If artificial intelligence ever gets to the point where it is greater than humans, won't it be capable of producing even better AI
Depends on if it is created by understanding intelligence, or through an accident. We have one example of intelligence created by accident, which has as of yet not managed to create another intelligence.
Even understanding may not suffice. If humans create a human-plus AI and understand how they did it, then it is natural to assume that the human-plus AI could create a human-plus-plus AI, but just as there are many problems with creating a human-plus AI, there might be many problems in creating a human-plus-plus AI, and the improvements we make may not be improvements which lend themselves to the task of creating the next generation of AI. It may be that creating a super-human AI involves solving hundreds of isolated problems over a very long period, and each problem solved may only provide a tiny bit of gain in solving the next problem.
As a specific example, what if we created a human-level AI which had the additional ability to over the long run beat any human at poker? In that way, this AI would be super-human, but the AI may not provide any advantage in creation of the next generation. There are lots of specific human talents one could imagine perfecting in an AI without improving that AI's ability to create a new AI (in fact, many of those talents may be distractions from that job!).
2. Basestation/routers need a simple-to-configure mode where they will let others into a separate subnet that goes straight out to the Internet but does not see my home computers directly.
Easier to just have a single open-access mode, and configure your own laptops to VPN into your home network. Also better alignment of the incentives.
As someone else mentioned, it was more likely Tangents, in a short-story collection of the same name by Greg Bear. The neighbor was based on Alan Turing, he's doing research on visualizing higher dimensions with music, and the kid is able to actually see higher dimensions directly. Really similar to "Mimsy Were the Borogoves" in flavor.
By "Had to explain", I'm sure the article means "As I was running on and on and on about historical crap, someone threw out an 'insight' to distract me."
I liked SUNW. I have fond memories about learning to program on SUN and HP workstations.
What I think is really funny is that I've been developing for workstations and servers for basically 2 decades, and it was only with this thread that I realized that the "W" in "SUNW" is for "Workstation". I always just figured it was a random letter to disambiguate Sun-the-computer-company from Sun-some-other-company. You know, like how AAPL doesn't really stand for Aple? Go figure, this problem probably is the most important thing for Sun to be working on just now.
The event I compare this to is Borland -> Inprise. Basically, the company has finally recognized that it's in dire straits, so it's taking the only thing keeping it ticking in the technical world (the respect engendered by its long and colorful history), and burying it.
SCO is so dirt-cheap right now that you can basically play with it without risk.
Why was this modded insightful? The price of the stock is completely irrelevant to how much risk you're taking. If SCO did a 10000-to-1 reverse split, would they then be much more risky because their stock was trading at $10k? Contrawise, various brokers allow fractional share purchases, but just because you can buy $100 worth of any stock doesn't mean they're equally risky.
You're right that there's little evidence to believe that something will evolve to replace it. More likely, we'll go extinct and then some other intelligent species might evolve, but given the billions of years evolution that previously took place, not very likely.
It took billions years for Earth to develop life of any sort, then most of a billion more to develop beyond single cells, then hundreds of millions to develop the diversity of things we recognize as the animal kingdom. But if humans could evolve from the point of the Cretaceous-Tertiary extinction event, then there wouldn't seem to be any reason something similarily-advanced couldn't evolve in 100 million years from a similar event at this time.
Can you imagine getting on a "soft-key" elevator? I think it would be cool at first, then really annoying.
An elevator should so one thing, and only one thing. You don't watch YouTube videos on your elevator. You don't browse web pages on your elevator. You don't have to adjust the picture-in-picture setting on your elevator.
Also, an elevator is a fixed implementation. Your elevator goes to 23 floors, or it goes to 15 floors, or it goes to 3 floors. You'll never download a firmware update to enable access to three more floors.
Gone are the days when a cellphone did one thing, and you literally only needed the numeric pad to dial with. Nowadays, your cellphone performs dozens of functions, and dialing phone numbers may no longer be the primary function.
I doubt Apple's current iPhone interface will be the one we all think of when we think back on things in five years. They're going to iterate, but given what I've seen comparing the iPod to competitors, I wouldnt' bet against them. Whether a specific implementation is right is a different question than whether their overall design aesthetic is right.
Back in the 80's, as a high school student I used valuable earnings to subscribe to BYTE, Dr Dobbs, and Computer Language. These were all great at the time (BYTE faded a bit, but was great in the mid-80s). Kept them up through college, but then CMP swallowed them all up and crushed them into pulp. They screwed up subscriptions, closed things down and randomly changed publishing schedules, and got rid of anyone with half a brain. So I dropped all those subscriptions, and didn't bother to go back.
Really, I suspect the Internet killed good technical mags. Blogs are just the scapegoat, but once you could look all this stuff and discuss it online, there wasn't as much demand for in-depth technical articles. That left magazines full of vendor-written articles, which nobody was interested in, but at least they were cheap.
Does Moore have $9,000,000 lying around somewhere?
A websearch on "Michael-Moore net-worth" certainly makes it easy to believe that he's worth a couple tens of millions of dollars. Unfortunately, he's apparently secretive on this front, so there's little in the way of unbiased estimation.
Great idea! There should be an MMO where you get to spend your day as a virtual trader!
Re:Yeah, UCSD p-code Pascal! :)
on
The Apple II At 30
·
· Score: 2, Insightful
You must have been using a different USCD than I recall. My memory was of swapping disks. Edit, save, swap disks, compile, swap disks, run, ad nauseum. You could arrange certain combinations to live on the same disk, but that combination did not include "Everything you needed to do stuff".
Which meant that when we got a Z80 card in one machine and ran Turbo Pascal, which was just everything-in-one-place, it was like heaven.
I can't wait to surf Mars. With moons that close, there ought to be tidal swells that one could ride forever.
Dude, Earth's moon is millions of times heavier than the moons of Mars. They're going to have to be pretty damn close to get a tidal swell worth riding, even with the reduced gravity.
I guess that the 9/11 hijackers used library computers doesn't help, nor does the current "Library 2.0" movement to offer customized services.
This doesn't sound right, but... why _shouldn't_ the 9/11 hijackers have used library computers? I mean, it's terrible that library computers were used, but it's not like that made them complicit. The hijackers probably also travelled on public roads, and drank water from municipal water supplies, and benefitted from living in a safe neighborhood due to local law enforcement, and used dozens or hundreds of other public services. That's what public services _are_, after, all. Beyond that, they probably bought food in a grocery store, etc, etc. If we start cracking down on libraries because of this - where do we stop?
[I'm not suggesting that you agree that the above is a good reason to crack down on libraries in any way, I'm just being annoyed that people seem to think we should "crack down" in this kind of thing. I suppose most such people don't even know where their local library is located:-).]
I've encountered CS students recently who in their third year are unable to do such basic things as understand memory allocation.
Right on. But I wouldn't qualify it with "recently". When I was in school, we didn't have this new-fangled Internet stuff - when I was in high school, getting access to a computer _period_ was rough. Which is why my early CS memories are about doing things like writing a Forth kernel in 6502 assembly and simulating execution longhand (ie, on sheets of paper), and spending a long afternoon working around trig tables to implement a 3D rotation library using entirely fixed-point arithmetic.
There is really nothing more frustrating than doing an interview of someone with a masters degree or years of experience, and realize that they don't understand hashtables. I mean, _hashtables_, man! Sigh. Then again, I thought data structures and algorithms was super fun in school, there were these different ways to accomplish things, and knowing the characteristics of each made your programs smaller and faster. I'm just glad that knowing this stuff has turned out to be so rewarding in the real world.
Indian and Japanese kids kick ALL amercian kids asses hard in education. to the point that the brightest graduating today are considered mentially retarted compared to the average students overseas.
This is clearly why India and Japan so completely dominate the world markets. The only way to end the massive depression the US has been in for the past 20 years is to improve our educational system, otherwise we'll just become a subsidiary of Japan, Inc.
I'm a fan of the Abiogenic theory of oil.... I think crude oil is just a part of the earth's carbon cycle.
Does it matter? Under one set of theories, we have to bury a bunch of biomass, wait a couple hundred million years, and we have more oil. Under the other theory, the oil is a natural part of planetary development, so we have to wait... a couple billion years? It's a little like arguing about the cosmic microwave background radiation, it's an interesting argument, but not terribly practical.
*Shrug*. I like Shrek and Shrek 2 a great deal, but they _are_ just a series of set pieces strung together which only works due to the casting. For the most part, Pixar creates pretty amazing movies which feel greater than the sum of their parts in many ways. That doesn't mean that Dreamworks absolutely sucks, it's just a differe style.
It's like Disney versus Warner Brothers. Initially, Warner Brothers was cheap and wanna-be, but over time they really came into their own by developed a cutting wit which simply wasn't present in the Disney pieces. I don't think Dreamworks is there yet, but it could happen.
I mean, this is Kirk. Obviously they'll have to have some sort of deathmatch involving laser boundaries and electrified monkey bars, and perhaps an improvised trebuchet or flintlock.
MacOS X embedded? WTF? I mean, I'm sure it's a wet dream to imagine that you could run the same thing across a bunch of platforms, but... no, it's not going to be the same thing, or even a very similar thing. In fact, there's a word for it: it will be a different thing.
I mean, look at Windows CE. The main similarity it has to Windows XP is that they both have Windows in their names. Sure, there are APIs which are similar between them - that's because if you have an existing API to do a particular job and it's working fine, you'd be silly to create an entirely new API to do the exact same job. Likewise for code. Just as Solaris and Linux have similar APIs in some places. But nobody would describe Linux as "Open-Source Solaris", except to idiots.
It makes the house less hot. If 40% of the sun is converted to electricity, then that's 40% which is *not* converted to heat. Decreases the demand for AC.
You're kidding, right? You ever looked at a non-solar roof? You can see the individual shingles, often colored, implying that it already is reflecting at least some sunlight. Just because shingles have 0% conversion to electricity doesn't mean that it's 100% converted to heat.
In fact, the net heat generation could easily be higher with solar. The solar panels are likely specifically designed to maximize absorbtion of sunlight, in the interests of maximizing conversion to electricity, and you can certainly get shingles designed to minimize absorbtion of solar energy.
Many of the points in the interview implied that software was simply soaking up all the hardware performance, and perhaps we could squeeze more out of the software. I completely agree, except...
The problem is that the software is an order of magnitude slower than it needs to be because the hardware has increased in performance by 2 and 3 and 4 orders of magnitude. If we had held the software to the same standards as we used to back when the hardware cost more than the programmers, it would be more efficient - but would only be able to make use of a couple megabyte of RAM and disk. The looseness of current software is part and parcel of harnessing the hardware. The hardware didn't just allow us go loose with the software we wrote - it allowed us to use abstractions which were measurably less efficient, but which had the side effect of allowing us to harness the hardware in the first place.
As a pair of trivial examples, take arrays and dictionaries. When I ask interview questions like "Design a hashtable" or "Reverse a linked list", many candidates have to actually step back and think about the question! 30 years ago, designing a good hashing function was the mark of true talent, and gains were to be had by selecting the linked-list scheme which best suited the problem at hand. These days, many people don't really know why you'd use a map versus a hash_map, or a vector versus a deque. And, for the most part, they don't really need to.
I run Linux because it lets me do things I want to do. There are enough people like me out there to provide a critical mass of contributors, so it's always improving. Ten years ago, it took me three days to a week to get the right combination of kernel patches and hardware to get some particular function operating, nowadays it's more like 15 minutes to an hour.
Is this still intimidating for grandma? Sure is. Who cares? Ford doesn't make the F250 (heavy pickup truck) in the hopes that the average user wants to drive it. Analysts don't come out complaining about it because the average driver doesn't want it.
It doesn't even matter if Linux grows faster than Windows (or MacOS X). All that matters is that Linux grows fast enough to sustain the interest of people developing for Linux. BTW, very very few of those developers are doing it because they want to depose Windows. They're doing it because they like doing it.
Slashdot Mirror
If artificial intelligence ever gets to the point where it is greater than humans, won't it be capable of producing even better AI
:-).
Depends on if it is created by understanding intelligence, or through an accident. We have one example of intelligence created by accident, which has as of yet not managed to create another intelligence.
Even understanding may not suffice. If humans create a human-plus AI and understand how they did it, then it is natural to assume that the human-plus AI could create a human-plus-plus AI, but just as there are many problems with creating a human-plus AI, there might be many problems in creating a human-plus-plus AI, and the improvements we make may not be improvements which lend themselves to the task of creating the next generation of AI. It may be that creating a super-human AI involves solving hundreds of isolated problems over a very long period, and each problem solved may only provide a tiny bit of gain in solving the next problem.
As a specific example, what if we created a human-level AI which had the additional ability to over the long run beat any human at poker? In that way, this AI would be super-human, but the AI may not provide any advantage in creation of the next generation. There are lots of specific human talents one could imagine perfecting in an AI without improving that AI's ability to create a new AI (in fact, many of those talents may be distractions from that job!).
Myself, I await the Singularity
-scott
2. Basestation/routers need a simple-to-configure mode where they will let others into a separate
subnet that goes straight out to the Internet but does not see my home computers directly.
Easier to just have a single open-access mode, and configure your own laptops to VPN into your home network. Also better alignment of the incentives.
-scott
As someone else mentioned, it was more likely Tangents, in a short-story collection of the same name by Greg Bear. The neighbor was based on Alan Turing, he's doing research on visualizing higher dimensions with music, and the kid is able to actually see higher dimensions directly. Really similar to "Mimsy Were the Borogoves" in flavor.
By "Had to explain", I'm sure the article means "As I was running on and on and on about historical crap, someone threw out an 'insight' to distract me."
-scott [a not-20-something Googler]
I liked SUNW. I have fond memories about learning to program on SUN and HP workstations.
What I think is really funny is that I've been developing for workstations and servers for basically 2 decades, and it was only with this thread that I realized that the "W" in "SUNW" is for "Workstation". I always just figured it was a random letter to disambiguate Sun-the-computer-company from Sun-some-other-company. You know, like how AAPL doesn't really stand for Aple? Go figure, this problem probably is the most important thing for Sun to be working on just now.
The event I compare this to is Borland -> Inprise. Basically, the company has finally recognized that it's in dire straits, so it's taking the only thing keeping it ticking in the technical world (the respect engendered by its long and colorful history), and burying it.
-scott
SCO is so dirt-cheap right now that you can basically play with it without risk.
Why was this modded insightful? The price of the stock is completely irrelevant to how much risk you're taking. If SCO did a 10000-to-1 reverse split, would they then be much more risky because their stock was trading at $10k? Contrawise, various brokers allow fractional share purchases, but just because you can buy $100 worth of any stock doesn't mean they're equally risky.
-scott
You're right that there's little evidence to believe that something will evolve to replace it. More likely, we'll go extinct and then some other intelligent species might evolve, but given the billions of years evolution that previously took place, not very likely.
It took billions years for Earth to develop life of any sort, then most of a billion more to develop beyond single cells, then hundreds of millions to develop the diversity of things we recognize as the animal kingdom. But if humans could evolve from the point of the Cretaceous-Tertiary extinction event, then there wouldn't seem to be any reason something similarily-advanced couldn't evolve in 100 million years from a similar event at this time.
-scott
Can you imagine getting on a "soft-key" elevator? I think it would be cool at first, then really annoying.
An elevator should so one thing, and only one thing. You don't watch YouTube videos on your elevator. You don't browse web pages on your elevator. You don't have to adjust the picture-in-picture setting on your elevator.
Also, an elevator is a fixed implementation. Your elevator goes to 23 floors, or it goes to 15 floors, or it goes to 3 floors. You'll never download a firmware update to enable access to three more floors.
Gone are the days when a cellphone did one thing, and you literally only needed the numeric pad to dial with. Nowadays, your cellphone performs dozens of functions, and dialing phone numbers may no longer be the primary function.
I doubt Apple's current iPhone interface will be the one we all think of when we think back on things in five years. They're going to iterate, but given what I've seen comparing the iPod to competitors, I wouldnt' bet against them. Whether a specific implementation is right is a different question than whether their overall design aesthetic is right.
Back in the 80's, as a high school student I used valuable earnings to subscribe to BYTE, Dr Dobbs, and Computer Language. These were all great at the time (BYTE faded a bit, but was great in the mid-80s). Kept them up through college, but then CMP swallowed them all up and crushed them into pulp. They screwed up subscriptions, closed things down and randomly changed publishing schedules, and got rid of anyone with half a brain. So I dropped all those subscriptions, and didn't bother to go back.
Really, I suspect the Internet killed good technical mags. Blogs are just the scapegoat, but once you could look all this stuff and discuss it online, there wasn't as much demand for in-depth technical articles. That left magazines full of vendor-written articles, which nobody was interested in, but at least they were cheap.
Good riddance.
-scott
I could see someone wanting to do blog entries from wherever, and not tolerating the on-screen keyboard.
Someone should almost invent an alternative to text input for such cases. *thinking*
Does Moore have $9,000,000 lying around somewhere?
A websearch on "Michael-Moore net-worth" certainly makes it easy to believe that he's worth a couple tens of millions of dollars. Unfortunately, he's apparently secretive on this front, so there's little in the way of unbiased estimation.
Great idea! There should be an MMO where you get to spend your day as a virtual trader!
You must have been using a different USCD than I recall. My memory was of swapping disks. Edit, save, swap disks, compile, swap disks, run, ad nauseum. You could arrange certain combinations to live on the same disk, but that combination did not include "Everything you needed to do stuff".
Which meant that when we got a Z80 card in one machine and ran Turbo Pascal, which was just everything-in-one-place, it was like heaven.
It really would make a great movie, though I don't think that the translation to big screen would offer anything that the game doesn't already have.
Would you change your mind if they cast "The Rock" in some role or other?
I can't wait to surf Mars. With moons that close, there ought to be tidal swells that one could ride forever.
Dude, Earth's moon is millions of times heavier than the moons of Mars. They're going to have to be pretty damn close to get a tidal swell worth riding, even with the reduced gravity.
I guess that the 9/11 hijackers used library computers doesn't help, nor does the current "Library 2.0" movement to offer customized services.
... why _shouldn't_ the 9/11 hijackers have used library computers? I mean, it's terrible that library computers were used, but it's not like that made them complicit. The hijackers probably also travelled on public roads, and drank water from municipal water supplies, and benefitted from living in a safe neighborhood due to local law enforcement, and used dozens or hundreds of other public services. That's what public services _are_, after, all. Beyond that, they probably bought food in a grocery store, etc, etc. If we start cracking down on libraries because of this - where do we stop?
:-).]
This doesn't sound right, but
[I'm not suggesting that you agree that the above is a good reason to crack down on libraries in any way, I'm just being annoyed that people seem to think we should "crack down" in this kind of thing. I suppose most such people don't even know where their local library is located
-scott
I've encountered CS students recently who in their third year are unable to do such basic things as understand memory allocation.
Right on. But I wouldn't qualify it with "recently". When I was in school, we didn't have this new-fangled Internet stuff - when I was in high school, getting access to a computer _period_ was rough. Which is why my early CS memories are about doing things like writing a Forth kernel in 6502 assembly and simulating execution longhand (ie, on sheets of paper), and spending a long afternoon working around trig tables to implement a 3D rotation library using entirely fixed-point arithmetic.
There is really nothing more frustrating than doing an interview of someone with a masters degree or years of experience, and realize that they don't understand hashtables. I mean, _hashtables_, man! Sigh. Then again, I thought data structures and algorithms was super fun in school, there were these different ways to accomplish things, and knowing the characteristics of each made your programs smaller and faster. I'm just glad that knowing this stuff has turned out to be so rewarding in the real world.
-scott
Indian and Japanese kids kick ALL amercian kids asses hard in education. to the point that the brightest graduating today are considered mentially retarted compared to the average students overseas.
This is clearly why India and Japan so completely dominate the world markets. The only way to end the massive depression the US has been in for the past 20 years is to improve our educational system, otherwise we'll just become a subsidiary of Japan, Inc.
I'm a fan of the Abiogenic theory of oil. ... I think crude oil is just a part of the earth's carbon cycle.
... a couple billion years? It's a little like arguing about the cosmic microwave background radiation, it's an interesting argument, but not terribly practical.
Does it matter? Under one set of theories, we have to bury a bunch of biomass, wait a couple hundred million years, and we have more oil. Under the other theory, the oil is a natural part of planetary development, so we have to wait
*Shrug*. I like Shrek and Shrek 2 a great deal, but they _are_ just a series of set pieces strung together which only works due to the casting. For the most part, Pixar creates pretty amazing movies which feel greater than the sum of their parts in many ways. That doesn't mean that Dreamworks absolutely sucks, it's just a differe style.
It's like Disney versus Warner Brothers. Initially, Warner Brothers was cheap and wanna-be, but over time they really came into their own by developed a cutting wit which simply wasn't present in the Disney pieces. I don't think Dreamworks is there yet, but it could happen.
I mean, this is Kirk. Obviously they'll have to have some sort of deathmatch involving laser boundaries and electrified monkey bars, and perhaps an improvised trebuchet or flintlock.
MacOS X embedded? WTF? I mean, I'm sure it's a wet dream to imagine that you could run the same thing across a bunch of platforms, but ... no, it's not going to be the same thing, or even a very similar thing. In fact, there's a word for it: it will be a different thing.
I mean, look at Windows CE. The main similarity it has to Windows XP is that they both have Windows in their names. Sure, there are APIs which are similar between them - that's because if you have an existing API to do a particular job and it's working fine, you'd be silly to create an entirely new API to do the exact same job. Likewise for code. Just as Solaris and Linux have similar APIs in some places. But nobody would describe Linux as "Open-Source Solaris", except to idiots.
Wait. Oh, OK, I get it. Carry on.
It makes the house less hot. If 40% of the sun is converted to electricity, then that's 40% which is *not* converted to heat. Decreases the demand for AC.
You're kidding, right? You ever looked at a non-solar roof? You can see the individual shingles, often colored, implying that it already is reflecting at least some sunlight. Just because shingles have 0% conversion to electricity doesn't mean that it's 100% converted to heat.
In fact, the net heat generation could easily be higher with solar. The solar panels are likely specifically designed to maximize absorbtion of sunlight, in the interests of maximizing conversion to electricity, and you can certainly get shingles designed to minimize absorbtion of solar energy.
Many of the points in the interview implied that software was simply soaking up all the hardware performance, and perhaps we could squeeze more out of the software. I completely agree, except ...
s /NoSilverBullet.html
http://www-inst.eecs.berkeley.edu/~maratb/reading
The problem is that the software is an order of magnitude slower than it needs to be because the hardware has increased in performance by 2 and 3 and 4 orders of magnitude. If we had held the software to the same standards as we used to back when the hardware cost more than the programmers, it would be more efficient - but would only be able to make use of a couple megabyte of RAM and disk. The looseness of current software is part and parcel of harnessing the hardware. The hardware didn't just allow us go loose with the software we wrote - it allowed us to use abstractions which were measurably less efficient, but which had the side effect of allowing us to harness the hardware in the first place.
As a pair of trivial examples, take arrays and dictionaries. When I ask interview questions like "Design a hashtable" or "Reverse a linked list", many candidates have to actually step back and think about the question! 30 years ago, designing a good hashing function was the mark of true talent, and gains were to be had by selecting the linked-list scheme which best suited the problem at hand. These days, many people don't really know why you'd use a map versus a hash_map, or a vector versus a deque. And, for the most part, they don't really need to.
I run Linux because it lets me do things I want to do. There are enough people like me out there to provide a critical mass of contributors, so it's always improving. Ten years ago, it took me three days to a week to get the right combination of kernel patches and hardware to get some particular function operating, nowadays it's more like 15 minutes to an hour.
Is this still intimidating for grandma? Sure is. Who cares? Ford doesn't make the F250 (heavy pickup truck) in the hopes that the average user wants to drive it. Analysts don't come out complaining about it because the average driver doesn't want it.
It doesn't even matter if Linux grows faster than Windows (or MacOS X). All that matters is that Linux grows fast enough to sustain the interest of people developing for Linux. BTW, very very few of those developers are doing it because they want to depose Windows. They're doing it because they like doing it.