecule can lead us to the conclusion that it *must* be that molecule, or that I should be able to intuit the vague definition of "insignificant" and "normal fluctuations of CO2", or that if human CO2 *doesn't* have an effect, I need to show that it is in fact countered by some other novel human effect.
Look, for evolution, falsification is simple - show me a more complex life form that precedes simpler ones. A rabbit fossil in the pre-cambrian for example. This is science.
So if I understand you correctly, you're making two assertions. One is that intelligence is defined by learning from mistakes. The second is that the human brain can simply be defined as a Bayesian filter.
Understanding that you see human intelligence as merely an implementation of a complex Bayesian filter gives me an insight as to why you would see a spam filter as intelligent.
I think your criteria probably needs some further filtering, though. For example, my spell checker starts off with a dictionary of words. When it makes a mistake, I tell it that it made a mistake, and it learns a new allowed word or spelling. I'll assert that this fits your definition of "learns on its own", even though you have to tell it that it made a mistake (just like you tell a spam filter to add something to a white list or black list). I'm not sure if you'd count a spell checker as "intelligent".
You might refine it to the following criteria - "if it learns on its own, by recognizing its own mistakes without being *told* it made a mistake, it is intelligent." This would exclude the spell checker, and probably to a great degree the spam filter.
Now, if the spam filter *wasn't told* something was "spam" or "not spam", but deduced from say, the fact that you deleted certain things, or spent more time reading certain things than others, or some other pattern recognition in your behavior (besides explicit marking of certain emails as spam), I'd agree that it was intelligent. In this case, it wouldn't be *told* it made a mistake, it would figure it out by non-trivial observation.
I think this revised criteria would include the human brain as "intelligent", for the most part, but I would be hard pressed to say that the ability to understand that you made a mistake without outside prompting is something that can be modeled by a Bayesian filter.
I'm sorry, perhaps I've misunderstood your position. Correct me if I'm wrong, but it seems as if you're arguing that your typical spam filter is intelligent because it appears intelligent, and that simply appearing intelligent is the whole point of the Turing Test, so your typical spam filter therefore has passed the Turing Test.
From my point of view, you're making a leap in rationale that cannot be justified, so I'm trying to better understand your position by asking a different question, that is, what would you consider as *not* intelligent? If you can sufficiently expound upon what you think would *not* fit the definition of artificial intelligence, then maybe you can explain your point better. Perhaps "intelligence" is simply a matter of speed (if the spam filter was slow, you'd consider it unintelligent). Or maybe your criteria is the % of false positives, or the % of false negatives.
I can only speculate wildly about what your criteria would be, though - you haven't made clear at all what you think would *not* be artificial intelligence.
Now perhaps you simply haven't thought of the question in those terms before, and don't have a ready answer, and I can understand that. But I definitely think it's something you should at least try to think about someday:)
One, there's no way you can assert that either CO2 concentration changes or temperature changes over the last few decades are unprecedented - there are certainly similar rates observed, in the historical record, just within the past 150 years, well before the industrial revolution.
Second, simply because you haven't been able to observe with sufficient accuracy, the past several hundred thousand years, doesn't mean you can take your detailed measurements of the past few decades and assert something extraordinary. Now, maybe if we had several hundred thousand years of weather station data, you could plausibly make that argument, but any comparison of modern instrumental temperatures against historical proxies is weak at best.
That all being said, what observations of CO2 and temperature over the next few decades would you consider a falsification of your hypothesis?
Let's be specific - what observations are you talking about?
1) CO2 does not contribute [positively] to global temperature - what would you *observe* to show that? Increasing CO2, all other factors held constant and no increase of temperature? (Have fun finding all other factors held constant:) )
2) Humans do not contribute enough CO2 to cause global warming - again, what would you *observe* to show that? First you have to come up with what the minimum amount necessary to cause global warming - do you have that number? What would falsify *that* number?
3) Human contributed CO2 is negated by other human factors - what *observation* would show that?
You're offering up "prove this, and then I'm wrong" - that's *not* how a falsifiable hypothesis works. I could say "prove that God doesn't exist by showing that all miracles are frauds", but that's still *not* a falsifiable hypothesis.
Put more bluntly, is there any observation of say, CO2 levels, extreme weather events, and global average temperature, that would falsify your hypothesis of Catastrophic Anthropogenic Global Warming? If we saw more CO2, less extreme weather, and lower global average temperature, over any period of time, would that be sufficient to falsify?
Give me the concise list of observations that in combination would falsify your hypothesis. Until then, you're talking religion.
Your argument reminds me of conversations I used to have about robots. You can, through the same sort of reasoning you're using, declare that essentially, anything is a robot. Not just your typical "build a car" robots, but even your microwave, or your wristwatch, or even just a single magnet.
How about this - what is the difference between a non-intelligent spam filter, and an intelligent one? If I create a spam filter that simply hardcodes a table, and it *appears* intelligent to you, is it therefore intelligent? Where would *you* draw the line?
1) plants certainly increase in yield if the previous temperature was too cold for them to grow at all; 2) there is no evidence that warmer means "less regular climate" - in fact, most of the warming has been documented to be a *narrowing* of the temperature range, rising the cools, and doing very little to the highs; 3) http://wattsupwiththat.com/2011/06/26/global-hurricane-activity-at-historical-record-lows-new-paper/
1) climate has always changed - you've got to show that this *particular* change is going to cause your *particular* harms 2) you've got to deal with the particular benefits that come with this *particular* change in order to do a proper analysis (more flooding might be bad here and there, but increased food production and crop yields could easily provide more benefit than the cost of occasional floods) 3) you've got to deal with the evidence that shows that while global temperature has increased, natural disasters have either had no change in incidence, or a positive change in incidence.
So what do you do when the risk assessment that is being made is in fact 100% backwards?
Tangential question - if you became thoroughly convinced that the data showed a safer, more benign planet when things were warmer, would you then begin encouraging CO2 emissions as a mandatory responsibility of world governments?
The alternative hypothesis is the null hypothesis - global warming and global cooling and even global temperature stability are natural phenomena. More specifically, global climate changes occur for the same reasons they did for the millions of years before humanity even existed - we might not be able to deconstruct every last variable, of course, but we don't need to have any specific cause in mind just because we've refuted the "CO2 drives temperature" hypothesis.
Or do you deny that climate ever changed before humans came along?:)
Or more importantly, what observations would convince you that the risk *wasn't* high? Historical incidences of higher temperatures and a more vibrant biosphere hospitable to mankind? Contemporary incidences of more harm from cold weather than harm from hot weather? Some future decrease in harm per capita during a period of increasing warming?
The problem we have is that the precautionary principle can fuck things up royally if you get it wrong. Take DDT - safe, non-toxic to humans and animals, and responsible for eliminating the scourge of malaria from countries world wide. Someone posited that there *might* be some risk associated with it, and *just to be safe*, it's been banned. As a consequence, malaria has raged throughout Africa for decades, killing tens of millions of babies.
Based on what I've read on the subject, I am unconvinced that there is any appreciable risk, and the relatively high costs and limited benefits of moving to alternative fuels and capping emissions is going to have tremendously negative unintended consequences.
I'll add another corollary - in order to be qualified to disregard the idea of a falsifiable hypothesis, and to be allowed to redefine the null hypothesis, one simply must be: A Global Warming Activist; A Member of some Environmentalist Organization; Democrat, or "politically correct".
I find it *hilarious* that the same people who are getting down on someone for believing in a non-falsifiable hypothesis (Intelligent Design), cannot see that their own religious belief of Catastrophic Anthropogenic Global Warming suffers from *the exact same flaw*. It's like watching a monotheist talk about how irrational it is to believe in all the ancient Greek Gods...hypocrisy, it seems, is a trait which knows no bounds.
For some "scientist" to claim that Intelligent Design is a science (hint: it cannot be falsified so it is not) does call into question all their other "scientific" claims.
So isn't it true also that for some "scientist" to claim that AGW is a science (hint: it cannot be falsified so it is not), calls into question all of their other "scientific" claims?
Please, concisely state your falsifiable hypothesis of AGW, and we'll start from there.
Any claim that CO2 *is* causing global temperature increase is an unsubstantiated claim. It may be a favored hypothesis by certain scientists who have a specific political agenda, but let's not pretend for a moment that this claim has been substantiated in any way shape or form.
Before you can substantiate such a claim, you need a falsifiable hypothesis, and thus far, I've heard absolutely zero instances of falsifiable hypothesis from the alarmist camp.
Warmer probably has more positive effects than negative effects. I can certainly imagine how someone who truly believed warmer was going to be on the whole worse, this kind of thing might be particularly scary, but what if the opposite was true?
To play the science game for a moment, how would you falsify the hypothesis "warmer is better" (or even "colder is better")?
I suppose it could also be true that overall, warmer or colder just doesn't matter, and everything sort of evens out, but we might as well start with some sort of procedure for determining what is better or worse for humanity (or even all life on the planet) as a whole.
Computerized spam filters work better than a smart human being because of pure volume, not because of any sort of intelligence.
A spam filter will never realize that you really are interested in a story from Huff Post because it happens to be about one of your relatives, or heck *about you*. It has no inherent understanding about what makes something desirable, or undesirable - it's simply a collection of hard coded filters with some fuzzy matching. It may *appear* intelligent to you, but I'm afraid that's a false positive:)
Do you think your spam filter could look at three paragraphs written by you, and three paragraphs written by your mother, and decide which one is which?
Fuzzy matching based on hard coded input is not what we mean when we say "AI" - I'm looking for a program that can pass the Turing Test, not one that can brute force through a million hard coded filters.
I'd argue at this point, the biggest problem with spam filters are probably false positives - those mistakes generally aren't noticed unless you're regularly checking your Spam box. There usually isn't a lot of opportunity for false positives (new contact with odd email address and certain keywords but important to you), so you might not have it occur often, but it represents a very big weakness in current filter logic.
To be clear, I'm talking about categories being created through intelligent deduction, not through hard coding. Deciding that a specific type of email is spam for Gina, but not spam for Fred, based on an understanding of who Gina and Fred are and how they relate to the particular email is one thing - deciding that a specific type of email is spam for Gina but not spam for Fred because Gina or Fred specifically told the SpamFilter so is something much less impressive.
The problem is that besides being told directly, even Google's spam filters cannot determine by deduction what other rules it should know based on that. For example, if email from ebay is spam to you in one account, but vital and important in another, does that also mean craigslist email should be treated the same way? Could it tell the difference between a falsified spam ebay email, and a legitimate one, based on content (rather than headers or other technical traces)? Can it discern that because *you* have two separate buckets that *I* should also have two separate buckets?
Hardcoding rules is easy. Jumping to an interesting and novel conclusion based on some existing rules is not:)
Categorizing spam has no analog component to it at all. No matter how many categories you decide to define, you'll never have an analog continuum - you'll have a discrete set of numbers.
In any case, the very definition of "spam" is a subjective one (dependent on the reader and the content), and currently our spam filters can only do the most basic pass (even if they do it incredibly fast). When you can create categories like "certainly spam for Gina, but not spam for Fred", and "a joke spam that Bob would enjoy, but one that would offend Lisa", then you're talking intelligence.
You cannot blithely assert that the brain works the way you posit. While the brain may very well be simply a collection of electro-stimulated biochemicals, that gives us no insight as to how you could possibly organize those biochemicals or simulate their action or function into discrete computational work. What we discern as randomness may actually have a pattern we are simply too dull to appreciate quite yet.
We can't even come close to simulating the 300,000 - 400,000 neurons in an ant's brain, much less the exact behavior of even a single-celled organism (which can have over 200 trillion molecules you'd have to simulate). I've got very little hope that we'll ever find an algorithmic path to AI - most likely, we'll simply hop on with bio-engineering, and simply create new life forms and just consider that whole brain thing a black box we can tweak here and there without fully understanding it.
Well, just because one God comes down and says "you all evolved", doesn't really mean anything, now does it? It could have created been a *different* God, and the God informing you of evolution is simply misinformed. Or maybe, the God that comes down and says "you all evolved" is actual Satan, and mere mortals simply cannot tell the difference.
The problem with your prediction that no God will come down and state "I didn't create you, you all evolved", is that you could use that for *anything*. If you wanted to assert that the hypothesis that aliens from Alpha Centauri controlled the countries of Singapore, Thailand and Greece, you could just state "I predict that no God will come down and state that Singapore, Thailand and Greece are not controlled by aliens from Alpha Centauri".
Slashdot Mirror
Fair enough. Just to be clear, though, what kind of timeframe do you put around that? 10 years of no warming trend with rising CO2? 15 years of no warming trend with rising CO2? http://blogs.forbes.com/patrickmichaels/2011/07/15/why-hasnt-the-earth-warmed-in-nearly-15-years/
My observation has been that typically, believers in CAGW will take any lack of warming trend as simply a minor setback to be ignored - they come up with some ad hoc explanation that makes the cooling somehow anomalous. With this kind of approach, CAGW can be infinitely justified (30 years of cooling? must've been a volcano. 45 years of cooling? some orbital cycle. 75 years of cooling? A spurious trend caused by solar activity.).
As for myself, it's very likely that the climate is fairly insensitive to most factors, and that natural cycles dominate on nearly every timescale we can imagine. It will be interesting, though, if the solar minimum predictions of cooling till 2050 hold up.
1) You can't assert that simply because CO2 has certain spectral properties, that therefore, as it acts in the atmosphere, it must be an overwhelming positive driver. As for an experiment that shows that some unknown effect balances out any spectral absorption contribution it could have, isn't that *exactly* what this data currently shows? That is, the theoretical model asserted that CO2 would cause one type of behavior, and for some reason, we observed something different.
2) Define "insignificant" and the "normal fluctuation of CO2". We've got a historical record that shows wild swings in CO2 levels, and on top of that, we've got huge uncertainties about CO2 sinks: http://www.science20.com/news_releases/where_does_co2_go_mystery_missing_sinks
3) Why would it have to be human activity?
You still haven't made any statement about what *observation* of temperatures, CO2, weather events, or anything else, that would falsify your hypothesis. The "heads I win, tails you lose" argument is the hallmark of religion, not science.
And *that's* exactly the point. How do you falsify a specific miracle? How do you prove that the reason why I found my keys wasn't because of Divine Intervention? What could you possibly observe that could prove that the chain of events that led up to me finding my keys *wasn't* due to a deity?
I'm asking you for any observation that will falsify the miracle of Human emissions of CO2 causing Catastrophic Global Warming. Instead of offering up any type of observation that would prove you wrong, you've instead asked me to either believe that the spectral properties of a single molecule can lead us to the conclusion that it *must* be that molecule, or that I should be able to intuit the vague definition of "insignificant" and "normal fluctuations of CO2", or that if human CO2 *doesn't* have an effect, I need to show that it is in fact countered by some other novel human effect.
Look, for evolution, falsification is simple - show me a more complex life form that precedes simpler ones. A rabbit fossil in the pre-cambrian for example. This is science.
Now try to do the same exercise for CAGW.
So if I understand you correctly, you're making two assertions. One is that intelligence is defined by learning from mistakes. The second is that the human brain can simply be defined as a Bayesian filter.
Understanding that you see human intelligence as merely an implementation of a complex Bayesian filter gives me an insight as to why you would see a spam filter as intelligent.
I think your criteria probably needs some further filtering, though. For example, my spell checker starts off with a dictionary of words. When it makes a mistake, I tell it that it made a mistake, and it learns a new allowed word or spelling. I'll assert that this fits your definition of "learns on its own", even though you have to tell it that it made a mistake (just like you tell a spam filter to add something to a white list or black list). I'm not sure if you'd count a spell checker as "intelligent".
You might refine it to the following criteria - "if it learns on its own, by recognizing its own mistakes without being *told* it made a mistake, it is intelligent." This would exclude the spell checker, and probably to a great degree the spam filter.
Now, if the spam filter *wasn't told* something was "spam" or "not spam", but deduced from say, the fact that you deleted certain things, or spent more time reading certain things than others, or some other pattern recognition in your behavior (besides explicit marking of certain emails as spam), I'd agree that it was intelligent. In this case, it wouldn't be *told* it made a mistake, it would figure it out by non-trivial observation.
I think this revised criteria would include the human brain as "intelligent", for the most part, but I would be hard pressed to say that the ability to understand that you made a mistake without outside prompting is something that can be modeled by a Bayesian filter.
I'm sorry, perhaps I've misunderstood your position. Correct me if I'm wrong, but it seems as if you're arguing that your typical spam filter is intelligent because it appears intelligent, and that simply appearing intelligent is the whole point of the Turing Test, so your typical spam filter therefore has passed the Turing Test.
From my point of view, you're making a leap in rationale that cannot be justified, so I'm trying to better understand your position by asking a different question, that is, what would you consider as *not* intelligent? If you can sufficiently expound upon what you think would *not* fit the definition of artificial intelligence, then maybe you can explain your point better. Perhaps "intelligence" is simply a matter of speed (if the spam filter was slow, you'd consider it unintelligent). Or maybe your criteria is the % of false positives, or the % of false negatives.
I can only speculate wildly about what your criteria would be, though - you haven't made clear at all what you think would *not* be artificial intelligence.
Now perhaps you simply haven't thought of the question in those terms before, and don't have a ready answer, and I can understand that. But I definitely think it's something you should at least try to think about someday :)
One, there's no way you can assert that either CO2 concentration changes or temperature changes over the last few decades are unprecedented - there are certainly similar rates observed, in the historical record, just within the past 150 years, well before the industrial revolution.
Second, simply because you haven't been able to observe with sufficient accuracy, the past several hundred thousand years, doesn't mean you can take your detailed measurements of the past few decades and assert something extraordinary. Now, maybe if we had several hundred thousand years of weather station data, you could plausibly make that argument, but any comparison of modern instrumental temperatures against historical proxies is weak at best.
That all being said, what observations of CO2 and temperature over the next few decades would you consider a falsification of your hypothesis?
Let's be specific - what observations are you talking about?
1) CO2 does not contribute [positively] to global temperature - what would you *observe* to show that? Increasing CO2, all other factors held constant and no increase of temperature? (Have fun finding all other factors held constant :) )
2) Humans do not contribute enough CO2 to cause global warming - again, what would you *observe* to show that? First you have to come up with what the minimum amount necessary to cause global warming - do you have that number? What would falsify *that* number?
3) Human contributed CO2 is negated by other human factors - what *observation* would show that?
You're offering up "prove this, and then I'm wrong" - that's *not* how a falsifiable hypothesis works. I could say "prove that God doesn't exist by showing that all miracles are frauds", but that's still *not* a falsifiable hypothesis.
Put more bluntly, is there any observation of say, CO2 levels, extreme weather events, and global average temperature, that would falsify your hypothesis of Catastrophic Anthropogenic Global Warming? If we saw more CO2, less extreme weather, and lower global average temperature, over any period of time, would that be sufficient to falsify?
Give me the concise list of observations that in combination would falsify your hypothesis. Until then, you're talking religion.
Your argument reminds me of conversations I used to have about robots. You can, through the same sort of reasoning you're using, declare that essentially, anything is a robot. Not just your typical "build a car" robots, but even your microwave, or your wristwatch, or even just a single magnet.
How about this - what is the difference between a non-intelligent spam filter, and an intelligent one? If I create a spam filter that simply hardcodes a table, and it *appears* intelligent to you, is it therefore intelligent? Where would *you* draw the line?
SELECT * FROM TABLE WHERE ISSPAM="false";
I call BS.
1) plants certainly increase in yield if the previous temperature was too cold for them to grow at all;
2) there is no evidence that warmer means "less regular climate" - in fact, most of the warming has been documented to be a *narrowing* of the temperature range, rising the cools, and doing very little to the highs;
3) http://wattsupwiththat.com/2011/06/26/global-hurricane-activity-at-historical-record-lows-new-paper/
Good. One anecdote. One that isn't even worth considering, because it's false:
http://wattsupwiththat.com/2009/03/19/despite-popular-opinion-and-calls-to-action-the-maldives-is-not-being-overrun-by-sea-level-rise/
Now, a citizen of Greenland might associate higher temperatures that allowed agriculture with good, and cooler temperatures and no agriculture as bad.
Want to trade another set of anecdotes? Or shall we play the science game and come up with a falsifiable hypothesis?
1) climate has always changed - you've got to show that this *particular* change is going to cause your *particular* harms
2) you've got to deal with the particular benefits that come with this *particular* change in order to do a proper analysis (more flooding might be bad here and there, but increased food production and crop yields could easily provide more benefit than the cost of occasional floods)
3) you've got to deal with the evidence that shows that while global temperature has increased, natural disasters have either had no change in incidence, or a positive change in incidence.
So what do you do when the risk assessment that is being made is in fact 100% backwards?
Tangential question - if you became thoroughly convinced that the data showed a safer, more benign planet when things were warmer, would you then begin encouraging CO2 emissions as a mandatory responsibility of world governments?
The alternative hypothesis is the null hypothesis - global warming and global cooling and even global temperature stability are natural phenomena. More specifically, global climate changes occur for the same reasons they did for the millions of years before humanity even existed - we might not be able to deconstruct every last variable, of course, but we don't need to have any specific cause in mind just because we've refuted the "CO2 drives temperature" hypothesis.
Or do you deny that climate ever changed before humans came along? :)
Why do you believe the risk is sufficiently high?
Or more importantly, what observations would convince you that the risk *wasn't* high? Historical incidences of higher temperatures and a more vibrant biosphere hospitable to mankind? Contemporary incidences of more harm from cold weather than harm from hot weather? Some future decrease in harm per capita during a period of increasing warming?
The problem we have is that the precautionary principle can fuck things up royally if you get it wrong. Take DDT - safe, non-toxic to humans and animals, and responsible for eliminating the scourge of malaria from countries world wide. Someone posited that there *might* be some risk associated with it, and *just to be safe*, it's been banned. As a consequence, malaria has raged throughout Africa for decades, killing tens of millions of babies.
Based on what I've read on the subject, I am unconvinced that there is any appreciable risk, and the relatively high costs and limited benefits of moving to alternative fuels and capping emissions is going to have tremendously negative unintended consequences.
Maybe he'd have been a Mormon or Scientologist :)
I'll add another corollary - in order to be qualified to disregard the idea of a falsifiable hypothesis, and to be allowed to redefine the null hypothesis, one simply must be: A Global Warming Activist; A Member of some Environmentalist Organization; Democrat, or "politically correct".
I find it *hilarious* that the same people who are getting down on someone for believing in a non-falsifiable hypothesis (Intelligent Design), cannot see that their own religious belief of Catastrophic Anthropogenic Global Warming suffers from *the exact same flaw*. It's like watching a monotheist talk about how irrational it is to believe in all the ancient Greek Gods...hypocrisy, it seems, is a trait which knows no bounds.
So isn't it true also that for some "scientist" to claim that AGW is a science (hint: it cannot be falsified so it is not), calls into question all of their other "scientific" claims?
Please, concisely state your falsifiable hypothesis of AGW, and we'll start from there.
Any claim that CO2 *is* causing global temperature increase is an unsubstantiated claim. It may be a favored hypothesis by certain scientists who have a specific political agenda, but let's not pretend for a moment that this claim has been substantiated in any way shape or form.
Before you can substantiate such a claim, you need a falsifiable hypothesis, and thus far, I've heard absolutely zero instances of falsifiable hypothesis from the alarmist camp.
Warmer probably has more positive effects than negative effects. I can certainly imagine how someone who truly believed warmer was going to be on the whole worse, this kind of thing might be particularly scary, but what if the opposite was true?
To play the science game for a moment, how would you falsify the hypothesis "warmer is better" (or even "colder is better")?
I suppose it could also be true that overall, warmer or colder just doesn't matter, and everything sort of evens out, but we might as well start with some sort of procedure for determining what is better or worse for humanity (or even all life on the planet) as a whole.
Computerized spam filters work better than a smart human being because of pure volume, not because of any sort of intelligence.
A spam filter will never realize that you really are interested in a story from Huff Post because it happens to be about one of your relatives, or heck *about you*. It has no inherent understanding about what makes something desirable, or undesirable - it's simply a collection of hard coded filters with some fuzzy matching. It may *appear* intelligent to you, but I'm afraid that's a false positive :)
Do you think your spam filter could look at three paragraphs written by you, and three paragraphs written by your mother, and decide which one is which?
Fuzzy matching based on hard coded input is not what we mean when we say "AI" - I'm looking for a program that can pass the Turing Test, not one that can brute force through a million hard coded filters.
I'd argue at this point, the biggest problem with spam filters are probably false positives - those mistakes generally aren't noticed unless you're regularly checking your Spam box. There usually isn't a lot of opportunity for false positives (new contact with odd email address and certain keywords but important to you), so you might not have it occur often, but it represents a very big weakness in current filter logic.
Spam filters don't even *attempt* the Turing Test.
http://en.wikipedia.org/wiki/Turing_test
I've never had a Spam filter engage in a text-based conversation with me to convince me it was human.
To be clear, I'm talking about categories being created through intelligent deduction, not through hard coding. Deciding that a specific type of email is spam for Gina, but not spam for Fred, based on an understanding of who Gina and Fred are and how they relate to the particular email is one thing - deciding that a specific type of email is spam for Gina but not spam for Fred because Gina or Fred specifically told the SpamFilter so is something much less impressive.
The problem is that besides being told directly, even Google's spam filters cannot determine by deduction what other rules it should know based on that. For example, if email from ebay is spam to you in one account, but vital and important in another, does that also mean craigslist email should be treated the same way? Could it tell the difference between a falsified spam ebay email, and a legitimate one, based on content (rather than headers or other technical traces)? Can it discern that because *you* have two separate buckets that *I* should also have two separate buckets?
Hardcoding rules is easy. Jumping to an interesting and novel conclusion based on some existing rules is not :)
Categorizing spam has no analog component to it at all. No matter how many categories you decide to define, you'll never have an analog continuum - you'll have a discrete set of numbers.
In any case, the very definition of "spam" is a subjective one (dependent on the reader and the content), and currently our spam filters can only do the most basic pass (even if they do it incredibly fast). When you can create categories like "certainly spam for Gina, but not spam for Fred", and "a joke spam that Bob would enjoy, but one that would offend Lisa", then you're talking intelligence.
You cannot blithely assert that the brain works the way you posit. While the brain may very well be simply a collection of electro-stimulated biochemicals, that gives us no insight as to how you could possibly organize those biochemicals or simulate their action or function into discrete computational work. What we discern as randomness may actually have a pattern we are simply too dull to appreciate quite yet.
We can't even come close to simulating the 300,000 - 400,000 neurons in an ant's brain, much less the exact behavior of even a single-celled organism (which can have over 200 trillion molecules you'd have to simulate). I've got very little hope that we'll ever find an algorithmic path to AI - most likely, we'll simply hop on with bio-engineering, and simply create new life forms and just consider that whole brain thing a black box we can tweak here and there without fully understanding it.
Well, just because one God comes down and says "you all evolved", doesn't really mean anything, now does it? It could have created been a *different* God, and the God informing you of evolution is simply misinformed. Or maybe, the God that comes down and says "you all evolved" is actual Satan, and mere mortals simply cannot tell the difference.
The problem with your prediction that no God will come down and state "I didn't create you, you all evolved", is that you could use that for *anything*. If you wanted to assert that the hypothesis that aliens from Alpha Centauri controlled the countries of Singapore, Thailand and Greece, you could just state "I predict that no God will come down and state that Singapore, Thailand and Greece are not controlled by aliens from Alpha Centauri".