It's not just that. If you look at heart rate versus life span you find that the number of heart beats in a lifespan for different species is a constant (so think of it as related to the metabolic time). A physicist named Geoffrey West is the one who worked that (and a bunch of other scaling laws) out:
Um, the periodic table is just a consequence of the mathematics of 3D spherically symmetric waves. Namely, the number of electrons per period is just the number of eigenvalues with a unique energy in the energy well formed by the nucleus. And the shape of that energy well is determined by the same mathematics, so really it's all just the consequence of having a central point and standing waves which interact with that central point, and having three dimensions.
Here's how it works. You have a central point, with some flux of force-carrying particles. If those particles are conserved (mass-less) as they are in the case of photons, the density of those particles must go as 1/r^2 from the source point. The energy of something that interacts with those force carriers will be:
integral[r0..r1]{F dr} = J/r0-J/r1 where J is the coupling constant. If that potential is attractive (J negative) then the energy with respect to infinity at a point r is -|J|/r.
Now, we want to solve for the standing-wave states around such a potential well. The energy of a particle in such a well is the sum of its kinetic energy (p^2/2m) and that potential (-|J|/r). The operator p_x corresponds to hbar/i d/dx, and so on (which you can get by looking at the properties of an operator which translates the system in space, and showing that for the behavior of a wave with only a momentum part to translate the wave according to its momentum, it requires that the momentum operator be hbar/i d/dx...)
If you want to see more about how that's derived using the generator of translations and so on, check out:
Where lapl is the laplacian operator (d^2/dx^2 + d^2/dy^2 + d^2/dz^2 in cartesian coordinates)
Write the whole thing in polar coordinates, it's a separable linear second order eigenvalue problem (solve for the permissible values of E). So you have different solutions in the three directions (r, theta, phi). The radial solutions have different energies - these are the 'n' levels in the periodic table, each of which is a row. The solutions along theta and phi are the 'l' and 'm' quantum numbers, which correspond to different elements within a row, and why there are a different number as 'n' increases. It is because the values of l and m are bounded by the values of n.
There is another piece of information you need, which is that the electrons surrounding the nucleus have an internal degree of freedom - spin. So everything gets multiplied by two (one state for spin up, one state for spin down).
This is why the first row has two elements, the second row has eight, etc.
Now, the trick is that by this logic, the third row should have eighteen elements, but it again only has eight. The reason here is that the many electrons around the nucleus aren't really non-interacting as has been assumed up to this point, and there are other small effects one needs to take into account to get the energies exactly right (the momentum operator I've used is non-relativisitic, and there are interactions between the spins of the electrons and the spins of particles in the nucleus) So the higher angular momentum (l) states don't have quite the same energy as the lower ones. In fact, n=2, l=3 has a higher energy than n=3, l=2 so n=3, l=2 fills up first.
And thats my attempt at cramming the highlights of a one-semester QM course into a/. post to explain the periodic table. I've left out a few things, like how to derive the Schroedinger equation (that Hpsi=Epsi bit) and how to actually solve the thing you get. A quick search revealed the following sites, which are probably a good place to go for those wan
There's a lot of information in genetic material - you'd need to wait for a fluctuation which just happens to produce that information. The kind of causality loop you're talking about would be much more likely to occur on atomic levels only (for instance, a photon forming a positron and electron, which then re-collide forming a photon again would qualify).
For it to occur macroscopically you'd have to have some way to engineer the loop (i.e. make it so that one particular set of fluctuations is the only one which can create a consistent loop), and you'd probably have to put enough energy into the thing to form everything that occurs in the loop (so a human-mass or two worth of energy).
Of course, if you're looking at it in the sense of photon->matter+antimatter->photon you don't really get what's traditionally thought of as timetravel - its not the kind of thing where you could say 'I want to go to the year 1532 AD'
It's hard to accelerate a person uniformly. Best way to do this would be to fall down a steep and very long gravity well. Of course, you also have to consider tidal forces there, but they can be reduced if you make the system even larger. However, a traditional means of acceleration (i.e. car ride) will accelerate with a very sharply falling off force, applied first to the layer of atoms at your back/feet/whatever part, which then must push the rest of your body. In extreme cases, the resulting wave of compression isn't very good for your internal organs.
That experiment is like a single organism floating around and waiting till its better, which isn't really what natural selection gives you. A better experiment would be to take 500 dice and 500 counters, roll the dice, update the counters, and now: throw away the lowest 100 counters and replace them with 100 counters taken from the group of 400 remaining. Now repeat. I guarantee you'll get to 20 faster than 222,155,644 generations.
Short chains of RNA can replicate themselves because each base tends to bond selectively with its conjugate base. It doesn't require any external machinery, just the presence of bases in solution. Of course, the process can take a LONG time (doubling time about a year). The paper I'm thinking of (Trinks, Schroeder, Biebricher "Ice and the Origin of Life" Kluwer Academic Publishers, 2004) studies this process occuring in arctic ice floes.
If you're willing to allow for the presence of a single enzyme, then you can get self-replicating DNA near hydrothermal vents. The hot spot sets up circulation, so the entire thing acts like a PCR chamber with a doubling time of about a minute (Braun, Goddard, Libchaber "Exponential DNA Amplification by Laminar Convection" Phys. Rev. Lett 91, 158103)
Well, since most people who review papers don't get paid, and editing is done by the person who submitted the paper in the first place... I really don't see much cost anywhere in there. Unless the cost of an email has suddenly become a few hundred bucks.
I believe that most of the cost is in fact related to publishing, since the cost of publishing a paper (to the person submitting it), depends strongly on the contents, i.e. how many figures, whether they're in color, etc. Whether a figure is in color or not can make a difference of hundreds of dollars. But that wouldn't affect anything other than printing/publishing costs.
Well, it doesn't matter how the numbers are generated as long as they end up having the exact same distribution. Chaotic processes, or an entity outside the universe playing bingo and calling out the numbers. There is no measurable difference between the two. So if thats free will, then a sufficiently good random number generator also has free will. There's nothing measurable that can draw a distinction between a perfect RNG and 'unpredictability due to free will'. This is a very general argument - it simply works on the separation of things into what must depend on an earlier state and what cannot depend on an earlier state, it has nothing to do with the structure of the brain or any specific cognitive system.
For any sufficiently long period of time after event, the outcome of a single change is likely to be random anyhow, so it probably wouldn't really matter. Of course, the necessary time may be very long for a well-chosen point of change. So I wouldn't think it's really that much of a risk to change the past, but its probably not all that useful unless you make very well-studied choices. Killing a single historical figure is not likely to create such long-term significant changes if the things that let that figure have the opportunity to become renowned are still present. For example, if you went back and killed Newton, its unlikely that no one would ever uncover the mysteries of 'Newtownian' mechanics, it'd probably just take longer. If you killed Hitler, it might very well be that you'd end up getting another war shortly after, maybe even a war with similar participants and similar atrocities.
But randomness itself has predictable properties. - mean, deviation, higher moments of the distribution of outcomes... If there's a random element to a person's actions, that by definition means that that element has no connection to the person's current state. It can't really be seen as 'will' because there's nothing behind it. It's 'predictably unpredictable'.
Trying to put detailed controls on the things is probably a losing proposition. Think of it this way - a bacterium with those controls is less well-adapted than one without it, so if you're trying to introduce that control as a protection against a hostile mutation, then its just as likely that a mutation removes the control as it is that it introduces some other negative effect.
As far as the dangers of genetic engineering, I'm not all that convinced there's much we can do to directly make something that accidentally becomes worse than the stuff already out there. Bacteria have short life cycles so they're one of the few things that can evolve about as fast as we can think up new stuff. So really we have more to fear in the sense of introducing new selection pressures (i.e. antibacterial soap) rather than having some lab-produced superbacterium (if its that good, it would've been done some time in the last billion years). After all, bacteria evolved us over that time span, and thats a far more complex change than tinkering with a few base pairs is going to cause.
So we might as well start and see what we can come up with. Intestinal bacteria do seem like a good place to start: since they already have a symbiotic relation with us, you'd expect it to be easy to introduce human-useful variations into the population without causing a major upset like trying to introduce a totally foreign species of bacteria.
I'm curious though: it takes a fair amount of time to actually wipe out say, 10gb of data. So I wonder if you rm -rf / and then immediately hit ctrl-c just how much you lose (or if you even really lose anything since rm probably just removes the file entry, as opposed to overwriting the data with zeros or noise...)
In this case its the author of the code who's doing it. I'd be worried however if someone were pressing a case for a GPL violation without the author's consent, since even if someone chooses to license under the GPL they're also free to license under alternate schemes to select other groups at their whim.
Well, thats the point of this paper. In a nutshell, what this paper is saying is that there's an inconsistency between general relativity and quantum mechanics unless you meet a certain condition, and black holes don't meet that condition.
Now, the only thing is I'm not quite sure where the condition he's specifying comes from; the author says 'what time do you mean when you write down the Schroedinger equation' but that argument seems insufficient to me, since you could say the exact same thing about 'what time do you mean when you write down classical hydrodynamic equations', with the answer being 'if you do everything in local variables, you get something consistent with GR'. So I can't immediately see why that doesn't apply to the Schroedinger equation as well since its just a wave equation, and those can certainly be done in GR.
The more telling example I think is the second one, which is an inconsistency with quantum measurements - the 'non-local correlations' he talks about, in which the inconsistency would be a consequence of quantum measurement theory (i.e. entanglement). I wish he gave a reference on that, or at least more detail, since I could easily imagine that that problem 'goes away' the same way it goes away if you try to use entanglement to send information faster than light. The result is that even though the entanglement effect is instantaneous, it can't transmit information. So I have to wonder if a similar thing would occur as to the problems introduced by not having an absolute time. However, it may very well be that it doesn't work out.
The thought experiment he talks about in this paper is rather interesting and there might very well be something there. At the least, its very 'cute' (in so far as such a thing can be considered cute) as it pulls in most areas of modern physics (astrophysics, condensed matter physics, particle physics...) into a single case, where the math just happens to be pretty much identical between all three views (its the lowest order theory for interacting 'stuff', so thats not too surprising).
Well anyhow, thats just this physics grad student's 3am analysis.
This isn't necessarily a recent development though. I'm sure we had the same sort of paranoia during the cold war - the very essence of the cold war was paranoia about communism and communist world powers at some future point acting against the US.
Well we do have programs that can parse english sentences (and those from other languages) and create a probable tree structure (noun, verb, adj, etc). There are of course sentences which are meaningless without 'understanding' of context even though they're grammatically valid, and so on, but its a start.
So the next step would be, as they're trying to do, adding that context-sensitivity which is a form of understanding. But understanding is not the same as thinking: thinking would involve taking all the things which have been processed and using them to achieve a goal (or at least synthesizing the ideas into a new idea). It'd be interesting if that could be achieved 'simply' by applying transformations on the sentences designed to preserve logical relations, like the way a symbolic math program simplifies and solves equations (it just rearranges the symbols in logically equivalent ways).
Depending on what exactly you're trying to figure out, you might not even need to simulate multiple molecules. If you just want to know if two relatively small organic compounds will react to form other organic compounds, you could probably at this point in time do it with a quantum chemistry package like MPQC (among others). Ghemical (http://www.uku.fi/~thassine/ghemical/) will interface with those. But I should warn you, even calculating a 10-atom system with dynamics will take ages. But this is pretty much as 'complete' as you're going to get. If you could do it on a beaker, you'd get all the fluid dynamics, crystallization, etc without having to add in any extra rules.
If you know what reactions will occur, then you can model those with rate equations and go up to the fluids level (so no more dealing with individual molecules). This will be a lot faster, and you can get such interesting things as reactions that form spirals and other spatially-extended patterns.
What about a job in academia (tenure), government,... thats about all I can think of but there's probably more examples where the bottom line isn't a guillotine.
Slightly offtopic, but I have a story from my experience of playing around with electrolysis and driveway cleaning compounds in my youth...
A solution of 6M HCl will slowly acquire copper ions when copper wires are used to electrolyse it. I have a suspicion this might be from chlorine formed at the anode reacting with the copper, but I don't know for sure. You get a green liquid, though I don't know whether its just Cu(1+) or if its some complex ion. Cu(2+) gives a blue liquid, I think. I may have those backward though, its been awhile.
Anyways, the resulting mix seems to eat through alumina with no problem, which was fun to demonstrate using aluminum foil to produce hydrogen gas. The 6M HCl alone won't get through the alumina. I don't think I ever found out exactly how that worked (you get what looks like copper metal in the resulting mush, but that could just be competition between CuCl and HCl in reacting with the aluminum metal beneath.) Any inorganic chemists here who want to hazard a guess?
What about using an amorphous form of tin rather than crystaline? I.e., quench the solder very quickly to room temperature rather than letting it cool slowly. If its due to either compressive stresses or oxygen, the problem is that when it expands, it expands preferentially along certain directions and at certain sites. If its amorphous, it shouldn't have any preferred directions, so instead of forming a whisker it just grows or shrinks a bit all over.
I imagine thats the sort of thing that adding a few impurities does, i.e. preventing a perfect crystal from forming. One question I have is what determines the radius of the whisker. If the entire sample is a perfect crystal, do you still get a localized whisker, or is it only if you have many grains of crystaline tin, each with a different orientation (so the whisker corresponds to a single grain).
Of course, if its some kind of diffusive dendritic growth thing, then that process will form crystals regardless of whether the underlying structure is amorphous or perfectly crystaline (grains might matter). But coatings should probably prevent processes of that kind.
I'd disagree. If you know the bias, you can just dismiss everything that seems unpleasant to you by saying 'oh, thats just their X bias' whether or not it actually is part of their bias or not. If their bias takes form entirely in omission/inclusion selectivity (I'll assume for now they don't write anything thats actually false) then you'll stand a good chance of dismissing facts because of a disagreement of bias.
On the other hand, if the bias is totally random then you really do have to take each thing stated and ask yourself 'is this true? is this the _only_ thing I have to know to understand this situation? is this an emotional argument or a rational one?'. And you have to do it every paragraph. So as a result you end up thinking critically about the ideas, rather than about the author. And hopefully if the bias is random enough then you approach neutrality simply by averaging, the same way that you (should) do to get a balanced view of a situation (i.e. consulting many different sources).
Re:Seriously... Why would you use this?
on
GIMP 2.2 Released
·
· Score: 1
The scale you need to use varies depending on the outputs. What you're describing is basically gamma correction. For instance, my laptop display shows differences in dark colors very well, so red 0 and red 1 are easily distinguishable, but not 254 vs 255. However, my CRT is the opposite. You can easily distinguish 255 vs 254 on my CRT, but not 0 and 1 (or even 0 and 20 unless the lights are out and its night time and you sit about 4 inches from the screen).
One problem (and I don't know how digital cameras handle this) is that some times you want to accurately store intensities that go way beyond the scale (logarithmic intensities). With 8 bits you can only store 3 orders of magnitude change. But if you were to take a picture of a grassy field or something, the sky is (really off the top of my head here) maybe a thousand times brighter than the grass. So you end up with either normal sky and black grass or green grass and pure white sky. With a greater color depth you can correctly store these values and then use postprocessing to try and correct for the effect. But with 8 bits, it would simply get clipped at 255 or 0 and you'd lose that data.
The lack of free will does not necessarily mean that rights are meaningless. Instead of looking at rights as something fundamental, granted by the universe or god or whatever, instead you can look at rights as a means to stabilize social structures. That is, if I and those around me don't respect other people's right to live, then we will tend to destabilize the society around me, and it will either act to restabilize itself (i.e. kill/imprison those that kill) or will break apart (everyone starts killing eachother).
So based on that, the question of machine rights would simply be 'if we do give machines rights, how will it change their behavior compared to not giving machines rights?' In a way, we already do give machines rights, except that they're tied in with the rights of their owner. They have the right to not be broken into by a third party, they have a right to not be attacked via some denial-of-service scheme, etc. However, we see these simply as continuations of the rights of the owner. The way to introduce further machine rights then is, if an AI is written which its owner feels attached to, that owner will demand those rights for the AI: not to be deactivated without permission, the ability to hold financial resources (for when the user wants his or her AI to do some stock trading for them), the right to not incriminate onesself (this would be defense against law enforcement demanding crypto keys, seizing the machine, etc), and so on.
Of course, another way would be to develop the technology that lets a computer emulate the structures in a particular human's brain, so that the computer's output is indistinguishable from that person. Then simply have a computer emulate a dead person for 10 years or so, pretending to be that person living in seclusion, and when it comes to court there will be a good body of evidence that a computer and a person can be interchangeable.
Without power to gain, there is very little motivation for war. However, removing economic power still leaves other kinds of power to gain. If what you were saying were true, that is, that services were no longer necessary, then it _would_ remove most of the motivation for wars (why have slaves?).
However, services will still be necessary since being able to build something atom-by-atom doesn't mean that you can _design_ the thing atom-by-atom yourself. You still need to get someone to figure out the layout for you, you still need to get someone (or some machine at least) to get the raw materials to your doorstep, you still need people to transport you places (unless of course everyone learns to fly a plane - not an instantaneous thing).
The problem with being cautious here is, if we take things slowly, then people who stand to lose status by the introduction of new technology will stand in its way, legislate it to death, and it will end up being something that only a few people get the benefit from (or, a thing which everyone benefits from illegally). If we just jump into it then there's no time for that, and whatever we settle into is much closer to what you'd get if we had had that technology all along.
Slashdot Mirror
It's not just that. If you look at heart rate versus life span you find that the number of heart beats in a lifespan for different species is a constant (so think of it as related to the metabolic time). A physicist named Geoffrey West is the one who worked that (and a bunch of other scaling laws) out:
. html
http://www.physicscentral.com/action/action-03-01
Um, the periodic table is just a consequence of the mathematics of 3D spherically symmetric waves. Namely, the number of electrons per period is just the number of eigenvalues with a unique energy in the energy well formed by the nucleus. And the shape of that energy well is determined by the same mathematics, so really it's all just the consequence of having a central point and standing waves which interact with that central point, and having three dimensions.
...)
/. post to explain the periodic table. I've left out a few things, like how to derive the Schroedinger equation (that Hpsi=Epsi bit) and how to actually solve the thing you get. A quick search revealed the following sites, which are probably a good place to go for those wan
Here's how it works. You have a central point, with some flux of force-carrying particles. If those particles are conserved (mass-less) as they are in the case of photons, the density of those particles must go as 1/r^2 from the source point. The energy of something that interacts with those force carriers will be:
integral[r0..r1]{F dr} = J/r0-J/r1 where J is the coupling constant. If that potential is attractive (J negative) then the energy with respect to infinity at a point r is -|J|/r.
Now, we want to solve for the standing-wave states around such a potential well. The energy of a particle in such a well is the sum of its kinetic energy (p^2/2m) and that potential (-|J|/r). The operator p_x corresponds to hbar/i d/dx, and so on (which you can get by looking at the properties of an operator which translates the system in space, and showing that for the behavior of a wave with only a momentum part to translate the wave according to its momentum, it requires that the momentum operator be hbar/i d/dx
If you want to see more about how that's derived using the generator of translations and so on, check out:
http://www.math.ohio-state.edu/~gerlach/math/BVtyp set/node111.html
So anyhow, we want to solve: Hpsi = Epsi where H is p^2/2m - |J|/r
p^2/2m = hbar^2 / 2m ( d^2/dx^2 + d^2/dy^2 + d^2/dz^2 )
H = p^2/2m - |J|/r
hbar^2/2m lapl(psi) - |J|psi/r = E psi
Where lapl is the laplacian operator (d^2/dx^2 + d^2/dy^2 + d^2/dz^2 in cartesian coordinates)
Write the whole thing in polar coordinates, it's a separable linear second order eigenvalue problem (solve for the permissible values of E). So you have different solutions in the three directions (r, theta, phi). The radial solutions have different energies - these are the 'n' levels in the periodic table, each of which is a row. The solutions along theta and phi are the 'l' and 'm' quantum numbers, which correspond to different elements within a row, and why there are a different number as 'n' increases. It is because the values of l and m are bounded by the values of n.
There is another piece of information you need, which is that the electrons surrounding the nucleus have an internal degree of freedom - spin. So everything gets multiplied by two (one state for spin up, one state for spin down).
This is why the first row has two elements, the second row has eight, etc.
Now, the trick is that by this logic, the third row should have eighteen elements, but it again only has eight. The reason here is that the many electrons around the nucleus aren't really non-interacting as has been assumed up to this point, and there are other small effects one needs to take into account to get the energies exactly right (the momentum operator I've used is non-relativisitic, and there are interactions between the spins of the electrons and the spins of particles in the nucleus) So the higher angular momentum (l) states don't have quite the same energy as the lower ones. In fact, n=2, l=3 has a higher energy than n=3, l=2 so n=3, l=2 fills up first.
And thats my attempt at cramming the highlights of a one-semester QM course into a
The main reason is entropy.
There's a lot of information in genetic material - you'd need to wait for a fluctuation which just happens to produce that information. The kind of causality loop you're talking about would be much more likely to occur on atomic levels only (for instance, a photon forming a positron and electron, which then re-collide forming a photon again would qualify).
For it to occur macroscopically you'd have to have some way to engineer the loop (i.e. make it so that one particular set of fluctuations is the only one which can create a consistent loop), and you'd probably have to put enough energy into the thing to form everything that occurs in the loop (so a human-mass or two worth of energy).
Of course, if you're looking at it in the sense of photon->matter+antimatter->photon you don't really get what's traditionally thought of as timetravel - its not the kind of thing where you could say 'I want to go to the year 1532 AD'
It's hard to accelerate a person uniformly. Best way to do this would be to fall down a steep and very long gravity well. Of course, you also have to consider tidal forces there, but they can be reduced if you make the system even larger. However, a traditional means of acceleration (i.e. car ride) will accelerate with a very sharply falling off force, applied first to the layer of atoms at your back/feet/whatever part, which then must push the rest of your body. In extreme cases, the resulting wave of compression isn't very good for your internal organs.
That experiment is like a single organism floating around and waiting till its better, which isn't really what natural selection gives you. A better experiment would be to take 500 dice and 500 counters, roll the dice, update the counters, and now: throw away the lowest 100 counters and replace them with 100 counters taken from the group of 400 remaining. Now repeat. I guarantee you'll get to 20 faster than 222,155,644 generations.
Short chains of RNA can replicate themselves because each base tends to bond selectively with its conjugate base. It doesn't require any external machinery, just the presence of bases in solution. Of course, the process can take a LONG time (doubling time about a year). The paper I'm thinking of (Trinks, Schroeder, Biebricher "Ice and the Origin of Life" Kluwer Academic Publishers, 2004) studies this process occuring in arctic ice floes.
If you're willing to allow for the presence of a single enzyme, then you can get self-replicating DNA near hydrothermal vents. The hot spot sets up circulation, so the entire thing acts like a PCR chamber with a doubling time of about a minute (Braun, Goddard, Libchaber "Exponential DNA Amplification by Laminar Convection" Phys. Rev. Lett 91, 158103)
Well, since most people who review papers don't get paid, and editing is done by the person who submitted the paper in the first place... I really don't see much cost anywhere in there. Unless the cost of an email has suddenly become a few hundred bucks.
I believe that most of the cost is in fact related to publishing, since the cost of publishing a paper (to the person submitting it), depends strongly on the contents, i.e. how many figures, whether they're in color, etc. Whether a figure is in color or not can make a difference of hundreds of dollars. But that wouldn't affect anything other than printing/publishing costs.
Well, it doesn't matter how the numbers are generated as long as they end up having the exact same distribution. Chaotic processes, or an entity outside the universe playing bingo and calling out the numbers. There is no measurable difference between the two. So if thats free will, then a sufficiently good random number generator also has free will. There's nothing measurable that can draw a distinction between a perfect RNG and 'unpredictability due to free will'. This is a very general argument - it simply works on the separation of things into what must depend on an earlier state and what cannot depend on an earlier state, it has nothing to do with the structure of the brain or any specific cognitive system.
For any sufficiently long period of time after event, the outcome of a single change is likely to be random anyhow, so it probably wouldn't really matter. Of course, the necessary time may be very long for a well-chosen point of change. So I wouldn't think it's really that much of a risk to change the past, but its probably not all that useful unless you make very well-studied choices. Killing a single historical figure is not likely to create such long-term significant changes if the things that let that figure have the opportunity to become renowned are still present. For example, if you went back and killed Newton, its unlikely that no one would ever uncover the mysteries of 'Newtownian' mechanics, it'd probably just take longer. If you killed Hitler, it might very well be that you'd end up getting another war shortly after, maybe even a war with similar participants and similar atrocities.
But randomness itself has predictable properties. - mean, deviation, higher moments of the distribution of outcomes... If there's a random element to a person's actions, that by definition means that that element has no connection to the person's current state. It can't really be seen as 'will' because there's nothing behind it. It's 'predictably unpredictable'.
Trying to put detailed controls on the things is probably a losing proposition. Think of it this way - a bacterium with those controls is less well-adapted than one without it, so if you're trying to introduce that control as a protection against a hostile mutation, then its just as likely that a mutation removes the control as it is that it introduces some other negative effect.
As far as the dangers of genetic engineering, I'm not all that convinced there's much we can do to directly make something that accidentally becomes worse than the stuff already out there. Bacteria have short life cycles so they're one of the few things that can evolve about as fast as we can think up new stuff. So really we have more to fear in the sense of introducing new selection pressures (i.e. antibacterial soap) rather than having some lab-produced superbacterium (if its that good, it would've been done some time in the last billion years). After all, bacteria evolved us over that time span, and thats a far more complex change than tinkering with a few base pairs is going to cause.
So we might as well start and see what we can come up with. Intestinal bacteria do seem like a good place to start: since they already have a symbiotic relation with us, you'd expect it to be easy to introduce human-useful variations into the population without causing a major upset like trying to introduce a totally foreign species of bacteria.
I'm curious though: it takes a fair amount of time to actually wipe out say, 10gb of data. So I wonder if you rm -rf / and then immediately hit ctrl-c just how much you lose (or if you even really lose anything since rm probably just removes the file entry, as opposed to overwriting the data with zeros or noise...)
And how often have you actually done that by accident?
In this case its the author of the code who's doing it. I'd be worried however if someone were pressing a case for a GPL violation without the author's consent, since even if someone chooses to license under the GPL they're also free to license under alternate schemes to select other groups at their whim.
Well, thats the point of this paper. In a nutshell, what this paper is saying is that there's an inconsistency between general relativity and quantum mechanics unless you meet a certain condition, and black holes don't meet that condition.
...) into a single case, where the math just happens to be pretty much identical between all three views (its the lowest order theory for interacting 'stuff', so thats not too surprising).
Now, the only thing is I'm not quite sure where the condition he's specifying comes from; the author says 'what time do you mean when you write down the Schroedinger equation' but that argument seems insufficient to me, since you could say the exact same thing about 'what time do you mean when you write down classical hydrodynamic equations', with the answer being 'if you do everything in local variables, you get something consistent with GR'. So I can't immediately see why that doesn't apply to the Schroedinger equation as well since its just a wave equation, and those can certainly be done in GR.
The more telling example I think is the second one, which is an inconsistency with quantum measurements - the 'non-local correlations' he talks about, in which the inconsistency would be a consequence of quantum measurement theory (i.e. entanglement). I wish he gave a reference on that, or at least more detail, since I could easily imagine that that problem 'goes away' the same way it goes away if you try to use entanglement to send information faster than light. The result is that even though the entanglement effect is instantaneous, it can't transmit information. So I have to wonder if a similar thing would occur as to the problems introduced by not having an absolute time. However, it may very well be that it doesn't work out.
The thought experiment he talks about in this paper is rather interesting and there might very well be something there. At the least, its very 'cute' (in so far as such a thing can be considered cute) as it pulls in most areas of modern physics (astrophysics, condensed matter physics, particle physics
Well anyhow, thats just this physics grad student's 3am analysis.
This isn't necessarily a recent development though. I'm sure we had the same sort of paranoia during the cold war - the very essence of the cold war was paranoia about communism and communist world powers at some future point acting against the US.
Well we do have programs that can parse english sentences (and those from other languages) and create a probable tree structure (noun, verb, adj, etc). There are of course sentences which are meaningless without 'understanding' of context even though they're grammatically valid, and so on, but its a start.
So the next step would be, as they're trying to do, adding that context-sensitivity which is a form of understanding. But understanding is not the same as thinking: thinking would involve taking all the things which have been processed and using them to achieve a goal (or at least synthesizing the ideas into a new idea). It'd be interesting if that could be achieved 'simply' by applying transformations on the sentences designed to preserve logical relations, like the way a symbolic math program simplifies and solves equations (it just rearranges the symbols in logically equivalent ways).
Depending on what exactly you're trying to figure out, you might not even need to simulate multiple molecules. If you just want to know if two relatively small organic compounds will react to form other organic compounds, you could probably at this point in time do it with a quantum chemistry package like MPQC (among others). Ghemical (http://www.uku.fi/~thassine/ghemical/) will interface with those. But I should warn you, even calculating a 10-atom system with dynamics will take ages. But this is pretty much as 'complete' as you're going to get. If you could do it on a beaker, you'd get all the fluid dynamics, crystallization, etc without having to add in any extra rules.
If you know what reactions will occur, then you can model those with rate equations and go up to the fluids level (so no more dealing with individual molecules). This will be a lot faster, and you can get such interesting things as reactions that form spirals and other spatially-extended patterns.
What about a job in academia (tenure), government, ... thats about all I can think of but there's probably more examples where the bottom line isn't a guillotine.
Slightly offtopic, but I have a story from my experience of playing around with electrolysis and driveway cleaning compounds in my youth...
A solution of 6M HCl will slowly acquire copper ions when copper wires are used to electrolyse it. I have a suspicion this might be from chlorine formed at the anode reacting with the copper, but I don't know for sure. You get a green liquid, though I don't know whether its just Cu(1+) or if its some complex ion. Cu(2+) gives a blue liquid, I think. I may have those backward though, its been awhile.
Anyways, the resulting mix seems to eat through alumina with no problem, which was fun to demonstrate using aluminum foil to produce hydrogen gas. The 6M HCl alone won't get through the alumina. I don't think I ever found out exactly how that worked (you get what looks like copper metal in the resulting mush, but that could just be competition between CuCl and HCl in reacting with the aluminum metal beneath.) Any inorganic chemists here who want to hazard a guess?
What about using an amorphous form of tin rather than crystaline? I.e., quench the solder very quickly to room temperature rather than letting it cool slowly. If its due to either compressive stresses or oxygen, the problem is that when it expands, it expands preferentially along certain directions and at certain sites. If its amorphous, it shouldn't have any preferred directions, so instead of forming a whisker it just grows or shrinks a bit all over.
I imagine thats the sort of thing that adding a few impurities does, i.e. preventing a perfect crystal from forming. One question I have is what determines the radius of the whisker. If the entire sample is a perfect crystal, do you still get a localized whisker, or is it only if you have many grains of crystaline tin, each with a different orientation (so the whisker corresponds to a single grain).
Of course, if its some kind of diffusive dendritic growth thing, then that process will
form crystals regardless of whether the underlying structure is amorphous or perfectly crystaline (grains might matter). But coatings should probably prevent processes of that kind.
I'd disagree. If you know the bias, you can just dismiss everything that seems unpleasant to you by saying 'oh, thats just their X bias' whether or not it actually is part of their bias or not. If their bias takes form entirely in omission/inclusion selectivity (I'll assume for now they don't write anything thats actually false) then you'll stand a good chance of dismissing facts because of a disagreement of bias.
On the other hand, if the bias is totally random then you really do have to take each thing stated and ask yourself 'is this true? is this the _only_ thing I have to know to understand this situation? is this an emotional argument or a rational one?'. And you have to do it every paragraph. So as a result you end up thinking critically about the ideas, rather than about the author. And hopefully if the bias is random enough then you approach neutrality simply by averaging,
the same way that you (should) do to get a balanced view of a situation (i.e. consulting many different sources).
The scale you need to use varies depending on the outputs. What you're describing is basically gamma correction. For instance, my laptop display shows differences in dark colors very well, so red 0 and red 1 are easily distinguishable, but not 254 vs 255. However, my CRT is the opposite. You can easily distinguish 255 vs 254 on my CRT, but not 0 and 1 (or even 0 and 20 unless the lights are out and its night time and you sit about 4 inches from the screen).
One problem (and I don't know how digital cameras handle this) is that some times you want to accurately store intensities that go way beyond the scale (logarithmic intensities). With 8 bits you can only store 3 orders of magnitude change. But if you were to take a picture of a grassy field or something, the sky is (really off the top of my head here) maybe a thousand times brighter than the grass. So you end up with either normal sky and black grass or green grass and pure white sky. With a greater color depth you can correctly store these values and then use postprocessing to try and correct for the effect. But with 8 bits, it would simply get clipped at 255 or 0 and you'd lose that data.
The lack of free will does not necessarily mean that rights are meaningless. Instead of looking at rights as something fundamental, granted by the universe or god or whatever, instead you can look at rights as a means to stabilize social structures. That is, if I and those around me don't respect other people's right to live, then we will tend to destabilize the society around me, and it will either act to restabilize itself (i.e. kill/imprison those that kill) or will break apart (everyone starts killing eachother).
So based on that, the question of machine rights would simply be 'if we do give machines rights, how will it change their behavior compared to not giving machines rights?' In a way, we already do give machines rights, except that they're tied in with the rights of their owner. They have the right to not be broken into by a third party, they have a right to not be attacked via some denial-of-service scheme, etc. However, we see these simply as continuations of the rights of the owner. The way to introduce further machine rights then is, if an AI is written which its owner feels attached to, that owner will demand those rights for the AI: not to be deactivated without permission, the ability to hold financial resources (for when the user wants his or her AI to do some stock trading for them), the right to not incriminate onesself (this would be defense against law enforcement demanding crypto keys, seizing the machine, etc), and so on.
Of course, another way would be to develop the technology that lets a computer emulate the structures in a particular human's brain, so that the computer's output is indistinguishable from that person. Then simply have a computer emulate a dead person for 10 years or so, pretending to be that person living in seclusion, and when it comes to court there will be a good body of evidence that a computer and a person can be interchangeable.
Without power to gain, there is very little motivation for war. However, removing economic power still leaves other kinds of power to gain. If what you were saying were true, that is, that services were no longer necessary, then it _would_ remove most of the motivation for wars (why have slaves?).
However, services will still be necessary since being able to build something atom-by-atom doesn't mean that you can _design_ the thing atom-by-atom yourself. You still need to get someone to figure out the layout for you, you still need to get someone (or some machine at least) to get the raw materials to your doorstep, you still need people to transport you places (unless of course everyone learns to fly a plane - not an instantaneous thing).
The problem with being cautious here is, if we take things slowly, then people who stand to lose status by the introduction of new technology will stand in its way, legislate it to death, and it will end up being something that only a few people get the benefit from (or, a thing which everyone benefits from illegally). If we just jump into it then there's no time for that, and whatever we settle into is much closer to what you'd get if we had had that technology all along.