You are a very stupid troll. There are plenty of aspects of evolution that are being challenged. Look at the work being done right now on the Breeder's equation, or really at any area. The problem with people like you is that you tend to treat "evolution" as a single concept.
That's completely wrong. It's like treating physics as a single concept. Evolution is a handful of equations, each of which is constantly being argued and refuted. For the most part, they are becoming more and more accurate (hello, genetic drift equation). Perhaps in Darwin's day you could have called evolution a single concept, but to do so today shows off your ignorance. At the moment there is an overwhelming amount of evidence for evolution of some form, and we (scientists) are simply arguing the details of it.
Mods, please mod this offtopic and mod parent flamebait/troll.
There are two competing ideas here. mestar (GGP post) says that intelligence selects for intelligence (sexual selection), while my evolutionary biology teacher says that there is some benefit to being intelligent (selection unknown). You agree with my evobio teacher, but say that statistical improbability was the cause(?!) of the selection.
I say you are correct that the chances are astronomically slim, and in trying to create an AI we need to break the problem down and find out exactly what it was that allowed those ancient geeks to succeed.
So lets say that a digital organism was placed in an environment where it could enter a "simulation mode" (aka imagination as you call it) and try doing things before it actually did them (and if they were reproductively profitable, follow through and actually do them). Memory is far better on a computer, and in a digital environment, sensors are energy free and perfect.
Then, we use machine learning algorithms (still subjected to random mutations) and attempt to have it predict outcomes of new circumstances.
Fuck, it can't be this simple. I'm missing some obvious barrier and it's going to occur to me five seconds after I press submit.
I disagree that our intelligence is just a program: we have the ability to learn and respond to completely new, unforeseen circumstances.
Machine learning has made leaps and bounds here, but it is still little more than data mining. I'm branching into philosophical territory here, but deductive reasoning and critical thinking are, as far as I know, impossible for a computer (I would love to be proven wrong. Journal articles please).
I've heard that before; however it's still not a reasonable answer to the question. That's like saying intelligence selects for intelligence (which is true). The problem is more a chicken-egg issue: What triggered the first bit of intelligent selection? It has to start somewhere. Peacock's tail works because a male had a brightly colored slightly larger feather that females could use as evidence that the male possessed greater fitness. Ditto to Diopsidae, and most other sexual selections.
How could this have worked in humans? An individual possess more neurons/better neurotransmitters is selected for? Great, we're back to hard coding a massive intelligence and making it subject to selection pressures. The worst of both parts of the AI world.
I'd like to point out that there is a broad range of neural nets beginning with a simple single neuron and working up to humans. Where did sexual selection begin? Are chimps sexually selected for intelligence? If so, what provided a greater pressure not to become as intelligent as us? What about birds? African Grey parrots clearly qualify for intelligence (if only as a starting point). The point I'm making here is that I suspect that sexual selection will undergo a major reworking in the next few years, as we figure out that simply calling something "sexual selection" is hardly better than just saying "god did it" (sorry for the flamebait, can't think of a better analogy). Yes, I've seen the sexual selection equations. Yes, I've worked with them. They need a rewrite, because they seem to be used whenever we can't find an immediate benefit to an adaptation.
On a related note, intelligence (as my evolutionary biology teacher told me) clearly increases fitness in a modern society (we're talking hunter-gatherer to Mesopotamia when we say modern. In the true modern society, all bets are off and we've pretty much rewritten selection).
What a great post. You've accurately summarized the entire AI problem in a couple of paragraphs - however, you're missing an incredibly simple aspect. You hint at it here:
will magically evolve
Ah, yes, that's the key, isn't it? The question to ask is not "how do humans think?", it's "what prompted selection for intelligence in humans?". Combined with massive (and I really do mean massive - the human brain has a faster clock speed and more cores (ugh, bad analogy) than a supercomputer - you'd be able to effectively set up an evolving algorithm and expose it to selection pressure for intelligence. If you know what selects for intelligence, by all means post it here; I've asked every biology teacher I've had since 9th grade and never gotten a reasonable answer.
One other point to consider; organic life works in generations; mutations do not discriminate on the basis of functionality, selection does that. Code that constantly rewrites itself replacing variable names at random and swapping if's for whiles and such (while still correcting syntax - almost every DNA sequence will "compile" into some sort of protein, just most of these new proteins will be useless (or deadly). Weight the randomization algorithm towards replacing commands with other similar commands, as most mutations will be replaced with similar amino acids (IE third base pair mutations for alanine are irrelevant, while second base pair mutations will often replace alanine with a different non-polar amino acid). Note that mutation rate is approximately my chances of getting laid, so you're code still has a good chance to compile in the next version. If it doesn't, consider that mutation selected against. Fork the code about a hundred times per generation, and you're bound to get at least one that's functional. If not, go back a generation and reroll. I'm not a coder, and I have no clue how to create code that self-modifies and self-compiles, but I'm pretty sure these are the basics.
I'm not saying learning algorythms will be easy, but they may just happen in our lifetimes. The memresistor may help speed things along; we'll see in the next couple of years when memresistor RAM comes out.
Not a problem. With sufficient brainwashing and hardware controls/overrides (or perhaps even in a software interpretation layer) the catbot wouldn't even think about getting out of line. There are two big problems here, though:
1) The technology is not yet ready to interpret signals from neurons in the brain. MRIs are close, but we basically need to be able to receive and decode thousands of signals per second and do it on a neuron by neuron basis. My home computer is probably fast enough to receive these signals, but we need some sort of yet-to-be-developed biochemical technique to gather them and even then we have no fucking clue what most of them do. See the work being done right now on Roadrunner with human vision systems, that has the potential to be groundbreaking here.
2) Cats are fucking stupid. Why not use a more intelligent animal? You might as well just use a VI to replace most of the stuff a cat could be taught to do; it'd be easier. Perhaps this idea would be viable with a human brain, but there are "moral issues" (whatever that means) involved with this idea.
Idiot. Space junk is a growing problem, but we are a long, long way from having it block sunlight.
Space is big. Really, really big. Satellites are small. I'm not really sure how many we have in orbit, but it was ~4000 10 years ago, and who knows how many now but we're still talking about a tiny fraction of a percent coverage.
Now the greenhouse effect is always at a tipping point since most radiation is held in or out by water vapor, which is why a CO2 increase will throw things off balance - I think you are implying that satellites will do the same thing. They will, just let me know when we have billions of tons of them (hint: that's decades away)
Re:Nov 23, 1987 - 1st documented use
on
Spam Is 30 Years Old
·
· Score: 2, Informative
I agree; and I tracked down the original for you to click on.
I read The Beak of the Finch a while ago (awful book, skip it) and it attempts to give a number of case studies of speciation (the only good one was an experiment involving breeding fruit flies in the dark; within a few generations they become blind (and therefore basically unable to mate with individuals of their original species - making them technically a new species)
Anyway, it's not hard to trace the changes leading one species to another. I read stuff about Jack Horner modeling hypothetical intermediate dinosaurs; IIRC, someone later found a new species very similar to what he predicted.
There's a few other cases; walking sticks with and without a white stripe on their backs found in different vegetation types (still technically the same species, but becoming different).
There's a million more examples; nothing perfect yet, but, our current understanding is that macroevolution takes centuries (this will be proven wrong by a case study sometime in the next decade, I guarantee it. We are quite close right now, and remember that Darwin thought microevolution took millenia; we now know it can take as little as a generation - see the Grant's study on Galapagos Finches).
Or, if you want decisions to have even more impact, The Witcher. It really felt like the (dozens of) major decisions you made in that game changed the world. Much better than most RPG's the same no matter what until a "good/bad/mercenary" ending.
Do sarcastometers use a logarithmic or linear scale?
Both. The OpenSarcasm-Dry fork uses logarithmic after v.98, while the vanilla uses linear to maintain computability with Apple's iWitty and Microsoft's SarcastXP/2008 formats.
But what is the advantage to learning how to fix nitrogen if there is already sufficient amounts around?
It's all about evolution. Sure, you* have enough NH3 to survive, even to grow, but there's millions of tons of N2 gas in the atmosphere, and if you could somehow use that as a fuel source, you'd be set for life.
So then, along comes a random mutation in an enzyme that pulls converts nitrites to nitrates (I'm making this up - but it was probably some enzyme to do with N). Rather than killing you, it allows you to pull N2 out of the air and turn it into ammonia, allowing you to reproduce more quickly. Now another mutation comes along, and it allows you to use Mo to push forward the reaction (mind you it worked before you had Mo: reactions can generally go forward without their cofactors, just more slowly.)
With this cofactor, you're able to reproduce much more quickly than your neighbors which don't have the mutation, and you become the bacteria we know today.
*you here refers to a now-extinct progenitor of nitrogen fixing bacteria. Individuals reproduce, populations evolve.
So maybe it was mostly nitrogen fixers around, but they weren't doing very well because of a lack of molybdenum?
Like I said, cofactors generally speed up a process. They are not generally required for the reaction to happen, they just speed it up (by several orders of magnitude) when they are present.
Nope. You lose. Bacteria/Prokaryotes/lots of protolife evolved first. I don't have a Biology textbook in front of me, but here is an online reference.
IAAB (student)
Overextending yourself into lots of unrelated areas, in the hope that a few of them will turn into cash cows, has been Microsoft's strategy for a long time now. It's a great way to lose mindshare (how many people equate Microsoft with bugs and crappy software?), and make massive profits (Windows, Office, hell, even the 360).
Google has, so far, enjoyed ample mindshare (almost everyone, myself included is a Google "fanboy"), but if they keep overextending themselves, they will crash and burn before they realize what's happening.
Besides, this entire Web 2.0 replacing operating systems debate is irrelevant anyway, because the majority of the world still doesn't have the high-speed (DSL or better) required for it. Let's reevaluate in 10 years or so, and then we can discuss whether or not Google docs is better for Joe Average than a pirated copy of Word (that is, I think, their prime competitor).
In ten years, you won't be able to send data over the internet unmonitored. The basis for this new law will be that "you can't physically take data across the border without it being searched."
Damn, I really, really, hope that's just my tinfoil hat talking.
Is that a serious question? The answer is no, because the only way to stop evolution is to extinguish all life as we know it.
As long as any organism is alive and has the ability to reproduce with genetic drift, life will continue to evolve. Besides, our predictions of global temperature increase by the end of the century are all below increases of 15C. Species which are adapted to higher temperatures, like Thermus aquaticus, will certainly not be wiped out by global warming, they will continue to evolve.
I'm browsing the internet right now on an 800mhz laptop with 256 megs of RAM. If Ubuntu + KDE can make a system that looks clean (better than Aero, IMHO) and runs well enough to browse the web (startup is a little slow, but Firefox runs great), why can't Microsoft?
I should note that this system came with XP, and that ran fine on here before Ubuntu. Given that everyone I know uses their PC for word processing, email and web browsing, why should they drop $1000 on a new PC when they can upgrade to Ubuntu, and get all the eye candy. Did I mention that they can stop paying for anti-spyware and anti-virus, too?
Mininova is a pretty good substitute for Demonoid. They are easily searchable, have a good selection, and a "thanks" feature which discourages spam and fakes.
The only problem is that their comments are on a separate page from the details of the torrent. Also, they are not private.
If your coworker really spent a thousand dollars on video cards, he is a moron. You can get a good gaming rig that will run anything except the very latest games at highest settings for less than $500 (less if you buy a $300 Dell and drop in a 7900GT). If you want to run everything at max settings, yes, you'll have to spend a thousand dollars, but compare that to consoles, where you must buy the console itself and a HDTV to go with it. IMHO, PC gaming is much, much less expensive than owning a 360.
Slashdot Mirror
Or, it could just provide more evidence that life originates in space.
You are a very stupid troll. There are plenty of aspects of evolution that are being challenged. Look at the work being done right now on the Breeder's equation, or really at any area. The problem with people like you is that you tend to treat "evolution" as a single concept.
That's completely wrong. It's like treating physics as a single concept. Evolution is a handful of equations, each of which is constantly being argued and refuted. For the most part, they are becoming more and more accurate (hello, genetic drift equation). Perhaps in Darwin's day you could have called evolution a single concept, but to do so today shows off your ignorance. At the moment there is an overwhelming amount of evidence for evolution of some form, and we (scientists) are simply arguing the details of it.
Mods, please mod this offtopic and mod parent flamebait/troll.
I say you are correct that the chances are astronomically slim, and in trying to create an AI we need to break the problem down and find out exactly what it was that allowed those ancient geeks to succeed.
Then, we use machine learning algorithms (still subjected to random mutations) and attempt to have it predict outcomes of new circumstances.
Fuck, it can't be this simple. I'm missing some obvious barrier and it's going to occur to me five seconds after I press submit.
Machine learning has made leaps and bounds here, but it is still little more than data mining. I'm branching into philosophical territory here, but deductive reasoning and critical thinking are, as far as I know, impossible for a computer (I would love to be proven wrong. Journal articles please).
How could this have worked in humans? An individual possess more neurons/better neurotransmitters is selected for? Great, we're back to hard coding a massive intelligence and making it subject to selection pressures. The worst of both parts of the AI world.
I'd like to point out that there is a broad range of neural nets beginning with a simple single neuron and working up to humans. Where did sexual selection begin? Are chimps sexually selected for intelligence? If so, what provided a greater pressure not to become as intelligent as us? What about birds? African Grey parrots clearly qualify for intelligence (if only as a starting point). The point I'm making here is that I suspect that sexual selection will undergo a major reworking in the next few years, as we figure out that simply calling something "sexual selection" is hardly better than just saying "god did it" (sorry for the flamebait, can't think of a better analogy). Yes, I've seen the sexual selection equations. Yes, I've worked with them. They need a rewrite, because they seem to be used whenever we can't find an immediate benefit to an adaptation.
On a related note, intelligence (as my evolutionary biology teacher told me) clearly increases fitness in a modern society (we're talking hunter-gatherer to Mesopotamia when we say modern. In the true modern society, all bets are off and we've pretty much rewritten selection).
Ah, yes, that's the key, isn't it? The question to ask is not "how do humans think?", it's "what prompted selection for intelligence in humans?". Combined with massive (and I really do mean massive - the human brain has a faster clock speed and more cores (ugh, bad analogy) than a supercomputer - you'd be able to effectively set up an evolving algorithm and expose it to selection pressure for intelligence.
If you know what selects for intelligence, by all means post it here; I've asked every biology teacher I've had since 9th grade and never gotten a reasonable answer.
One other point to consider; organic life works in generations; mutations do not discriminate on the basis of functionality, selection does that. Code that constantly rewrites itself replacing variable names at random and swapping if's for whiles and such (while still correcting syntax - almost every DNA sequence will "compile" into some sort of protein, just most of these new proteins will be useless (or deadly). Weight the randomization algorithm towards replacing commands with other similar commands, as most mutations will be replaced with similar amino acids (IE third base pair mutations for alanine are irrelevant, while second base pair mutations will often replace alanine with a different non-polar amino acid). Note that mutation rate is approximately my chances of getting laid, so you're code still has a good chance to compile in the next version. If it doesn't, consider that mutation selected against. Fork the code about a hundred times per generation, and you're bound to get at least one that's functional. If not, go back a generation and reroll. I'm not a coder, and I have no clue how to create code that self-modifies and self-compiles, but I'm pretty sure these are the basics.
I'm not saying learning algorythms will be easy, but they may just happen in our lifetimes. The memresistor may help speed things along; we'll see in the next couple of years when memresistor RAM comes out.
1) The technology is not yet ready to interpret signals from neurons in the brain. MRIs are close, but we basically need to be able to receive and decode thousands of signals per second and do it on a neuron by neuron basis. My home computer is probably fast enough to receive these signals, but we need some sort of yet-to-be-developed biochemical technique to gather them and even then we have no fucking clue what most of them do. See the work being done right now on Roadrunner with human vision systems, that has the potential to be groundbreaking here.
2) Cats are fucking stupid. Why not use a more intelligent animal? You might as well just use a VI to replace most of the stuff a cat could be taught to do; it'd be easier. Perhaps this idea would be viable with a human brain, but there are "moral issues" (whatever that means) involved with this idea.
Disclaimer IANAAIResearcher (who is, honestly?)
I find your ideas disturbing and wish to unsubscribe from your newsletter.
Space is big. Really, really big. Satellites are small. I'm not really sure how many we have in orbit, but it was ~4000 10 years ago, and who knows how many now but we're still talking about a tiny fraction of a percent coverage.
Now the greenhouse effect is always at a tipping point since most radiation is held in or out by water vapor, which is why a CO2 increase will throw things off balance - I think you are implying that satellites will do the same thing. They will, just let me know when we have billions of tons of them (hint: that's decades away)
http://groups.google.com/group/news.admin/msg/483c12f48d13225e?output=gpl
Slashdot should block *.tinyurl.com and that yahoo scripts place, as well as prevent AC's from posting live links.
Anyway, it's not hard to trace the changes leading one species to another. I read stuff about Jack Horner modeling hypothetical intermediate dinosaurs; IIRC, someone later found a new species very similar to what he predicted.
There's a few other cases; walking sticks with and without a white stripe on their backs found in different vegetation types (still technically the same species, but becoming different).
There's a million more examples; nothing perfect yet, but, our current understanding is that macroevolution takes centuries (this will be proven wrong by a case study sometime in the next decade, I guarantee it. We are quite close right now, and remember that Darwin thought microevolution took millenia; we now know it can take as little as a generation - see the Grant's study on Galapagos Finches).
Or, if you want decisions to have even more impact, The Witcher. It really felt like the (dozens of) major decisions you made in that game changed the world. Much better than most RPG's the same no matter what until a "good/bad/mercenary" ending.
Remember, XP was competing against Win 98! Windows 2000 wasn't very popular outside of servers.
Things start to look a bit different now, don't they?
Both. The OpenSarcasm-Dry fork uses logarithmic after v.98, while the vanilla uses linear to maintain computability with Apple's iWitty and Microsoft's SarcastXP/2008 formats.
It's all about evolution. Sure, you* have enough NH3 to survive, even to grow, but there's millions of tons of N2 gas in the atmosphere, and if you could somehow use that as a fuel source, you'd be set for life.
So then, along comes a random mutation in an enzyme that pulls converts nitrites to nitrates (I'm making this up - but it was probably some enzyme to do with N). Rather than killing you, it allows you to pull N2 out of the air and turn it into ammonia, allowing you to reproduce more quickly. Now another mutation comes along, and it allows you to use Mo to push forward the reaction (mind you it worked before you had Mo: reactions can generally go forward without their cofactors, just more slowly.)
With this cofactor, you're able to reproduce much more quickly than your neighbors which don't have the mutation, and you become the bacteria we know today.
*you here refers to a now-extinct progenitor of nitrogen fixing bacteria. Individuals reproduce, populations evolve.
Like I said, cofactors generally speed up a process. They are not generally required for the reaction to happen, they just speed it up (by several orders of magnitude) when they are present.
(Mo is used as a cofactor, meaning that it can be used over and over again without being depleted. You just need a single atom of Mo per enzyme.)
Nope. You lose. Bacteria/Prokaryotes/lots of protolife evolved first. I don't have a Biology textbook in front of me, but here is an online reference. IAAB (student)
Google has, so far, enjoyed ample mindshare (almost everyone, myself included is a Google "fanboy"), but if they keep overextending themselves, they will crash and burn before they realize what's happening.
Besides, this entire Web 2.0 replacing operating systems debate is irrelevant anyway, because the majority of the world still doesn't have the high-speed (DSL or better) required for it. Let's reevaluate in 10 years or so, and then we can discuss whether or not Google docs is better for Joe Average than a pirated copy of Word (that is, I think, their prime competitor).
Damn, I really, really, hope that's just my tinfoil hat talking.
Is that a serious question? The answer is no, because the only way to stop evolution is to extinguish all life as we know it.
As long as any organism is alive and has the ability to reproduce with genetic drift, life will continue to evolve. Besides, our predictions of global temperature increase by the end of the century are all below increases of 15C. Species which are adapted to higher temperatures, like Thermus aquaticus, will certainly not be wiped out by global warming, they will continue to evolve.
Where's the -1 Misinformed mod when you need it?
I should note that this system came with XP, and that ran fine on here before Ubuntu. Given that everyone I know uses their PC for word processing, email and web browsing, why should they drop $1000 on a new PC when they can upgrade to Ubuntu, and get all the eye candy. Did I mention that they can stop paying for anti-spyware and anti-virus, too?
The only problem is that their comments are on a separate page from the details of the torrent. Also, they are not private.
All in all, I consider them good enough. http://www.mininova.org/
Shift is force move,
Alt is force shortcut, and
Ctrl is force copy.
IIRC, the shortcuts in KDE are the same. I prefer KDE's menu, however, since Windows defaults to copy between hard drives, and move between folders.
If your coworker really spent a thousand dollars on video cards, he is a moron. You can get a good gaming rig that will run anything except the very latest games at highest settings for less than $500 (less if you buy a $300 Dell and drop in a 7900GT). If you want to run everything at max settings, yes, you'll have to spend a thousand dollars, but compare that to consoles, where you must buy the console itself and a HDTV to go with it. IMHO, PC gaming is much, much less expensive than owning a 360.