0955: "Neutrinos"

[LLK]

This right here is why I love XKCD. One day I'm laughing about a stupid joke and the next I'm learning about neutrinos well before my morning coffee.

But you'll only get paid if you manage to have these "excited believers" (read: probably fanatics) admit that they were wrong.

Yeah, if you bet someone that it wasn't true, you'll never see your money. (I once bet someone $50 that Dumbledore actually died in Book Six... I still haven't seen that money)

I think the best thing that's going to come out of this is the newspaper headlines. I seriously don't think they've had the opportunity to write headlines like these in a while: "Tiny Neutrinos May Have Broken Cosmic Speed Limit" (NY Times). The second best thing is they might give Fermilab more funding.

[snow5379]

I don't think this means Einstein was all that wrong.

Maybe c is the speed of a neutrino and not the speed of light... and we've only figured this out recently because our calculations have just gotten accurate enough? As far as I know the neutrino isn't very well understood and could possibly be massless. My personal belief has been that a neutrino is to a photon as a positron is to a proton for a while so this experiment hasn't at all surprised me.

However... what about supernovas and such? We get the neutrinos and photons at about the same time... if neutrinos are faster then we'd get them years before we see the explosion. But... we don't. So I have my doubts just because of that.

[endolith]

Same idea as Hawking's "insurance policy"

http://en.wikipedia.org/wiki/Thorne–Hawking–Preskill_bet

[IamaWalruswithPhD]

The webcast is live right now,
we might find some subtitles at anytime.

http://webcast.web.cern.ch/webcast/

I would like to add that there would be a lot of explanation needed when Photons don't have max speed, as there has to be a reason they are slower than c_o. The seminar is quite easy to follow till it is over one's head (solid state physician here).

[Samik]

Only really really stupid people would take a bet like that

Well now you've just gone and called both of my parents stupid.


I have an outstanding bet with my father for $1,000 that the technological singularity won't have hit by 2045, and an outstanding bet with my mother for $5,000 that the US won't announce within the year that we've been in contact with aliens for a long time.

I just commented to my father a few weeks ago that I thought I should start making a living off of making bets with conspiracy theorists, so I suppose this is my first ever 'Randall GOOMH" moment.


People "want to believe" in conspiracies, and miracles, and the supernatural, and all sorts of other things along those lines - the more so the older they get. I've never been entirely positive why - I mean, I understand why some might fall into the "there's got to be something more out there" mindset as they move closer towards having to deal with the reality of their own mortality, but, for me, there's more than enough about the physical world and the human experience as it is to keep my interest / keep me occupied for a lifetime. Not so for everyone, it seems.




Edit: Obligatory link to Conspiracy Theories. "The theories are of course occasionally true, but their truth is completely uncorrelated with the believer's certainty." <-- I have learned over time that it is completely, utterly impossible to make conspiracy theorists understand this principle. So, like Randall, I just gave up and started making bets with them.

[Eebster the Great]

Funny, I actually read these results 60 ns before they were published.

[darkwombat]

The idea in the comic wouldn't work. In order to get the $200 they owe you, they would have to admit that they lost the bet. So you'd have to argue with them that they were wrong, and then it's the same thing all over again .

I don't really like to argue with people, but I'll argue with you if you pay me $200 a throw.

But you'll only get paid if you manage to have these "excited believers" (read: probably fanatics) admit that they were wrong.

-"See? The very authors admitted that their paper was flawed!"
-"But they could be wrong! That doesn't prove that it's not true!"
-"Every single scientific observation done in the last century is evidence that you are wrong!"
-"IDONTCAREHOWCANYOUBESOSUREABOUTTHATSCIENCEISNOTPERFECTANDITCHANGESALLTHETIMEIWONTGIVEYOUANYMONEYGOAWAY"

This could be averted by mutual agreement that a third party will hold the cash and arbitrate whether the bet is won or lost. Then all you have to do is be certain that your third party is not a fanatic.

[dash]

I have an outstanding bet with my father for $1,000 that the technological singularity won't have hit by 2045, and an outstanding bet with my mother for $5,000 that the US won't announce within the year that we've been in contact with aliens for a long time.

Long before 2045 what is known as the Singularity will have occured. A better question is will it be before 2015.

I think you'll be pocketing the $5000 though. There are no aliens around earth. Just us. Maybe your mother just thinks you need the money and you're just too proud to accept it as a gift, so she deliberately makes a losing bet with you.

[Samik]

http://io9.com/5843112/faster-than-light-neutrinos-not-so-fast

That io9 article that Mercuryseven quoted does a much, much better job of explaining why I think this 'discovery' is likely to be flawed than I could have done myself. Specifically with regards to Supernova 1987A:


"
This was an explosion about 160,000 light years from earth. The thing is, the neutrinos and the photons from the explosion reached us at almost exactly the same time. In the cause of intellectual honestly, I need to point out that the neutrinos were detected first, by about 3 hours, but this is because the envelope of the explosion was optically thick and the photons had to bounce around a while, while the neutrinos just streamed right out.

But how much of a delay between neutrinos and photons would we expect if the OPERA result applied?

[some math] = 3.2 years

In other words, if the effect really were this large, we would have seen the neutrinos from SN 1987A way back in 1984. Yeah, we would have noticed that.
"


It really becomes an issue of sample size / magnitude / whatever. On the one hand, we have neutrinos and photons covering 160,000 light years, and on the other, we have neutrinos traversing merely 730 km. In which case would any discrepancy be more likely to be observable?

[Samik]

Maybe your mother just thinks you need the money and you're just too proud to accept it as a gift, so she deliberately makes a losing bet with you.

She needs the money more than I do, by a long shot. I won't be making any attempt to collect when the end of the year rolls around. (Also, if you've been party to the debates I'd had with her about government conspiracies and alien cover ups and whatnot, you wouldn't be skeptical of her conviction.)

Long before 2045 what is known as the Singularity will have occured. A better question is will it be before 2015.

....

Want to make a bet?



Ok, all kidding aside, this cuts right to the center of the issue - that people seem to have drastically different interpretations (or, in some cases, no real solid understanding at all) of what, exactly, the technological singularity is supposed to be. My understanding is that it refers to an event where an AI is constructed that has the ability to improve it's own intelligence / functionality at a faster rate than humans could improve it, leading to runaway growth in intelligence / functionality as it constantly improves itself and therefor improves its ability to improve itself.

I think something like this has to be the definition of the singularity, because most of the alternatives I've heard suggested are pretty meaningless. If your definition of "technological singularity" is exponential growth of technology, then I would point out that technology has been improving exponentially for several million years, at least - it took us almost a millions years to go from Oldowan to Acheulean tools. That's just not an interesting claim to me in any way.

[J Thomas]

Is there anybody who actually believes that GPS wouldn't work without relativity?

Is there anybody who believes that GPS couldn't be made to work if relativity was true and we didn't know it?

How about if relativity was wrong or imprecise but we thought it was right? Would that make GPS stop working?

Isn't it true in general that technology can sometimes be informed by theory, but develops in large measure by trial and error? Moore's Law works because chipmakers reduce size at a predictable rate, finding by trial and error how to get results at each new size. Wouldn't it be crazy to say that quantum electrodynamics has to be correct because computer chips work?

[SpringLoaded12]

My favorite thing about the fora is that whenever the strip suggests an interesting idea or brings up a certain controversial geeky topic, tons of people start discussing plausible setups to make the idea real, suggest ways to improve it, or say why it won't work, or in the case of the controversial topic, they go into debate about it.

My least favorite thing about the fora is actually the same thing. The thread for the strip that mentioned atheism and fundamentalism was thirty-something pages long. I sampled two pages and saw debates going on over things tangent to the atheism vs. fundamentalism debate, often going on for paragraphs over the tiniest details of others' posts. A lot of the debate was over evolutionism (which it should be noted is not synonymous with atheism, though they often overlap). Of course, the entire thread missed the point of the strip, which was that the endless debate was wasteful and futile. I can only imagine what the threads would look like if they were over SMBC strips, which often involve some mockery towards fundamentalism or towards Christianity as a whole.

My second least favorite thing about the fora is the tendency of many readers to take serious offense when the strip's joke involves harm to unnamed people, or lightly ribs a certain demographic.

[Samik]

they go into debate about it ... often going on for paragraphs over the tiniest details of others' posts.

Is... isn't that why we're here? On the XKCD forum? To debate geeky things?

[belliott4488]

The timing for this strip was perfect. Someone posted a link to an article about the neutrino observations on my Facebook profile page, so I posted a link to this strip in response.

[jcsalomon]

I like the way the paper ends:

… the potentially great impact of the result motivates the continuation of our studies in order to investigate possible still unknown systematic effects that could explain the observed anomaly. We deliberately do not attempt any theoretical or phenomenological interpretation of the results.

This is definitely a “that’s funny…” moment.

[lylebot]

Relativity is a theory (two theories actually); they're either true or not true. General relativity is probably not true, since it's not compatible with another very well-tested physical theory (quantum mechanics). Physicists have known this for a long time.

Even if it's not true, it's still a really, really good approximation. Good enough for GPS, for instance. But it really is likely that at some point someone will observe something that contradicts it. I have no idea if this neutrino thing is that observation (though I would also bet that it isn't).

[chironomidae]

The thing is, we already have proof that Einstein's theory isn't 100% complete - it can't describe the very small. For that we need quantum theory, and uniting the two has been the ongoing plight of physicists for the last several decades. I will remain skeptical of this finding until it can be corroborated, but in the back of mind I can't help but wonder if this is the missing piece of data that has heretofore kept us from discovering the "theory of everything".

The difference between this experiment and the 1987A supernova is the fact that most of the neutrinos from 1987A traveled through a relative vacuum while the neutrinos in this experiment traveled through matter. Perhaps the interactions (albeit very weak) between the neutrino and the mass of the earth somehow sped them up, like quantum tunneling or some other yet-unknown interaction. Who knows. Any way you cut it, it's hard not to be a little bit excited.

[dash]

...what, exactly, the technological singularity is supposed to be. My understanding is that it refers to an event where an AI is constructed that has the ability to improve it's own intelligence / functionality at a faster rate than humans could improve it, leading to runaway growth in intelligence / functionality as it constantly improves itself and therefor improves its ability to improve itself.

I agree with that interpretation. The details of how it would play out are subject to debate. Until we have real AI it's all just handwaving.

I object to any prediction of anything more than around 2 years in the future. It's like trying to predict the weather more than 2 weeks ahead. Utterly meaningless. Between now and then anything could happen.

Mainstream science is hopelessly corrupt. Read Halton Arp's book "Seeing Red: Redshifts, Cosmology and Academic Science". Or Eric Lerner's "The Big Bang Never Happened". I think Halton Arp's argument against quasars being far away is very persuasive. If they are closer, the Big Bang is an epic failure.

I didn't like Randall's tacit acceptance that all non-mainstream claims will turn out to be false. It's a sort of herd mentality, taking comfort in being on the side of the majority. But sometimes the herd gets driven over the cliff!

Anyway that's it for me, I don't want to get drawn into any debates.

[Samik]

I object to any prediction of anything more than around 2 years in the future. It's like trying to predict the weather more than 2 weeks ahead. Utterly meaningless. Between now and then anything could happen.

I can understand this point of view. It just didn't seem like this is what you were saying when you said:

Long before 2045 what is known as the Singularity will have occured. A better question is will it be before 2015.

That strongly implies that you personally believe it will happen sooner rather than later, which is a different thing than saying that "later" cannot be predicted with any real degree of certainty.

I didn't like Randall's tacit acceptance that all non-mainstream claims will turn out to be false. It's a sort of herd mentality, taking comfort in being on the side of the majority. But sometimes the herd gets driven over the cliff!

I don't think that's what he's doing at all. I think the principle at play here is the one I quoted in my post on the first page, from Consiracy Theories: "The theories are of course occasionally true, but their truth is completely uncorrelated with the believer's certainty."

Skepticism does not necessitate close-mindedness.




If you think about it, this is pretty much the reason the scenario in the strip involved gambling - even if the stick protagonist occasionally loses some his bets, he's likely to come out significantly on top over the long haul when betting against studies with results that purport to "[overturn] all of physics".

[Aelfyre]

I have an outstanding bet with my father for $1,000 that the technological singularity won't have hit by 2045, and an outstanding bet with my mother for $5,000 that the US won't announce within the year that we've been in contact with aliens for a long time.

Long before 2045 what is known as the Singularity will have occured. A better question is will it be before 2015.

I think you'll be pocketing the $5000 though. There are no aliens around earth. Just us. Maybe your mother just thinks you need the money and you're just too proud to accept it as a gift, so she deliberately makes a losing bet with you.

I do think 2045 is a likely timeframe +/- 5 years.. but I think 2015 is perhaps a *tad* optimistic..

To be clear, I would *love* to wrong on this one.. and see the Singularity occur while I was reasonably young but I think there are several breakthrus yet to be made that rely upon several technologies yet to be developed.

here's to hoping tho.. LOL

[gingermrkettle]

On a non-technical level, some of the coverage has noted that relativistic physics is necessary for GPS receivers to attain the accuracy they do. There seems to be a slightly odd logical sequence of events if the measurements that appear to disprove a theory are reliant on that theory being correct.

[frezik]

Yeah, I know, I'm sounding like one of those crackpots I love to hate, but still...

Yes, yes, you are.

There's basically one area of physics considered unassailable: thermodynamics. It's the closest thing physics has to a mathematical axiom.

Everything else lies on a range of probability. As Randall's strip alluded to, corrections for relativity (involving the fact that satellites move faster relative to objects on the ground) are built into every GPS system. The correction factor is tiny, but it would be enough to completely throw off the system. Each time someone uses a GPS to get somewhere, they're doing a small, informal experiment for Einstein's theories.

So which is more likely--that an extremely well tested theory is wrong, or that a hugely complicated set of instruments had a manufacturing defect in one little piece and ended up producing bad data?

It will be very interesting if this data does pan out. But there are lots of interesting things in the world, and I'll be lucky to experience even one tenth of them. I don't feel the need to chase after unlikely cases on the potential that they'll be interesting.

Mainstream science is hopelessly corrupt. Read Halton Arp's book "Seeing Red: Redshifts, Cosmology and Academic Science". Or Eric Lerner's "The Big Bang Never Happened". I think Halton Arp's argument against quasars being far away is very persuasive. If they are closer, the Big Bang is an epic failure.

Halton Arp's ideas have not held up under new data.

[Samik]

To be clear, I would *love* to wrong on this one.. and see the Singularity occur while I was reasonably young but I think there are several breakthrus yet to be made that rely upon several technologies yet to be developed.

I just personally believe that we are not even anywhere in the vicinity of creating anything that can be meaningfully called AI, even in a weak sense.

Sure, we can program computers, with application of massive amounts of processing power and storage capacity, to perform single task objectives, like chess or jeopardy, reasonably well. But that's just, in my opinion, not AI. As long as you're brute forcing the functionality, you haven't created anything that is meaningfully intelligent. Or, at the very least, you haven't laid any of the conceptual groundwork for AI that can step beyond the single objective they've been brute force designed to accomplish (which is something any "singularity" inducing AI would need to be able to do).

And this is the crux of it for me, and where I agree with you that we are missing some key bits of knowledge that we'd need to design an AI capable of brining about the singularity: Our ability to design such an AI is currently hampered less by our lack of technological proficiency than by our lack of understanding of just exactly how "intelligence" functions in the first place.

Any continuing advancements we make re: computational power/efficiency will be made at a pre-singularity, human-level (as opposed to 'super-human') rates until we know a lot more about the underlying structure of systems that generate 'intelligence' than we currently do.



EDIT: Also, consider that even the creation of strong AI isn't sufficient to bring about the technological singularity. After all, there are (although this is potentially debatable, depending on who you talk to) around 6 billions strong non-artificial intelligences walking around on the planet right now, but none of us have to ability to achieve singularity-inducing rates of self-improvement.

To be capable of inducing the singularity, it isn't sufficient for an AI to just be smarter than us - it will need to have a vastly, vastly different and improved spin on "intelligence" than anything we've created or postulated thus far.

Considering that we haven't even come close to creating any AI that can even begin to approximate human-level cognition, it's hard to see how we can be very near creating an AI that vastly outstrips such.

[Samik]

On a non-technical level, some of the coverage has noted that relativistic physics is necessary for GPS receivers to attain the accuracy they do. There seems to be a slightly odd logical sequence of events if the measurements that appear to disprove a theory are reliant on that theory being correct.

Now that is an interesting point.

[eran_rathan]

On a non-technical level, some of the coverage has noted that relativistic physics is necessary for GPS receivers to attain the accuracy they do. There seems to be a slightly odd logical sequence of events if the measurements that appear to disprove a theory are reliant on that theory being correct.

Indeed.

Each GPS satellite has an atomic clock onboard (which is part of the way that GPS works, by re-sectioning your (the reciever's) position based on the distances to each satellite, and each satellite's relative position to each other based on their orbital position at a particular time). The tiny amount of relativistic effects from moving from Terra's surface to orbit, and the simultaneous increase in acceleration, must be taken into account (along with certain atmospheric effects).

Now if only we can get those idiots at Lightsquared to lead our GPS bandwidth alone...

[dash]

I do think 2045 is a likely timeframe +/- 5 years.. but I think 2015 is perhaps a *tad* optimistic..

When I hear numbers in this realm I assume it comes back to Kurzweil's prediction of that being the point where, due to Moore's law and the steady doubling of computational power every 18 months or so that a cheap desktop computer will have enough power to model a human brain.

I counter with that whole argument being ridiculous. We have supercomputers. The problem we face is the theory, what to actually do with the computational power if it's available.

There is a TED talk or something like that where a guy talks about the implementation of algorithms from the standpoint of computational steps per watt of power. It goes like this:

If you have a general purpose computer (like all our PC's) and you give it a task, it performs X operations per watt.

If you create an FPGA (field programmable gate array) solution that does just the one critical function, you get 100 times the performance, or 100X operations per watt.

And if you go to creating an ASIC (application specific integrated circuit) you get another 100 factor improvement. So you're at the 10000X operations per watt.

History is full of examples. Cell phones are one. ASICS are on the order of 10,000 times more efficient in terms of power consumed per computation.

So in conclusion, if we knew the theory of what needs to be computed to make real AI, actually doing it is the easy part. It's easy to improve upon already working technology. That's the trivial part.

Program Blue Gene the right way and you've got real AI. A year or two later and you've got a toaster that can carry on a conversation.

[Samik]

So in conclusion, if we knew the theory of what needs to be computed to make real AI, actually doing it is the easy part. It's easy to improve upon already working technology. That's the trivial part.

Program Blue Gene the right way and you've got real AI. A year or two later and you've got a toaster that can carry on a conversation.

Exactly.

The roadblocks standing in the way of the singularity have precisely nothing to do with our ability to bring computational power/efficiency to the table.



EDIT: Wait a minute, it's you again, Dash. Oh great. Now we're using the exact same argument to argue for exactly opposite things.

[dash]

Halton Arp's ideas have not held up under new data.

Arp talks about such responses to his work. From a quick glance at that link, which I think you mean:

"... the publicly available data from the Sloan Digital Sky Survey and 2dF QSO redshift survey to test the hypothesis that QSOs are ejected from active galaxies with periodic noncosmological redshifts. For two different intrinsic redshift models, [...] and find there is no evidence for a periodicity at the predicted frequency in log(1+z), or at any other frequency."

I think the subject in question is Halton Arp's observation that redshifts are quantized (I think something like 37.5 km/second is the amount, all redshifts are multiples of this according to his assertions). That quantization is a sideline, whether it is true or not the important point is that the quasars are ejected from central bodies that have far less redshift than the quasars ejected. So we have 3 objects all in a line, again and again and again, with lots of evidence they're very close to each other, yet they have drastically different redshifts, always the same way.

So Arp argues persuasively that the quasars are ejected from the central body and their redshift is intrinsic to them, not due to the fact that they are very, very much further away and receding very quickly.

Personally I think Arp runs into trouble when he presents his own ideas of what's going on. He ought to confine himself to trying to disprove the Big Bang instead of presenting his own theory. What happens is people look at his theory and reject it, and so don't even look at his refutation of the Big Bang. It's kind of funny, actually. Tragically funny...

It's as if "I don't know" is not a viable answer in mainstream science. Mainstream science has to always believe in something.

[Ptharien's Flame]

I'd just like to bring everyone's attention to the following article, which I feel is especially relevant now that the media in general have gotton wind of "FTL neutrinos".
http://www.newyorker.com/reporting/2010/12/13/101213fa_fact_lehrer

[dash]

EDIT: Wait a minute, it's you again, Dash. Oh great. Now we're using the exact same argument to argue for exactly opposite things.

No need to capitalize my username.

I guess I don't really understand what the opposite things are?

[Samik]

I guess I don't really understand what the opposite things are?

Oh, I was just finding it funny that we both seem to agree on the point that Kurzweil-style singularity predictions focusing largely on processing power are flawed, but that you seem to move from that understanding towards a near-term singularity, whereas I use the same point to argue against one (see my earlier post, if you hadn't already).

[blowfishhootie]

Should we really view Einsteins theories as a hard, uncompromising limit on the universe?

I can't speak for the media, or the common person, (or really the scientific community either) but the scientific community doesn't view Einstein as complete truth.

Consider it this way: Either Einstein or this new study has made an error. Einstein's theories have a lot more supporting evidence, therefore it is far more likely for this study to be in error. Therefore the scientific community will now try to find that error. If the new study is in fact correct, they will not find an error, and will become strong enough that it will become more likely that Einstein made the error. Then the scientific community will devote their time to trying to find Einstein's error.

Einstein's theories aren't regarded as true, they're regarded as most likely to be true. Same as Newton before him.

I think this is a pretty great explanation, especially the last line. The person to whom you are responding seems to have a fundamental misunderstanding as to the nature and purpose of scientific theories.

[dash]

but that you seem to move from that understanding towards a near-term singularity, whereas I use the same point to argue against one

Kurzweil can't conceive of a singularity before 2045 timeframe. I refute his argument. You have made no argument as to why it should be even later than 2045. Your "argument" is just the assertion, "This is a very hard problem! It's going to take a long time!"

We're just at the level of opinion at this point. Until it happens it's all just speculation. There are no experts.

[Samik]

Kurzweil can't conceive of a singularity before 2045 timeframe. I refute his argument. You have made no argument as to why it should be even later than 2045. Your "argument" is just the assertion, "This is a very hard problem! It's going to take a long time!"

We're just at the level of opinion at this point. Until it happens it's all just speculation. There are no experts.

Well, no, of course my position is hardly rock solid. As you say, trying to lock down the "truth" of what's going to happen over the next several decades is "like trying to predict the weather more than 2 weeks ahead."


You state, matter of factly, "I refute his argument.", as if you've done some additional argumentative work for your position than I. I refute his arguement as well, for exactly the same reasons as you, which my prior two posts should make clear. If my refutation is invalid, so is yours, but I don't think this is what you meant. It's what's beyond this refutation that we disagree on:


You seem to take the point that our ability to generate singularity-inducing AI's is not being hung up by a lack of available processing power, and desire to move from that towards a near-term singularity. I take the same point and desire to move away from a near-term singularity.

If my "argument" is just a weak assertion, then yours is even less - you haven't suggested any reason whatsoever for anyone to believe that such a powerful AI can be produced by 2015, or before 2045. I at least attempt to make some justification for why I believe we're a long ways off. You haven't even done that much.


EDIT: What I mean to say is that saying "the result X is not being held up by variable A" is not the same thing as saying "the result X is not being held up". Refuting, even successfully, Kurzweil's beliefs about the singularity and computational power is not the same thing as making any effective argument that the singularity is imminent, likely, or even possible. All you've done is argued against "variable A". In that sense, you've done even less argumentative work for your position than I.


Edit2: I should point out that when I say "you haven't suggested any reason whatsoever for anyone to believe... etc", I am not implying that I or anyone else asked or expected you too. I never attacked the integrity of your position, I merely pointed it out that I found it amusing how we could both look at the same argument and derive from it opposite conclusions. That was entertaining to me. However, once you attacked the integrity of my position, I wasn't about to refrain from using the same style of attack to demonstrate that yours was weaker still.

[cellocgw]

On a non-technical level, some of the coverage has noted that relativistic physics is necessary for GPS receivers to attain the accuracy they do. There seems to be a slightly odd logical sequence of events if the measurements that appear to disprove a theory are reliant on that theory being correct.

Now that is an interesting point.

Interesting but incorrect. GPS position validity (not accuracy) is, and has been, easily tested by repeatedly recording the GPS estimate of a static location on the ground. More simply put: turn on your GPS while sitting on your couch. Do the same every hour, or day, or whatever. If you remove the relativistic corrections to the GPS's calculations, you will have data suggesting your couch moved many meters. The only question might be, as others posted, that GR is not quite correct and this nanosecond discrepancy might be related to that.

[dash]

If my "argument" is just a weak assertion, then yours is even less - you haven't suggested any reason whatsoever for anyone to believe that such a powerful AI can be produced by 2015, or before 2045. I at least attempt to make some justification for why I believe we're a long ways off. You haven't even done that much.

Kurzweil claims we won't get real AI until 2045 (or thereabouts) and the reason is that's when a desktop PC will be as powerful as a human brain. There are various assumptions he's making in that conclusion. The one I'm interested in is the concept that making real AI happen is limited by the computing power of a desktop PC.

I give an example of how technology can advance very quickly, far beyond Moore's law rates. Hard AI could happen tomorrow. It might have already happened, but it's kept secret. It could happen before 2015. That's just my informed opinion.

Kurzweil seems to advocate a policy of "Let's just sit on our hands until computing power gets to where it needs to be before we even think about tackling this problem." I have almost no respect for Kurzweil, actually. And even less for Minsky, for that matter.

I agree the singularity can't happen until after real AI occurs. I also agree with you that what passes for AI is really just computer science and the exploration of algorithms. It's not anything like how the human brain achieves intelligence. And that's part of why I dismiss Minsky.

I'm of the opinion the real AI problem (actually I call it Machine Intelligence) is much simpler than anyone thinks. But never underestimate the stupidity of the human race or its ability to bark up the wrong tree and make simple things complicated.

The nature of the singularity and details are just matters of opinion. There's no point in debating opinion. To each his own.

[ramosfer]

Chang Kee Jung, a neutrino physicist at Stony Brook University in New York, says he’d wager that the result is the product of a systematic error. “I wouldn’t bet my wife and kids because they’d get mad,” he says. “But I’d bet my house.”
- Read in Wired magazine

[Samik]

The things dash said.

Again, I have no relevant disagreement with anything in your paragraphs 1, 3 or 4. Neither of us have any respect for Kurzweil. Enough said.

Where we disagree appears to be on how relatively easy it is / will be to generate "real AI / machine intelligence". You say -

I'm of the opinion the [machine intelligence] problem ... is much simpler than anyone thinks.

- I cannot help but point out that this, currently, is the most substantive "argument" you've made so far in this thread about that.

Demonstrating scenarios where technology has advanced beyond the rate of Moore's law doesn't get you very far, considering that Moore's law refers specifically to transistor density on integrated circuits. If you're extracting the "doubles every 18 months / 2 years / whatever" component, then of course there will be examples of faster progress, because not every technological endeavor is subject to the exact same challenges and constraints as putting transistors on integrated circuits.

Developing strong AI / machine intelligence, we both seem to agree, is one of these unrelated endeavors. And just because there's no reason to believe that our progress in this regard is constrained to within a "doubles every 2 years" framework (What would that even mean in this context? That the AIs we design are twice as "realistic" every 2 years?) doesn't mean there's any reason to believe results will occur more rapidly.


Right now, I'm saying "it's hard", and you're saying "it's easy", and that's about as far as we've gotten. Yet we both seem to agree that we're not currently really on the right track -

I also agree with you that what passes for AI is really just computer science and the exploration of algorithms.

In the absence of any other evidence or valid points of argument (not saying there aren't any - just not getting ahead of myself - we certainly have yet introduced any), not being very close and not being on the right track do more to suggest distance from the goal than proximity.



EDIT:
RE:

The nature of the singularity and details are just matters of opinion. There's no point in debating opinion. To each his own.

Just because neither of us has gone very far into constructing any kind of valid argument does not mean that it is entirely impossible to approach questions about the future with any sort of valid critical process. We're certainly not likely to be able to pin down the date of the singularity's arrival, but do you really want to say that no one can make any educated guesses about what the future holds whatsoever? No estimates of the probability of certain events that are any better than random guesses?

Tell me that I don't know enough about the subject matter to argue about it intelligently, and I may not be able to disagree with you. But suggest that "there's no point" in even trying to make educated guesses, and you lose me very quickly.

[dash]

Right now, I'm saying "it's hard", and you're saying "it's easy", and that's about as far as we've gotten. Yet we both seem to agree that we're not currently really on the right track -

It's just my opinion, as I've said repeatedly. It is also my opinion that my opinion is an informed one.

You know, I certainly didn't mean to offend you, which it seems I did. All I can think of is that it bothered you that I put quotes around the word "argument". I did that to try to convey that it wasn't an argument, it was just you stating your opinion. And we don't want to debate opinion, I think.

[shadowoftruth]

Can someone help out an engineering student with a love of physics?

I read up on the results and from what I'm reading it looks like the only thing to exceed C are neutrinos -(Muon Neutrinos to be precise)? From my understanding of relativity haven't we already seen it break down on a subatomic scale, many many times? (Examples I can think of off the top of my head are, quantum entanglement, electron spin shells, and anything to do with gravity on a quantum scale). From all the research I've read it seems like we pretty well understand that relativity is a great equation for governing and predicting large scale interactions, but isn't a valid theory on a quantum scale. This was an issue even in Einstein's time and all that research spent on finding a theory of quantum gravity seems to demonstrate that it has been pretty well understood that realities doesn't work on quantum scales.

So could somebody explain to me why this study is so amazing and ground breaking? I find it interesting that we might have found particles moving faster then light but still we are talking about relativity here, not something like String theory or modern atomic theory, being broken at a quantum scale.