And the variation in 1 million lines of valid computer code has significantly less variation than the potential variation of text files containing 1 million lines of random ascii text. So we'd still be better off with the million-monkeys-million-typewriters scenario for doing this.
Except, I think we'll find that none of the million monkeys on a million typewriters will produce a working brain program.
1. We're not trying to mimic the brain at the protein level, rather at the broader, inter-neuron level (and whatever complex intra neuron behaviour we discover).
TFA makes the point that in order to understand & emulate inter-neuron behavior, we must first understand the proteins that regulate & govern the growth and function of those neurons. So yeah, maybe we don't need to "mimic" the proteins, but we need to "understand" their behavior so we can encode it somehow into the system. And there's the rub - we have not even come CLOSE to understanding the behavior of the numerous chemicals that float around our brain. Being able to do *just this* in the next 10 years is a wildly optimistic prediction.
Yes, and past performance is no guarantee of future success. This is the problem with making these wild-ass predictions. There is a theoretical limit to the size we can shrink current electronics to, unless you really think we're going to be walking around with subatomic-sized transistors?
Moore's law could very well collapse.
He also assumes that we'll be able to design a computer that is more intelligent than we are. Based on the software I've seen & worked with, I'd say we may be able to design a computer that is "almost" as intelligent as a human, but I suspect getting it to be as intelligent as we are is going to turn out to be a continuously elusive goal, because of inevitable limitations of human insight and the fuckups that happen when a group of people are working on something together and don't communicate perfectly. Something that amazingly complex will be plagued with bugs.
So... if we design a computer, even if it is the self-aware equivalent of a C high school student... is that computer going to suddenly be able to take over the process of designing ever-smarter, ever-faster computers, better than the best PhDs and CS theorists can come up with?
Kurzweil assumes limitless upward growth. I think it's a lot more likely that we're going to approach an upper limit asymptotically, and end up with computers that are perpetually "almost" as smart as we are, but never quite reaching the same level.
It's certainly possible to perform the same computations as the human brain. In fact, many of us can (in pairs) create a new one (and it's usually pretty fun).
None of which means a thing to our ability to simulate it as a brain, rather than as a "big database that stores things."
The brain requires a quite reasonable amount of space and power, it's true. Because it functions on the cellular and molecular level. Given that a computer circuit is quite a bit larger than the average human cell, and the functionality is also significantly more constrained, I think your estimate of a "100x" scale up is ridiculously over-simplified, and simply doesn't take into account the vast differences in scale and architecture.
Sure, classic rock from obscure bands can qualify.
Obscure, 1-album bands are where it's at. You get extra points for having a concert t-shirt from them that you found at Goodwill. Preferably the 3/4-length sleeve variety.
The only way you could know this for certain is to check every million line program and see if it simulates a brain, which isn't ever going to happen, at least not in this universe.
Really, but we can realistically set a date on when we'll be able to simulate 100 billion individual pieces working together with their environment, to produce machines capable of higher thought?
I'd honestly say that writing a code generator to generate 1 million lines of code at random, and then analyze what it produces is probably the easier way to approach this - 1 million random lines of code have less potential variations than 100 billion neurons do, after all.
The fact that he demonstrates an appallingly oversimplified view of how the brain develops and functions, and almost no understanding of the fundamental biology that's required for genes to work?
I'm going to go with that. Yeah, that's my final answer.
There's also - technically - a finite number of 256-bit encryption keys... but I don't hear anybody claiming that we'll have a brute force solution to 256-bit encryption anytime soon.
But now, talking about 100 billion or so individual-and-very-much-non-binary neurons, and how they interact with one another, their environment, and how they change over time, and you think that there's a practical way to model *that* in a reasonable amount of time?
It's not technically "infinite", but it's a hell of a lot more complex than we're likely to be able to understand fully enough to build one ourselves in the next 10 years.
Maybe you should spend time scanning headlines, instead of attempting to make yourself look so very post-modern by posting snarky little regex comments that add nothing to the discussion other than to tell us, "Here's a guy who thinks he's so very much better than all of us, because he's never heard of the thing we're talking about."
To borrow a line from Chris Rock, you are clearly "keepin' it real - real dumb." If not knowing about something makes you feel better, great. Otherwise, you might try understanding the topic, and contributing a useful opinion.
I agree, the article doesn't do much to advance any sort of debate other than saying "OH NOES, FACEBOOK R TEH EVILZ!"
But services like FourSquare, and even Facebook, are fundamentally interesting pieces of technology with lots of interesting technological problems to solve. They have potential, and ignoring them in favor of "email - the way god intended for us to communicate" is a little short-sighted.
What people don't realize is that their life worked perfectly fine before this service, and it will go on afterwards as well. These services don't really add anything useful.
You're an idiot. You just summed up the entirety of human technological innovation, and discarded it with an offhand, "Your life was fine before you had it, you won't ever miss it."
See, when you say things like this, you sound like an idiot, and nobody takes idiots seriously when they state their firmly held opinions about things they've clearly never bothered to think about.
You do realize that 90% of something 100% of something... right?
You're free to claim that you believe the cup of coffee she's buying isn't worth the 100% price in the first place, but a 10% discount is still a savings for her, regardless of whether or not you consider it a worthwhile expense.
No, these services continue to exist because people find them useful, and so continue to use them, despite their flaws & risks.
You may have noticed that I followed that statement about concerns with a qualifier: "those are interesting technological challenges."
How do you build a system that gives people the interesting functionality they're looking for while allowing them to maintain some privacy and keeping their info secure? Why would this sort of a question not interest a geek who is interested in things like technology and security?
It's been discussed here on Slashdot previously, and many of the stories involving twitter, geotagging, facebook, and location-aware applications have had discussions mentioning it (and similar services, like Gowalla).
You post here on Slashdot, and yet you don't even read the headlines?
That's so 2001. The new hotness is to listen to bands that only released one album, and then broke up after the lead singer's death by heroin overdose.
It's pretty specific criteria, that's how you know the music must be good. And you can spend hours speculating on your blog about how amazing the music would have been, if the guy hadn't died.
Shit, I just spilled vanilla soy chai on my hoodie.
With all the soldiers, sailors, and airmen who don't particularly think that geeks making cutesy geek words out of their acronyms are amusing? Great plan!
I'll take the Seal team on the left, you take the Ranger battalion on the right.
As someone who ostensibly cares about technology, being here on Slashdot, why would you not be interested in hearing about geolocation applications, what they can do, and why people are using them?
Is it that you're "too cool" for anything that might involve (or even *EASE*, for those of us who are a bit awkward) social interaction, even if it is one of the more interesting recent developments in consumer tech? Take a look at all of the "Augmented Reality" type apps out there, and tell me there's not some interesting technological potential in them. The idea that you can have a device in your pocket, pull it out, and within 30 seconds be looking for "cool shit to do near where I'm standing," is amazing, because if it's built up enough to have data, you're going to start seeing more and more of the cool local shit that never gets much advertising, but is still really cool to experience - think little local restaurants that don't advertise, but have a rabid local following - wouldn't it be neat to be able to find those places easily, no matter where you are, instead of another bland steak at Friday's, because "Well, I recognize the sign, and I don't know this town."
There are obvious privacy and security concerns relevant to these kinds of apps - those are interesting technological challenges. The apps themselves are a really fascinating application of multiple technologies in a novel way. So really, the question is: why would you NOT be interested to hear a bit about the apps, and how they're being used, if you're interested enough in technology to be here reading this stuff?
Because he's arguing he's always operating the vehicle at 70%, and then claiming he'd be "just as safe" at 50%.
The arguments being used don't assume that everybody goes home and gets a good nights rest before going out, they assume that "I'm already impaired most of the time, so why is it a big deal if I'm impaired like normal, and then I impair myself a bit more?"
I'm hoping to move back closer to the city soon, where there is better public transit. Moved out this way for work, and the short commute is nice, but the inconvenience on the weekends does get old. But it's only that - a minor inconvenience.
Better public transit would be nice, but the distances that would need to be covered would make it pretty expensive, prohibitively so, probably, outside the cities.
Yes, above.05 is when you begin to be affected and your depth perception, reasoning, and peripheral vision are reduced.
I don't drink and drive, and I live a suburban life - my buddies and I pick a designated driver when we're heading out, and that person "travels free" for the night (only restriction is drinks must be non-alcoholic - everything else, dinner, any activities, gas, etc. are covered by the rest of the guys), while everybody else whoops it up. Having been the DD multiple times, I can attest that it's still a hell of a lot of fun, sometimes messing with your drunk friends is more fun than drinking itself.
I've even been pulled over on one of these nights, and asked to take a breathalyzer test - in fairness, the car did smell like it was full of drunks. I passed with a 0.0, because I'd had nothing to drink - the officer looked genuinely surprised. I told him "I'm the designated driver, no drinks for me tonight," he laughed and said "good plan, keep out of trouble then, and keep your speed down, son." And that was it. We had a fun night which was also safe, and didn't involve anybody getting thrown in jail.
It's just not that hard to do, this is what I don't understand about all the people moaning about how unreasonable this is. It's quite possible to have a (safe) fun night out that involves no alcohol for the person driving the vehicle. We rotate the duty, unless somebody volunteers to be DD on any given night.
The rest of your argument does not follow logically from those two statements at all. Your argument is that the only reason this could have happened is because of lax security or a poorly designed system, and therefore: 1) PFC Manning had access to data far beyond and above his role as an intel analyst; 2) It is the fault of IT Security & the people who built the system that this leak happened;
Your one example of a cable from Iceland that he *might* have leaked (I have not seen any confirmation of this, and in fact have seen wikileaks deny that they received the "260k" diplomatic documents that PFC Manning claimed he had leaked to Mr. Lamo, which would presumably have included this memo) is not enough evidence to declare that he had access - IN THE INTEL SYSTEM - to see, and leak, that memo. It's possible, and it bears investigation - and I'm certain that the military is investigating this leak fully to find and plug holes like this, should they exist.
Which brings us to the Afghan War Diaries. The leaks of the data pertaining to Afghanistan, you have failed to demonstrate in the least that he "should not" have had access to, and I notice that you have glossed over my response to that, wherein I explained quite clearly and coherently why and how he would have had access to that data.
So at this point, what we know is that Manning leaked 90k documents *about Afghan operations.* You have failed to show that he had access to data far above and beyond what he would have needed for his job, and you have failed to show that his job would have required access to the 90k documents he's alleged to have leaked.
Thus, you have failed to advance your point that the blame for the leak lies with the IT security people, rather than with PFC Manning's social engineering attack.
In your rush to shift the focus away from Mr. Assange and his role in publicizing the names of Afghan civilians who have cooperated with the NATO military forces there, you have ignored the arguments which effectively undermine your points. Not surprising, this is Slashdot, after all, but still, maybe you should try addressing an argument before you declare it irrelevant - it helps with the whole "chain of logic" thing required to prove a point.
Slashdot Mirror
And the variation in 1 million lines of valid computer code has significantly less variation than the potential variation of text files containing 1 million lines of random ascii text. So we'd still be better off with the million-monkeys-million-typewriters scenario for doing this.
Except, I think we'll find that none of the million monkeys on a million typewriters will produce a working brain program.
TFA makes the point that in order to understand & emulate inter-neuron behavior, we must first understand the proteins that regulate & govern the growth and function of those neurons. So yeah, maybe we don't need to "mimic" the proteins, but we need to "understand" their behavior so we can encode it somehow into the system. And there's the rub - we have not even come CLOSE to understanding the behavior of the numerous chemicals that float around our brain. Being able to do *just this* in the next 10 years is a wildly optimistic prediction.
Yes, and past performance is no guarantee of future success. This is the problem with making these wild-ass predictions. There is a theoretical limit to the size we can shrink current electronics to, unless you really think we're going to be walking around with subatomic-sized transistors?
Moore's law could very well collapse.
He also assumes that we'll be able to design a computer that is more intelligent than we are. Based on the software I've seen & worked with, I'd say we may be able to design a computer that is "almost" as intelligent as a human, but I suspect getting it to be as intelligent as we are is going to turn out to be a continuously elusive goal, because of inevitable limitations of human insight and the fuckups that happen when a group of people are working on something together and don't communicate perfectly. Something that amazingly complex will be plagued with bugs.
So... if we design a computer, even if it is the self-aware equivalent of a C high school student... is that computer going to suddenly be able to take over the process of designing ever-smarter, ever-faster computers, better than the best PhDs and CS theorists can come up with?
Kurzweil assumes limitless upward growth. I think it's a lot more likely that we're going to approach an upper limit asymptotically, and end up with computers that are perpetually "almost" as smart as we are, but never quite reaching the same level.
None of which means a thing to our ability to simulate it as a brain, rather than as a "big database that stores things."
The brain requires a quite reasonable amount of space and power, it's true. Because it functions on the cellular and molecular level. Given that a computer circuit is quite a bit larger than the average human cell, and the functionality is also significantly more constrained, I think your estimate of a "100x" scale up is ridiculously over-simplified, and simply doesn't take into account the vast differences in scale and architecture.
Been there, already have the ironic trucker hat.
Sure, classic rock from obscure bands can qualify.
Obscure, 1-album bands are where it's at. You get extra points for having a concert t-shirt from them that you found at Goodwill. Preferably the 3/4-length sleeve variety.
Really, but we can realistically set a date on when we'll be able to simulate 100 billion individual pieces working together with their environment, to produce machines capable of higher thought?
I'd honestly say that writing a code generator to generate 1 million lines of code at random, and then analyze what it produces is probably the easier way to approach this - 1 million random lines of code have less potential variations than 100 billion neurons do, after all.
The fact that he demonstrates an appallingly oversimplified view of how the brain develops and functions, and almost no understanding of the fundamental biology that's required for genes to work?
I'm going to go with that. Yeah, that's my final answer.
There's also - technically - a finite number of 256-bit encryption keys... but I don't hear anybody claiming that we'll have a brute force solution to 256-bit encryption anytime soon.
But now, talking about 100 billion or so individual-and-very-much-non-binary neurons, and how they interact with one another, their environment, and how they change over time, and you think that there's a practical way to model *that* in a reasonable amount of time?
It's not technically "infinite", but it's a hell of a lot more complex than we're likely to be able to understand fully enough to build one ourselves in the next 10 years.
Maybe you should spend time scanning headlines, instead of attempting to make yourself look so very post-modern by posting snarky little regex comments that add nothing to the discussion other than to tell us, "Here's a guy who thinks he's so very much better than all of us, because he's never heard of the thing we're talking about."
To borrow a line from Chris Rock, you are clearly "keepin' it real - real dumb." If not knowing about something makes you feel better, great. Otherwise, you might try understanding the topic, and contributing a useful opinion.
I agree, the article doesn't do much to advance any sort of debate other than saying "OH NOES, FACEBOOK R TEH EVILZ!"
But services like FourSquare, and even Facebook, are fundamentally interesting pieces of technology with lots of interesting technological problems to solve. They have potential, and ignoring them in favor of "email - the way god intended for us to communicate" is a little short-sighted.
You're an idiot. You just summed up the entirety of human technological innovation, and discarded it with an offhand, "Your life was fine before you had it, you won't ever miss it."
See, when you say things like this, you sound like an idiot, and nobody takes idiots seriously when they state their firmly held opinions about things they've clearly never bothered to think about.
You do realize that 90% of something 100% of something... right?
You're free to claim that you believe the cup of coffee she's buying isn't worth the 100% price in the first place, but a 10% discount is still a savings for her, regardless of whether or not you consider it a worthwhile expense.
No, these services continue to exist because people find them useful, and so continue to use them, despite their flaws & risks.
You may have noticed that I followed that statement about concerns with a qualifier: "those are interesting technological challenges."
How do you build a system that gives people the interesting functionality they're looking for while allowing them to maintain some privacy and keeping their info secure? Why would this sort of a question not interest a geek who is interested in things like technology and security?
It's been discussed here on Slashdot previously, and many of the stories involving twitter, geotagging, facebook, and location-aware applications have had discussions mentioning it (and similar services, like Gowalla).
You post here on Slashdot, and yet you don't even read the headlines?
That's so 2001. The new hotness is to listen to bands that only released one album, and then broke up after the lead singer's death by heroin overdose.
It's pretty specific criteria, that's how you know the music must be good. And you can spend hours speculating on your blog about how amazing the music would have been, if the guy hadn't died.
Shit, I just spilled vanilla soy chai on my hoodie.
With all the soldiers, sailors, and airmen who don't particularly think that geeks making cutesy geek words out of their acronyms are amusing? Great plan!
I'll take the Seal team on the left, you take the Ranger battalion on the right.
As someone who ostensibly cares about technology, being here on Slashdot, why would you not be interested in hearing about geolocation applications, what they can do, and why people are using them?
Is it that you're "too cool" for anything that might involve (or even *EASE*, for those of us who are a bit awkward) social interaction, even if it is one of the more interesting recent developments in consumer tech? Take a look at all of the "Augmented Reality" type apps out there, and tell me there's not some interesting technological potential in them. The idea that you can have a device in your pocket, pull it out, and within 30 seconds be looking for "cool shit to do near where I'm standing," is amazing, because if it's built up enough to have data, you're going to start seeing more and more of the cool local shit that never gets much advertising, but is still really cool to experience - think little local restaurants that don't advertise, but have a rabid local following - wouldn't it be neat to be able to find those places easily, no matter where you are, instead of another bland steak at Friday's, because "Well, I recognize the sign, and I don't know this town."
There are obvious privacy and security concerns relevant to these kinds of apps - those are interesting technological challenges. The apps themselves are a really fascinating application of multiple technologies in a novel way. So really, the question is: why would you NOT be interested to hear a bit about the apps, and how they're being used, if you're interested enough in technology to be here reading this stuff?
That's true. You should totally try my favorite bar, but it's pretty underground, so trust me, you've never heard of it.
Uh. I didn't say anything about the sex offender registry... did you mean to post this here?
Because he's arguing he's always operating the vehicle at 70%, and then claiming he'd be "just as safe" at 50%.
The arguments being used don't assume that everybody goes home and gets a good nights rest before going out, they assume that "I'm already impaired most of the time, so why is it a big deal if I'm impaired like normal, and then I impair myself a bit more?"
The logic fails.
I'm hoping to move back closer to the city soon, where there is better public transit. Moved out this way for work, and the short commute is nice, but the inconvenience on the weekends does get old. But it's only that - a minor inconvenience.
Better public transit would be nice, but the distances that would need to be covered would make it pretty expensive, prohibitively so, probably, outside the cities.
Yes, above .05 is when you begin to be affected and your depth perception, reasoning, and peripheral vision are reduced.
I don't drink and drive, and I live a suburban life - my buddies and I pick a designated driver when we're heading out, and that person "travels free" for the night (only restriction is drinks must be non-alcoholic - everything else, dinner, any activities, gas, etc. are covered by the rest of the guys), while everybody else whoops it up. Having been the DD multiple times, I can attest that it's still a hell of a lot of fun, sometimes messing with your drunk friends is more fun than drinking itself.
I've even been pulled over on one of these nights, and asked to take a breathalyzer test - in fairness, the car did smell like it was full of drunks. I passed with a 0.0, because I'd had nothing to drink - the officer looked genuinely surprised. I told him "I'm the designated driver, no drinks for me tonight," he laughed and said "good plan, keep out of trouble then, and keep your speed down, son." And that was it. We had a fun night which was also safe, and didn't involve anybody getting thrown in jail.
It's just not that hard to do, this is what I don't understand about all the people moaning about how unreasonable this is. It's quite possible to have a (safe) fun night out that involves no alcohol for the person driving the vehicle. We rotate the duty, unless somebody volunteers to be DD on any given night.
The rest of your argument does not follow logically from those two statements at all. Your argument is that the only reason this could have happened is because of lax security or a poorly designed system, and therefore:
1) PFC Manning had access to data far beyond and above his role as an intel analyst;
2) It is the fault of IT Security & the people who built the system that this leak happened;
Your one example of a cable from Iceland that he *might* have leaked (I have not seen any confirmation of this, and in fact have seen wikileaks deny that they received the "260k" diplomatic documents that PFC Manning claimed he had leaked to Mr. Lamo, which would presumably have included this memo) is not enough evidence to declare that he had access - IN THE INTEL SYSTEM - to see, and leak, that memo. It's possible, and it bears investigation - and I'm certain that the military is investigating this leak fully to find and plug holes like this, should they exist.
Which brings us to the Afghan War Diaries. The leaks of the data pertaining to Afghanistan, you have failed to demonstrate in the least that he "should not" have had access to, and I notice that you have glossed over my response to that, wherein I explained quite clearly and coherently why and how he would have had access to that data.
So at this point, what we know is that Manning leaked 90k documents *about Afghan operations.* You have failed to show that he had access to data far above and beyond what he would have needed for his job, and you have failed to show that his job would have required access to the 90k documents he's alleged to have leaked.
Thus, you have failed to advance your point that the blame for the leak lies with the IT security people, rather than with PFC Manning's social engineering attack.
In your rush to shift the focus away from Mr. Assange and his role in publicizing the names of Afghan civilians who have cooperated with the NATO military forces there, you have ignored the arguments which effectively undermine your points. Not surprising, this is Slashdot, after all, but still, maybe you should try addressing an argument before you declare it irrelevant - it helps with the whole "chain of logic" thing required to prove a point.
Exactly - find some people who have good driving records, and get a few serviceable minivans. Voila, suburban drunk bus is born.