Actually quantum computing is, by design, probabilistic. Every specifically quantum algorithm (even Shor's infamous factoring algorithm) gives incorrect results by design for the simple reason that it's really not possible to have quantum algorithms which succeed all the time (unless you forgeo their quantum properties). So long as the probability of a correct answer is strictly greater than 0.5, however, one only has to repeat the computation a constant number of times to get the probability of success arbitrarily close to 1.
Nitpick: though this is true for the most familiar quantum algorithms, i.e. the ones in BQP (Bounded-error Quantum Polynomial-time), there are also some quantum algorithms that give exact answers with 100% certainty, such as EQP (Exact Quantum Polynomial-time).
Yes, and our current problem in physics is that we have two forks of Newtonian Physics. In one branch, we fixed the description of gravitation, in the other branch we fixed the description of subatomic particles. Both branches are very successful in their respective area. Now we try to merge those branches, however it turns out that the patches are not compatible, and we don't know what is the right way to combine them.
To be pedantic, both General Relativity and Quantum Mechanics are branches of Special Relativity, not of Newtonian Relativity. Like SR, QM also assumes that Maxwell's Equations are literally true, and thus c is constant in all QM reference frames (which implies the use of the Lorentz transformation and the absence of universal clocks).
As a concrete example, muons (a heavy counterpart to the common electron) are deeply in the realm of QM, yet we observe that the half-life of a muon is affected by the muon's energy. A muon experiences a constant half-life (about 2.2 microseconds) as all particles must in QM, but the more kinetic energy it has (i.e. the faster it goes) the more intensely it sees the universe around it as "running slow" due to time dilation. Meanwhile, we see time as running normal for us, but when we look at the muon, we see the muon as "running slow", such that higher-energy muons have longer half-lives. Lest anyone think this is academic, muon decays are quite easy to create in a lab, and nature also creates them quite commonly in the particle showers that spill out when cosmic rays hit the Earth's upper atmosphere. Not only are SR's predictions of muon behavior easy to lab-confirm, but we also have the striking fact that cosmic ray muons in a Newtonian-limit QM would be limited to traveling about 0.7km/0.4mi, never reaching our detectors on the Earth's surface, yet real-world muons routinely hit our detectors by traveling much deeper into the atmosphere than that.
That said, QM doesn't include GR's deformable spacetime fabric, and thus QM uses a Newtonian approximation for acceleration. Due to SR, this means that both ourselves and the muon are "right": according to SR/QM, both our clocks are running slow, and we have no GR to properly resolve the paradox and decide what happens if the muon decelerates to match our speed. As soon as you add a relativistic treatment of acceleration and deceleration to QM, you start getting the infinities spoken of in the article, and at a fundamental level those infinities mean that QM doesn't know what happens when a particle gains or loses energy. Instead, QM says that particles gain or lose energy "instantaneously", or rather "so fast that you can't tell what happened" as per the Heisenberg uncertainty of dE*dt. This means GR is, in some sense, more complete as a mathematical theory than QM is, since we can't look GR square in the eye and watch it squirm like QM does.
Sidebar: this is also why String Theory is still considered quite promising by many, despite looking like a quagmire to an outsider: all physical theories generated by mathematical String Theory automatically include GR and generalized QM as basic assumptions, with no problematic infinities. The sole difficulty in String Theory is "merely" finding an instance of that Theory that looks like our universe.
You clearly don't know your 18th century American or European history.
America built its textile industry, and indeed its prominence in the Industrial Revolution, only by flagrantly violating the patents of established European countries, especially those of Britain (and Scotland, the Silicon Valley of the day). In fact, the blatant colonial-day flouting of patent law and the use of "industrial espionage" (i.e. brain drain) against Britain was actually a significant pain point in the years leading into America's Revolutionary War, and tangentially figured into Britain's enthusiasm for the War of 1812.
(Copyrights were violated as well, but weren't considered as important at the time. Copyright itself was still crawling out of the cradle and unfamiliar to many — the first modern copyright law, the Statute of Anne, was passed only at the start of that century in 1709. At that point in time, copyright wasn't considered an economic engine, the way that patents were.)
You seem to know your physics. I have a question.
If a "stationary" black hole gets hit by an object of comparable mass, and neglecting the effects of gravity between both objects, will the black hole move at all? Will it only get as much kinetic energy as the mass it absorbs had, none at all, or you could actually hit it?
This is actually quite easy to answer, because momentum is conserved (in both Newtonian physics and General Relativity). In the Newtonian model, which is accurate for the masses and velocities we're dealing with, momentum equals mass times velocity. A stationary black hole has mass m_0 > 0 and velocity v_0 = 0, for a momentum of m_0*v_0 = 0. An incoming object with mass equal to the black hole has mass m_1 = m_0 and velocity v_1 > 0, for a momentum of m_1*v_1 > 0. The joint system, after the black hole has completely absorbed the incoming object, has momentum m_joint*v_joint = (m_0*v_0 + m_1*v_1) = (0 + m_1*v_1) = m_1*v_1 and mass m_joint = (m_0 + m_1) = 2*m_1. Therefore, its velocity is v_joint = (m_joint*v_joint / m_joint) = (m_1*v_1 / 2*m_1) = (v_1 / 2), or exactly half the velocity of the original incoming object, traveling in the same direction that the original incoming object was traveling. This is simply a fully inelastic collision — that is, a collision where the two colliding objects stick together instead of bouncing off each other. The fact that one object is a black hole is immaterial.
i just dont trust this latest batch that has and is being rushed to market over the H1N1, i will take my chances without this vaccine.
This "latest batch" has been almost six months in the making, which is roughly the same length of time it takes to develop the seasonal flu vaccine each year. Flu already requires a new vaccine every year: different strains require different vaccines, and every new year brings one or more new strains of flu, and this year's swine flu vaccine is no different from the vaccine for any other strain of flu in that respect. When a new strain arises, scientists don't have to throw away all their knowledge about making flu vaccines or treat it as a from-scratch research project. The biggest difference this year is that scientists found out about the novel H1N1 after they'd already started work on this year's seasonal flu vaccine, so it was too late to include it. This is why they're running behind on vaccine production.
Some background is required to appreciate this. Flu mutates fairly rapidly, since it's a single-stranded RNA virus. A "strain" is actually hundreds of individual mutant lines inheriting from a common ancestor but with significant differences between them. From the perspective of creating a flu vaccine, two proteins are critically important: hemagglutinin and neuraminidase, "H" and "N" respectively in "H1N1". Both of these proteins are directly involved in the cell-to-cell spreading of the virus — and thus highly conserved across mutant lines, and even across strains — and both of these proteins can be recognized by the immune system on the exterior of the virus. If you create a vaccine for one mutant line, then that single vaccine will be effective against most of the other mutants lines within the same strain. That said, there are other important proteins besides "H" and "N", and these differences mean that a vaccine for one H1N1 strain (e.g. A/Brisbane/59/2007, a target of this year's seasonal flu vaccine) will offer only a tiny amount of protection against any other H1N1 strain (e.g. the new swine flu strain).
Immunity to this year's strain will be widespread by next year, either by vaccination or by infection. This is unfortunate because some other strain (often one that already exists today) will fill the empty niche and become the "new" strain next year, and a new vaccine will be needed to protect against the "new" strain. Happily, scientists can look at this year's epidemiological data and make some pretty good guesses about which strains are likely to infect lots of people. They pick the three to five strains that look positioned to cause the most harm, and they start work on a combined vaccine against all of them for next year's flu season. This happens as the current flu season winds down in late Winter/early Spring, about six to nine months before the vaccine will be needed for the next flu season, because it will take that long to grow the flu viruses in chicken eggs, to kill the viruses, and to make doses of vaccine out of the viral remains.
Once next year's strains have been identified, all the hard research work has already been accomplished: growing the virus in chicken eggs rarely changes from year to year, killing the virus is easy and repeatable (and easy to verify before making a vaccine out of it), and making doses from the dead virus is usually trivial. When complications arise, it's almost always in the viral growth stage. If the strain is strongly bird-associated (e.g. the H5N1 "bird flu" strain from a few years back), the flu strain might kill chicken eggs too quickly. (Thankfully, viruses with this problem rarely spread well from human to human, since the virus must make a trade-off between infecting birds well or infecting humans well.) Alternately, the virus might not grow very well at all in chicken eggs, which is a pretty rare event since chicken eggs have very few defenses and even human-adapted strains usually retain the ability
Topologically, the inner surface of your intestines is external to your body (think "the inside of a torus"); also, like your skin, they only allow selective passage of molecules.
That's a very middle-school way of putting it, but basically accurate. You leave out the fact that the immune system regularly intrudes into the gut, and in fact communicates chemically with the commensal bacteria that live in your gut (because they share a joint interest in protecting the gut from hostile bacteria). From the immune system's perspective, the gut is a front line of defense, moreso even than the tonsils and adenoids, and it's critical to pay attention to what's happening inside it.
When you eat peanut butter, you don't end up with massive gobs of peanut butter floating around in your arteries, veins and capillaries.
That's because your digestive system has sorted the peanut butter into separate sugars, lipids, and proteins, not to mention broken them up into tiny globules to maximize surface area and thus permit absorption within a reasonable timeframe.
But the proteins, while not fully intact thanks to the action of pepsin in the stomach, are still in very large polypeptide pieces since pepsin can only slice at predetermined weak points in the peptide chain. The remaining polypeptides are still large enough to behave as antigens and provoke an immune system response.
(This is, in fact, a piece of the immune system in action. The stomach successfully digests most bacteria hapless enough to fall into it. The digestion renders the bacterium harmless but the antigens from the bacterium's surface reach the intestine intact. Once in the gut, the antigens are absorbed by the intestine and presented to the immune system, allowing the adaptive immune system to start planning an antibody-and-T-cell directed attack at the invading bacteria in the nose/sinus/throat, before they can even successfully breach the epithelium.)
(There's an even better piece of evidence: mad cow / Creutzfeldt-Jakob disease. The disease is caused by a prion, a misfolded protein that catalyzes its own misfolding from healthy protein. If proteins were obliterated by the digestive system, it would be impossible for mad cow / CJD to spread: the fact that the protein can spread from brain to food to intestine to brain and continue the infection in another individual means that at least some proteins can be absorbed whole by the intestines, even without being digested.)
Try again. The shrinks claimed that it was impossible that peptic ulcers were caused by a bacterial infection. They were wrong. Get over it.
When did I ever say that "shrinks" were right? My post was purely about biologists — a shrink talking about the immune system deserves as much respect as a chiropractor claiming subluxations cause cancer (i.e. none at all). You're projecting your own interpretation on my words here.
Also, you apparently don't understand how evolution works - "weaken the immune system -- presumably as some sort of evolutionary adaptation to divert energy resources toward getting away from the source of the stress" - you have to be joking.
No, I'm not joking. The human brain is a very expensive piece of hardware -- it consumes roughly 20%-30% of our daily caloric intake, depending on how much time is spent thinking heavily. The immune system is cheaper but nonetheless not trivial to fuel: mounting an immune response is expensive enough that mice shut down their reproductive organs in order to fuel it in the face of constant energy input during an acute infection. Energy is a budget — energy in necessarily equals energy out in a healthy organism — and the budget must balance or the organism dies. Given how much energy our evolutionary environment was willing to devote to our brains, I find it outright implausible that the body never diverts energy from other systems toward the brain, when it clearly does this between other organ systems.
Weaker immune systems get culled from the gene pool all the time.
Having a strong immune system is useless to, say, an animal with its foot trapped under a rock for three days, until it can finally work itself free. Energy must be either conserved or diverted toward systems that will help the animal escape, and in that situation the only two organ systems that can assist in that are the brain and the skeletal muscles. If the animal fails to divert energy to one of these two systems, the animal will die (and be culled from the gene pool). Adrenaline is well-known for being able to divert the body's resources toward the skeletal muscles to perform feats-of-strength, and cortisol is increasingly known to trigger energy-conserving behaviors, up to and including the shutdown of the immune system. Just because cortisol is maladaptive to humans living now doesn't mean it was always so — we didn't evolve to survive committee meetings and TPS reports.
Elizabeth Blackburn et al. UCSF has shown work that reveals that mothers caring for their very sick children have shorter telomeres when they report that their emotional stress is at the greatest point. She also found telomerase active at the site of blockages in coronary artery tissue. This could be why heart attacks can come on so suddenly: Telomerase is driving the growth of the blockage.
Other work has shown that the poor of society have shorter telomeres than the rich.[14] Short telomeres can lead to telomeric crisis and the initiation of cancer if many other conditions are also met, or so the discussion goes at this point.[citation needed]
Blackburn and the two other co-discoverers of telomerase won the Lasker Award (2006), and the Nobel Prize (2009) for the discovery of telomerase and subsequent work on telomerase. Blackburn also won the 2006 Gruber Genetics Prize for same.
Seriously, wake up and smell the post-1995 research. You're living 15 years behind the state-of-the-art.
... and appears to be invented from thin air, same as the "your ulcers are caused by stress" bullshit that was finally disproved a couple of decades ago.
Uh, "ulcers are caused by stress" hasn't been disproven... in fact, it's been confirmed in a roundabout way. Stress is now clinically demonstrated to weaken the immune system — presumably as some sort of evolutionary adaptation to divert energy resources toward getting away from the source of the stress — and that leaves you more prone to bacterial infections, such as H. pylori in your stomach/duodenum.
You might like peanut butter, or bacon and eggs,or ketchup, but injecting any of them directly into your bloodstream isn't the same as eating them.
Yes it is. What do you think your intestines do? (Pepsin doesn't digest proteins all the way into their constituent amino acids, after all. And your intestinal cells are positively bathed in whatever peptides you just ate, and they're just as easily inflamed as any other epithelial tissue in the body.)
Richard Lenski could have saved himself a lot of time if he had asked himself "was any new information created when it mutated" . The answer of course is NO!
Even though you're a troll, I'm feeling generous today. This is completely and utterly wrong, and if you understood what information was, you'd agree with me — and all of biology — that evolution occurs within a species. (Speciation will have to wait for another day.)
Claude Shannon, of Bell Labs fame, invented Information Theory in the late 1940s for the utterly practical purpose of cramming more data onto copper wires. What he discovered, with a bit of a shock quite soon after, was that the equations were identical to those describing thermodynamic entropy. In fact, thermodynamic entropy turned out to be a special case of Information Theory. After discovering this, Shannon took to calling his discovery "information entropy".
Fundamentally, thermodynamic entropy is the unpredictableness of a physical system. The more unpredictable a physical system is, the more information it takes to describe the system. This was the link between the two.
About 10 years later in computer science, two researchers named Kolmogorov and Chaitin independently invented a hypothetical measure for the complexity of any arbitrary data: measure the length of the shortest possible computer program that can produce that data. Again, random data has the highest complexity: if the data has a pattern, then a short program can compute the pattern starting from a tiny piece of the data; but if there is no pattern in the data, the program must be large enough to duplicate a full copy of the data.
Getting back to biology: mutations add randomness to DNA. Therefore, they make the DNA less predictable, and therefore they add information and complexity to the DNA. After that, natural selection acts on that mutation: if the mutation was harmful for the cell, the cell makes fewer copies of itself; if the mutation was beneficial for the cell, the cell makes more copies of itself.
(Aside: It helps that DNA duplication is a fairly common event, especially in kingdoms like the animal kingdom where virus-like transposons infect all of our genomes. If you're a cell, and you have two copies of a gene, and one copy is mutated into something useless by a mutation, then nothing bad happens to you. In fact, if having the extra copy was a bad thing, making the copy shut up or do something else is a good thing. A lot of new proteins arose because the gene coding for them was copied then modified until it finally did something useful again, like the mammal blood clotting cascade or the photoreceptor pigments for color vision.)
Natural selection provides a filter: it layers meaning on top of the information in the DNA, in much the same way that "English" is a filter that layers meaning on top of "light-emitting screen that displays funny squiggly marks". If information "A" means "cell lives", and information "B" means "cell dies", then natural selection is the process that distinguishes between "A" and "B" by giving them meaning. Information is complexity. Information is unpredictability. Information is randomness. Information is not meaning. Meaning is something you do with information, not something the information inherently has.
Once you understand the difference between "information" and "meaning", you necessarily realize that DNA was the final, unequivocal proof that microevolution logically must exist in biology and that it's silly to argue otherwise. Your beliefs are contradicted by reality itself.
No, modern agriculture provides food as a function of how much energy and petrochemicals are put into each unit area of land. The oil price spike caused a global food crisis because not only did it increase the costs of fuel (for farm operation as well as transport) it also increased the cost of the oil derived chemical feedstocks used in fertilizer.
If you're going to repeat this myth, please take the time to come up with at least one fertilizer component made from a petroleum-derived feedstock before repeating it.
Mineral fertilizers contain nitrogen, phosphorous, and potassium (the "NPK" trio), sometimes plus other trace minerals like calcium. Oil contains none of these, so oil is worthless as a fertilizer feedstock. Only for nitrogen does this come even within spitting distance of the truth: the Haber process consumes hydrogen and atmospheric nitrogen to create ammonia, and the most common source for the hydrogen is natural gas — NOT petroleum, which is extremely carbon-rich and hydrogen-poor.
100% of petroleum in farming goes to the use of diesel: operating diesel-powered farming equipment and hauling the harvest in diesel-powered trucks. From a slightly broader perspective of all fossil fuels used in farming, the Haber process barely scratches the surface, as it's tiny compared to the need for diesel. And from an even broader net-carbon perspective, even the diesel use is massively outstripped by the conversion of natural carbon sinks (e.g. old-growth forests, peat bogs, wetlands) into farmland, which can hold far less carbon than the systems it replaced. The Haber process is too tiny to pick on, and calling fertilizers "petrochemicals" just serves as a factually incorrect "buzzword bingo" game to distract people from the real issues.
Essentially, *if* we knew the laws of physics in sufficient detail to build an accurate simulation — i.e. had a working quantum gravity theory — *and* we could probe all the entropy of the black hole to insert into the simulation — an feat that makes quantum gravity look like high school algebra — *then* we could build the simulation and run it backward (yay time-reversible physics) to figure out how the black hole was created, including the fact that one of the formative objects was indeed a shoe. Likewise, if we were to study *all* the outflowing Hawking radiation, we could model the emission of the radiation, and from that a model of the black hole that emitted it, and from that a model of the objects that formed it, shoe included.
This is of course a "spherical cow" thought experiment because of the immense amounts of entropy involved in a star-sized black hole: just the RAM in the simulating computer would necessarily occupy a volume greater than star-sized. But for a sufficiently small and short-lived black hole — for instance, one formed by colliding a proton and antiproton with the correct kinetic energy — the numbers suddenly become much more reasonable. It's conceivable that we could perform the experiment on a microscopic black hole and arbitrarily approach that limit of understanding in our models, even if we could never truly reach it.
(At this point, the idea reminds me a bit of X-ray crystallography: in crystallography, the image generated is a fuzzy, imperfect view of the spatial Fourier transform of a physical crystalline structure, with sharp lines indicating the distance between repeating units in the lattice (i.e. the spatial frequency). I imagine that interpreting the Hawking radiation from a microscopic lab-created black hole would be a roughly similar experience — with the exception that humans can, with practice, glean meaning from a spatial Fourier transform just by looking at it, whereas I doubt a similar thing would be possible with a pictoral representation of entropy in Hawking radiation.)
That said, even if we never did achieve such an experiment with microscopic black holes, the entropy continues to be information even when we're not looking at it, or even when we have no ability to derive meaning from it — which was my original point in my first reply. Though I've definitely found this to be a useful and thought-provoking discussion.
With all due respect, I still think the mistake is on your end, in that you seem to equate entropy, information, and informational entropy.
Entropy refers to the logarithm of the number of possible states a system could be in, based on what you can observe about it. (And of course to fully generalize it, you actually sum over the probability times log of the inverse of probability, which is Shannon entropy, to account for states not being equiprobable.)
Correct.
Information, however, is the opposite of entropy: the more you know about a system (i.e. the more information you have about it), the less entropy it has (because you've reduced the number of states it could possibly be in -- that's the point Less Wrong makes about how knowing more about particles makes them colder.)
Correct, with the caveat that you're only hinting at the underlying idea. Entropy is measured relative to a frame of reference (a closed system), just as position and velocity are measured relative to a frame of reference (an inertial/non-accelerating frame in Newtonian physics and SR).
The Maxwell's Demon refutation shows that *entropy* cannot be destroyed (summed over all the universe), including informational entropy. But *information* can be destroyed -- by its conversion into entropy, or increase in the number of possible states it can be in. (In the example, after using the knowledge the demon had, the mutual *information* between the demon and the system is indeed destroyed.) What physicists warn cannot happen is destruction of entropy: collapsing mutliple states into one state, which would violate Liouville's Theorem.
Wrong emphasis: *mutual* information, not mutual *information*. "Mutual information" is a special case of the broader category of "information". Some instances of *mutual* information even belong to the restricted and rarefied subcategory of *meaningful* mutual information, i.e. mutual information that human mental algorithms are capable of making direct use of. But, again, this has nothing to do with the Shannon definition of information, which makes no claims as to whether or not the information is meaningful [1].
Why am I emphasizing the word "mutual" in all of this? Because the mutual information is a property of the closed system itself, not of either component! If two components of a closed system are correlated, it doesn't mean the equality of entropy and information is broken, it just means that they cannot be fully examined in isolation from each other.
If you were to examine each component in isolation, as if each one existed inside a closed system that excluded the other, you would dutifully arrive at an entropy figure for each one. If you then incorrectly summed the apparent entropy of these two components, as if the fictitious isolation actually existed, the sum would double-count that portion of the entropy which is correlated between the two parts. The entropy figures you calculated were only apparent entropy, not true entropy. Instead, you have to subtract out the degree to which their entropies are correlated — that is, their "negentropy" or "mutual information" — to arrive at the correct figure for the closed system as a whole. But the negentropy does not physically exist: it is a mathematical artifact of your attempt to consider each component in isolation, just as Newtonian centrifugal force is an artifact of a non-inertial reference frame. The negentropy "exists" only because it cancels out your first mathematical artifact: when you tried to consider the parts independently of their correlation, you invented "apparent entropy" figures that had nothing to do with the "true entropy" figure (relative to the larger system in which the two components are correlated).
It's enlightening to think about entropy as being the physical representation of "history". If an irreversible action happens, entropy is the historical record that we exist in
I don't know what the Maxwell's Demon refutation has to do with what I posted. I agree with all of that, and am familiar with it. In fact, I've helped to promote that very link you gave about Engines of Cognition.
My apologies, but have you followed through on the consequences? Information entropy and thermodynamic entropy are the same thing -- rather, the latter is a special case of the former -- and if black holes could permanently destroy information, then one could build a Maxwell's Demon by using a black hole to destroy the thermodynamic entropy gained in learning the particle velocities. Because Maxwell's Demon is impossible, black holes must preserve information. QED.
But anyway, I agree that perhaps I went too far in saying that no information escapes. Rather, the minimum possible information escapes. You can infer the total mass-energy that went into a black hole based on what comes out, but that's it. A black hole still removes all constraints on the degrees of freedom of a system that keep it from having the entropy that a black hole would. The degrees of freedom are constrained when e.g. the object has shape, a specific pattern of motion, remains a solid rather than a liquid, etc. We recognize such constraints as information.
Again, you're missing my point. All the information entropy is preserved, and every bit of it is present in the resulting Hawking Radiation: for each bit of information entropy that you observed falling into the black hole, there is one bit of information entropy trapped in the event horizon that you can predict better than chance. The information has been smeared and jumbled with all the other bits, but it is still in there, and it still correlates with the circumstances that led to the black hole's formation.
A human confusion is the idea that "information" has anything to do with "meaning". One bit of information allows you to correctly predict the answer to one yes/no question with equal odds. Nothing more. It requires more information to describe hot water than it does to describe cold water, but that doesn't mean the hot water is more meaningful. Bulk properties like shape, state of matter, etc. are very tiny bits of information, vastly outnumbered by the bits of information that exist closer to the atomic level. That means, if you smear and jumble the bits together, it's hard to find the ones that a human is interested in. That doesn't mean the bits ceased to exist -- any of them.
And, yes, the dust is still settling on this interpretation of black holes, but it's increasingly the dominant position among physicists. This is why I mentioned Stephen Hawking: as the original proponent of the "black holes destroy information" hypothesis, the fact that he was convinced otherwise by the evidence and recanted his old hypothesis in 2004 is extremely indicative of the mood of the field. There are still some unresolved questions, yes, but it's not obvious how to proceed: first we need to discover a coherent description of what "entropy" means (either thermodynamic or informational) in a Quantum Mechanical context, because classical entropy doesn't have a sane interpretation in QM.
Exactly. Another way to think about it is this: the speed of light is the maximum speed information can travel. Since the escape velocity of a black hole is greater than the speed of light, information cannot escape either.
Because information cannot escape, it means you cannot infer what went into a black hole by looking at what comes out. (Note: information being "destroyed" in a black hole doesn't prevent you from e.g. keeping a record of what you tossed into a black hole. Once it goes inside, it won't somehow destroy your records, as I used to believe was the implication.)
Interestingly, this is wrong, and Stephen Hawking lost a bet over it.
The Second Law is actually very tightly coupled to Shannon-style Information Entropy: if you knew enough information (as a fait accompli) about the current state of a system at equilibrium, you could successfully build a Maxwell's Demon that used that information to separate the system into hot and cold reservoirs, allowing you to cancel out the entropy of the system with your information's "negentropy" (as it's called). Learning the information in the first place causes your information entropy (i.e. correlation with the system, negentropy) to increase, which by physical necessity also causes your thermodynamic entropy to increase in tandem. This is why a Maxwell's Demon doesn't work: the entropy undone in using the information is always less than (or, in a perfect system, equal to) the entropy done while learning it. (If blind faith provided non-tautological and accurate information about the universe with better-than-random chance, then you could build a Maxwell's Demon that broke the Second Law -- and since the Second Law is inviolable, it must be the case that blind faith tells you nothing... except possibly tautologies if your brain uses reversible computing. If you think hard enough about it, it also disproves substance dualism.)
In the specific case of information entropy and black holes, it turns out that the information never crosses the event horizon, and thus never has to break the speed of light limits when leaving it. As modern physics and Stephen Hawking have both discovered, all the entropy of a black hole's formative mass/energy is encoded in two dimensions as ripples in the event horizon of the black hole. Black holes have also been discovered to be maximum-entropy objects in modern physics, containing the largest amount of entropy physically possible for the volume of space enclosed by the event horizon. (This has interesting implications on the nature of reality -- look up the anti-de Sitter/CFT correspondence for all sorts of 2D/3D weirdness, like the universe being equivalent to a 2D hologram.) When a quantum of Hawking Radiation emerges from the event horizon, it carries off precisely the amount of entropy equal to the entropy carried by the change in the surface area of the event horizon when expelling the quantum, thus maintaining the invariant that the black hole is a maximum-entropy object. Because "information entropy" is another way of saying "too random to predict ahead of time" -- that's what information is, by Shannon's definition of it -- the radiation looks quite random indeed. But that doesn't mean it's uncorrelated with the history of the black hole.
... Plus, I honestly don't understand why receiving calls is considered so dangerous while on the road. Text messaging is a different beast entirely, one that requires your full attention, especially when they are being sent. But unless you are one of those people who can't drive and listen to a conversation in the car or the sound system, then I don't see why receiving a phone call is any more dangerous....
The problem, as you grasped with texting, is attention. Driving is filled with critical moments: brief events that are capable of leading to an accident if not dealt with swiftly. They happen a lot more than we realize in retrospect: if they are dealt with well, we don't associate them with strong emotion and thus we forget them; and because we estimate the probability of an event based on how many examples of it we can call to mind, we don't realize how frequently they happen.
Having music playing in the background requires no attention, allowing you to drop attention entirely in order to focus on a critical moment. Having a conversation with a person in the car is slightly more dangerous, because your attention is more likely to be diverted when a critical moment is approaching, but once the critical moment is noticed, the conversation will gracefully pause as you deal with it, because your passenger can see your body cues as well as the road ahead of you. Having a conversation with a person on the phone, in contrast, is orders of magnitude more dangerous: the person on the other end of the phone will converse as if he or she has your undivided attention, because neither your body language nor the road is visible to him/her.
This is greatly compounded by our social expectations surrounding telephone use, which arose in the context of landlines: it's considered extremely rude to stop paying attention to the person on the other end of the line. (This is why text messaging is taking over: it's more polite, because there's no expectation of undivided attention.) Accordingly, our brains innately rate this social risk of rudeness as justifying a higher level of attention than a conversation with a passenger in your car. Your brain thinks of being rude as a danger, just as it thinks of getting into a car accident as a danger, but doesn't recognize that the difference is several orders of magnitude — it severely discounts the risk of getting into an accident, because the brain isn't innately good at understanding low-probability events (because it has a hard time calling up examples). Therefore, the innate attention systems devote undue attention to the phone conversation, increasing the risk of an accident: both from failures to recognize critical moments early enough (a risk that was present with a passenger, but now greatly amplified), and from failures to devote enough attention on the critical moments before they evolve into near-accidents or accidents (a new risk entirely, because the brain still believes the accident is unlikely).
AFAIK, a classic electromechanical meter, which is overwhelmingly typical in the United States, requires 100% of the load to flow through an inductive coil. Since an inductive coil acts as a low-pass filter, radio frequencies are stopped at the meter itself. More modern solid-state meters are more likely to use a current transformer, which does not require the load to pass through a particularly strong inductance.
... For addiction science to progress AA has got to be publicly debunked as a religious cult-like group that stagnates science...
Ohfuckyes. AA and similar organizations have inflicted an amazing amount of damage through their cultish, quasi-religious tactics. AA's tactics almost always have 1:1 parallels with the evangelical Christian groups that they spawned from.
A prominent phenomenon in evangelical Christianity is the testimony, where the person in question describes (in gruesome detail) every wrong they did (or wish they did, or imagine themselves as wishing they did) prior to their most recent Born Again experience, baptism, or whatever ritual happens to be their flavor-of-the-month. Inevitably, they fall off the wagon in secret, get caught in public, beg for forgiveness, then start the cycle over again with another testimonial (possibly at a new church). The sin/confession cycle is deeply psychologically addictive because of the euphoric rush provided during the testimony phase. Combine that with a taboo against judging the current actions of current group members and a stubborn insistence that people are not responsible for their own actions, and they create the very helplessness that they preach, trapping people in the religion and keeping them from developing emotional stability in their lives as they lurch from one destructive behavior to the next, each time convinced that the church/God/Jesus has once again saved them from destruction.
Modulo a few changed words, this whole cycle plays itself out almost identically with AA and similar groups, except with the added disturbing component that their victims are people who have already proven themselves to be drawn to addictive patterns, making them perfect prey for the groups' emotional predation. Given that nine-tenths of any drug addiction is psychological, the actual chemical addictiveness of a given drug is almost irrelevant to the equation.
The SRY gene is responsible for only a small proportion of direct masculinization: it causes the gonads to develop as testicles, it causes production of anti-Müllerian hormone, it causes production of androgens (i.e. testosterone), and it causes some tiny changes in the brain. All the remaining primary and secondary characteristics are the result of either AMH or testosterone.
Supposing a 46-XY subject with CAIS (Complete Androgen Insensitivity Syndrome), note that AMH is still effective at de-feminizing the internal genitalia by triggering the reabsorption of the Müllerian ducts: no fallopian tubes, uterus, cervix, or upper vagina. However, because the androgen signal cannot be received, Wolffian ducts do not form: no efferent ducts, epididymis, vas deferens, or seminal vesicle. Also, the brain is not fully masculinized, as the bulk of brain masculinization is androgen-induced, not SRY-induced.
The final CAIS result is external female genitalia, a shallow vagina, active testes producing male gametes, male-normal androgen levels, and no internal genitalia (male or female) beyond the testes and vagina. The testes may remain undescended, or may descend into the labia majora (homologous to the scrotum) where they are easily palpable. Gender identity is most frequently female, due to largely absent masculinization of the brain. Secondary sexual characteristics are frequently hyper-feminine: in contrast to a CAIS individual, a typical woman express some androgen-induced characteristics (albeit vastly reduced from the same in a typical man). In particular, CAIS individuals frequently have little to no body/pubic hair, and highly feminine facial bone structure. For this reason, there is a disproportionate positive correlation between CAIS and supermodels.
PAIS (Partial Androgen Insensitivity Syndrome) is much less sharply defined, as it runs the gamut from nearly-CAIS-woman to man-with-micropenis, including everything between the two (depending on how badly impaired the androgen receptors are).
Note that the athlete in the article clearly does not have androgen insensitivity, either complete or partial, so this line of discussion is irrelevant to that topic. My best guess (as a well-educated layperson) is 46-XX, absence of SRY, with abnormally high testosterone production. The facial bone structure suggests high androgen exposure during puberty, which for reasons of exposure consistency would tend to rule out steroids or an androgenic endocrine tumor, and the female gender identity and (presumed) presence of external female genitalia rules out fetal androgen exposure from testicular tissue (e.g. chimeric absorption of a 46-XY twin brother). Perhaps some genetic disruption of the regulatory regions involved in testosterone production (more likely), or an abnormally strong response to female-typical androgen levels (less likely). My conclusion: in essence, an unusually butch woman.
A predictable response, but consider that if a person is clinically addicted to certain drugs, then not having those drugs regularly is highly likely to cause death. In that sense, there is some truth to the reasoning that "the drugs made me do it" - the drugs are necessary to avoid death, and if the person were not addicted, then the drugs would not be necessary.
Happily, even among highly chemically-addictive drugs, there are very few that can kill you if you discontinue them abruptly. Sadly, the worst of the lot (alcohol) is the most widely abused. Because alcohol is a powerful GABA agonist (i.e. amplifies the effects of GABA, an inhibitory neurotransmitter), the body responds to long-term alcohol use by decreasing sensitivity to GABA. During alcohol withdrawal, the reduced sensitivity means the body's natural GABA signals are no longer effective, causing glutamate and other excitatory transmitters to operate unchecked. Untreated withdrawal in a heavy alcoholic causes elevated heart rate, tremors, convulsions, permanent brain damage, then death.
Other highly addictive drugs with absolutely dreadful withdrawals (e.g. heroin) are rarely even remotely life threatening (however miserable the person feels while it's happening).
The concept underlying "gaydar" applies much more broadly than to just gay people. Humans naturally broadcast extremely subtle, ambiguous social signals indicating their membership in whichever subculture(s) they belong to, often through the use of word choice or references to common subcultural touchstones. Because each signal is ambiguous, they individually mean nothing and are normally tuned out as noise by non-members. However, a person who shares membership in one or more subcultures will spot the ambiguous signals and wonder, "Huh, I wonder if X is also a member of Y group". The received signal will prime them to look for more signals of the same kind, and cause them to semi-subconsciously broadcast return signals. If they see more signals in response, a feedback loop forms as they become increasingly certain, and eventually there's an unspoken knowledge by both parties that each knows the other knows that both are members of the subculture. At that point, they start a conversation.
This is an extremely broad human behavior, and applies to almost all subcultures, no matter how trivial: from ones as secretive as illegal drug users and 19th-century gay men, to ones as openly-declarable as churchgoers and swing music enthusiasts. The point at which a conversation is held depends on the risk and consequences of reading the signals incorrectly: members of more mainstream, well-known subcultures will start a conversation with each other more quickly, because the odds are good and the consequences are few. But all subcultures do it; even an enthusiast of something completely non-offensive, like crochet, won't start a conversation with a random acquaintance about crochet without first seeing a hint or two that the acquaintance actually has a crochet interest. The minimum consequence of being wrong is wasting the other person's time, which is rude and thus a social negative.
Seriously folks, is there no already existing file system that can already meet these needs? If not, then what are Google's competitors using?
Is that no one else has yet to face up to this issues properly and this is a huge competitive advantage for Google, or is it simply NIH?
HDFS (the Hadoop File System) is based on fairly similar concepts, AFAIK. I haven't seen anything that directly competes with Bigtable, but the article goes into detail about SSTables and compaction, so I don't see any barrier to writing an open source competitor to that, either.
(I've heard substantial rumors that there are already closed-source clones of both GFS and Bigtable, in addition to the significant and often well-publicized internal adoptions of Hadoop, e.g. at Yahoo. The GFS and Bigtable white papers are rather surprisingly complete.)
It's not a debate. Doing this turns those power lines into big antennas. You can't debate the laws of physics.
Note that the article specifically mentions that it's not about broadband-over-power-lines (BPL). The author is restricting himself to discussing home replacements for Ethernet, which are blocked at the meter and cannot leave the house. Yes, the wires inside the house become antennae, but it doesn't cause the widespread disruption that BPL would.
You do realize that this was in reference to Ogg Theora, right, and not Ogg Vorbis?
Theora the video codec is formerly known as VP3, and it was commercially sold by On2 Technologies before On2 open-sourced it. On2 owned (and owns) the patents on VP3/Theora, and they were confident enough that there were no third-party patents pertaining to it that they felt safe selling it commercially. If On2 says they've granted everyone a perpetual license to use the VP3 patents, then I trust On2's judgement that everyone has permission to use VP3, especially since commercial infringement opens you up to much greater patent liability than open source infringement. In a world where MJPEG probably has two or three submarine patents still floating around, it's not worth worrying about VP3/Theora.
Slashdot Mirror
Actually quantum computing is, by design, probabilistic. Every specifically quantum algorithm (even Shor's infamous factoring algorithm) gives incorrect results by design for the simple reason that it's really not possible to have quantum algorithms which succeed all the time (unless you forgeo their quantum properties). So long as the probability of a correct answer is strictly greater than 0.5, however, one only has to repeat the computation a constant number of times to get the probability of success arbitrarily close to 1.
Nitpick: though this is true for the most familiar quantum algorithms, i.e. the ones in BQP (Bounded-error Quantum Polynomial-time), there are also some quantum algorithms that give exact answers with 100% certainty, such as EQP (Exact Quantum Polynomial-time).
Yes, and our current problem in physics is that we have two forks of Newtonian Physics. In one branch, we fixed the description of gravitation, in the other branch we fixed the description of subatomic particles. Both branches are very successful in their respective area. Now we try to merge those branches, however it turns out that the patches are not compatible, and we don't know what is the right way to combine them.
To be pedantic, both General Relativity and Quantum Mechanics are branches of Special Relativity, not of Newtonian Relativity. Like SR, QM also assumes that Maxwell's Equations are literally true, and thus c is constant in all QM reference frames (which implies the use of the Lorentz transformation and the absence of universal clocks).
As a concrete example, muons (a heavy counterpart to the common electron) are deeply in the realm of QM, yet we observe that the half-life of a muon is affected by the muon's energy. A muon experiences a constant half-life (about 2.2 microseconds) as all particles must in QM, but the more kinetic energy it has (i.e. the faster it goes) the more intensely it sees the universe around it as "running slow" due to time dilation. Meanwhile, we see time as running normal for us, but when we look at the muon, we see the muon as "running slow", such that higher-energy muons have longer half-lives. Lest anyone think this is academic, muon decays are quite easy to create in a lab, and nature also creates them quite commonly in the particle showers that spill out when cosmic rays hit the Earth's upper atmosphere. Not only are SR's predictions of muon behavior easy to lab-confirm, but we also have the striking fact that cosmic ray muons in a Newtonian-limit QM would be limited to traveling about 0.7km/0.4mi, never reaching our detectors on the Earth's surface, yet real-world muons routinely hit our detectors by traveling much deeper into the atmosphere than that.
That said, QM doesn't include GR's deformable spacetime fabric, and thus QM uses a Newtonian approximation for acceleration. Due to SR, this means that both ourselves and the muon are "right": according to SR/QM, both our clocks are running slow, and we have no GR to properly resolve the paradox and decide what happens if the muon decelerates to match our speed. As soon as you add a relativistic treatment of acceleration and deceleration to QM, you start getting the infinities spoken of in the article, and at a fundamental level those infinities mean that QM doesn't know what happens when a particle gains or loses energy. Instead, QM says that particles gain or lose energy "instantaneously", or rather "so fast that you can't tell what happened" as per the Heisenberg uncertainty of dE*dt. This means GR is, in some sense, more complete as a mathematical theory than QM is, since we can't look GR square in the eye and watch it squirm like QM does.
Sidebar: this is also why String Theory is still considered quite promising by many, despite looking like a quagmire to an outsider: all physical theories generated by mathematical String Theory automatically include GR and generalized QM as basic assumptions, with no problematic infinities. The sole difficulty in String Theory is "merely" finding an instance of that Theory that looks like our universe.
You clearly don't know your 18th century American or European history.
America built its textile industry, and indeed its prominence in the Industrial Revolution, only by flagrantly violating the patents of established European countries, especially those of Britain (and Scotland, the Silicon Valley of the day). In fact, the blatant colonial-day flouting of patent law and the use of "industrial espionage" (i.e. brain drain) against Britain was actually a significant pain point in the years leading into America's Revolutionary War, and tangentially figured into Britain's enthusiasm for the War of 1812.
(Copyrights were violated as well, but weren't considered as important at the time. Copyright itself was still crawling out of the cradle and unfamiliar to many — the first modern copyright law, the Statute of Anne, was passed only at the start of that century in 1709. At that point in time, copyright wasn't considered an economic engine, the way that patents were.)
You seem to know your physics. I have a question. If a "stationary" black hole gets hit by an object of comparable mass, and neglecting the effects of gravity between both objects, will the black hole move at all? Will it only get as much kinetic energy as the mass it absorbs had, none at all, or you could actually hit it?
This is actually quite easy to answer, because momentum is conserved (in both Newtonian physics and General Relativity). In the Newtonian model, which is accurate for the masses and velocities we're dealing with, momentum equals mass times velocity. A stationary black hole has mass m_0 > 0 and velocity v_0 = 0, for a momentum of m_0*v_0 = 0. An incoming object with mass equal to the black hole has mass m_1 = m_0 and velocity v_1 > 0, for a momentum of m_1*v_1 > 0. The joint system, after the black hole has completely absorbed the incoming object, has momentum m_joint*v_joint = (m_0*v_0 + m_1*v_1) = (0 + m_1*v_1) = m_1*v_1 and mass m_joint = (m_0 + m_1) = 2*m_1. Therefore, its velocity is v_joint = (m_joint*v_joint / m_joint) = (m_1*v_1 / 2*m_1) = (v_1 / 2), or exactly half the velocity of the original incoming object, traveling in the same direction that the original incoming object was traveling. This is simply a fully inelastic collision — that is, a collision where the two colliding objects stick together instead of bouncing off each other. The fact that one object is a black hole is immaterial.
i am aware of how vaccines work
i just dont trust this latest batch that has and is being rushed to market over the H1N1, i will take my chances without this vaccine.
This "latest batch" has been almost six months in the making, which is roughly the same length of time it takes to develop the seasonal flu vaccine each year. Flu already requires a new vaccine every year: different strains require different vaccines, and every new year brings one or more new strains of flu, and this year's swine flu vaccine is no different from the vaccine for any other strain of flu in that respect. When a new strain arises, scientists don't have to throw away all their knowledge about making flu vaccines or treat it as a from-scratch research project. The biggest difference this year is that scientists found out about the novel H1N1 after they'd already started work on this year's seasonal flu vaccine, so it was too late to include it. This is why they're running behind on vaccine production.
Some background is required to appreciate this. Flu mutates fairly rapidly, since it's a single-stranded RNA virus. A "strain" is actually hundreds of individual mutant lines inheriting from a common ancestor but with significant differences between them. From the perspective of creating a flu vaccine, two proteins are critically important: hemagglutinin and neuraminidase, "H" and "N" respectively in "H1N1". Both of these proteins are directly involved in the cell-to-cell spreading of the virus — and thus highly conserved across mutant lines, and even across strains — and both of these proteins can be recognized by the immune system on the exterior of the virus. If you create a vaccine for one mutant line, then that single vaccine will be effective against most of the other mutants lines within the same strain. That said, there are other important proteins besides "H" and "N", and these differences mean that a vaccine for one H1N1 strain (e.g. A/Brisbane/59/2007, a target of this year's seasonal flu vaccine) will offer only a tiny amount of protection against any other H1N1 strain (e.g. the new swine flu strain).
Immunity to this year's strain will be widespread by next year, either by vaccination or by infection. This is unfortunate because some other strain (often one that already exists today) will fill the empty niche and become the "new" strain next year, and a new vaccine will be needed to protect against the "new" strain. Happily, scientists can look at this year's epidemiological data and make some pretty good guesses about which strains are likely to infect lots of people. They pick the three to five strains that look positioned to cause the most harm, and they start work on a combined vaccine against all of them for next year's flu season. This happens as the current flu season winds down in late Winter/early Spring, about six to nine months before the vaccine will be needed for the next flu season, because it will take that long to grow the flu viruses in chicken eggs, to kill the viruses, and to make doses of vaccine out of the viral remains.
Once next year's strains have been identified, all the hard research work has already been accomplished: growing the virus in chicken eggs rarely changes from year to year, killing the virus is easy and repeatable (and easy to verify before making a vaccine out of it), and making doses from the dead virus is usually trivial. When complications arise, it's almost always in the viral growth stage. If the strain is strongly bird-associated (e.g. the H5N1 "bird flu" strain from a few years back), the flu strain might kill chicken eggs too quickly. (Thankfully, viruses with this problem rarely spread well from human to human, since the virus must make a trade-off between infecting birds well or infecting humans well.) Alternately, the virus might not grow very well at all in chicken eggs, which is a pretty rare event since chicken eggs have very few defenses and even human-adapted strains usually retain the ability
Topologically, the inner surface of your intestines is external to your body (think "the inside of a torus"); also, like your skin, they only allow selective passage of molecules.
That's a very middle-school way of putting it, but basically accurate. You leave out the fact that the immune system regularly intrudes into the gut, and in fact communicates chemically with the commensal bacteria that live in your gut (because they share a joint interest in protecting the gut from hostile bacteria). From the immune system's perspective, the gut is a front line of defense, moreso even than the tonsils and adenoids, and it's critical to pay attention to what's happening inside it.
When you eat peanut butter, you don't end up with massive gobs of peanut butter floating around in your arteries, veins and capillaries.
That's because your digestive system has sorted the peanut butter into separate sugars, lipids, and proteins, not to mention broken them up into tiny globules to maximize surface area and thus permit absorption within a reasonable timeframe.
But the proteins, while not fully intact thanks to the action of pepsin in the stomach, are still in very large polypeptide pieces since pepsin can only slice at predetermined weak points in the peptide chain. The remaining polypeptides are still large enough to behave as antigens and provoke an immune system response.
(This is, in fact, a piece of the immune system in action. The stomach successfully digests most bacteria hapless enough to fall into it. The digestion renders the bacterium harmless but the antigens from the bacterium's surface reach the intestine intact. Once in the gut, the antigens are absorbed by the intestine and presented to the immune system, allowing the adaptive immune system to start planning an antibody-and-T-cell directed attack at the invading bacteria in the nose/sinus/throat, before they can even successfully breach the epithelium.)
(There's an even better piece of evidence: mad cow / Creutzfeldt-Jakob disease. The disease is caused by a prion, a misfolded protein that catalyzes its own misfolding from healthy protein. If proteins were obliterated by the digestive system, it would be impossible for mad cow / CJD to spread: the fact that the protein can spread from brain to food to intestine to brain and continue the infection in another individual means that at least some proteins can be absorbed whole by the intestines, even without being digested.)
Try again. The shrinks claimed that it was impossible that peptic ulcers were caused by a bacterial infection. They were wrong. Get over it.
When did I ever say that "shrinks" were right? My post was purely about biologists — a shrink talking about the immune system deserves as much respect as a chiropractor claiming subluxations cause cancer (i.e. none at all). You're projecting your own interpretation on my words here.
Also, you apparently don't understand how evolution works - "weaken the immune system -- presumably as some sort of evolutionary adaptation to divert energy resources toward getting away from the source of the stress" - you have to be joking.
No, I'm not joking. The human brain is a very expensive piece of hardware -- it consumes roughly 20%-30% of our daily caloric intake, depending on how much time is spent thinking heavily. The immune system is cheaper but nonetheless not trivial to fuel: mounting an immune response is expensive enough that mice shut down their reproductive organs in order to fuel it in the face of constant energy input during an acute infection. Energy is a budget — energy in necessarily equals energy out in a healthy organism — and the budget must balance or the organism dies. Given how much energy our evolutionary environment was willing to devote to our brains, I find it outright implausible that the body never diverts energy from other systems toward the brain, when it clearly does this between other organ systems.
Weaker immune systems get culled from the gene pool all the time.
Having a strong immune system is useless to, say, an animal with its foot trapped under a rock for three days, until it can finally work itself free. Energy must be either conserved or diverted toward systems that will help the animal escape, and in that situation the only two organ systems that can assist in that are the brain and the skeletal muscles. If the animal fails to divert energy to one of these two systems, the animal will die (and be culled from the gene pool). Adrenaline is well-known for being able to divert the body's resources toward the skeletal muscles to perform feats-of-strength, and cortisol is increasingly known to trigger energy-conserving behaviors, up to and including the shutdown of the immune system. Just because cortisol is maladaptive to humans living now doesn't mean it was always so — we didn't evolve to survive committee meetings and TPS reports.
And that doesn't even get into the recent telomerase discoveries:
Elizabeth Blackburn et al. UCSF has shown work that reveals that mothers caring for their very sick children have shorter telomeres when they report that their emotional stress is at the greatest point. She also found telomerase active at the site of blockages in coronary artery tissue. This could be why heart attacks can come on so suddenly: Telomerase is driving the growth of the blockage.
Other work has shown that the poor of society have shorter telomeres than the rich.[14] Short telomeres can lead to telomeric crisis and the initiation of cancer if many other conditions are also met, or so the discussion goes at this point.[citation needed]
Blackburn and the two other co-discoverers of telomerase won the Lasker Award (2006), and the Nobel Prize (2009) for the discovery of telomerase and subsequent work on telomerase. Blackburn also won the 2006 Gruber Genetics Prize for same.
Seriously, wake up and smell the post-1995 research. You're living 15 years behind the state-of-the-art.
... and appears to be invented from thin air, same as the "your ulcers are caused by stress" bullshit that was finally disproved a couple of decades ago.
Uh, "ulcers are caused by stress" hasn't been disproven... in fact, it's been confirmed in a roundabout way. Stress is now clinically demonstrated to weaken the immune system — presumably as some sort of evolutionary adaptation to divert energy resources toward getting away from the source of the stress — and that leaves you more prone to bacterial infections, such as H. pylori in your stomach/duodenum.
You might like peanut butter, or bacon and eggs,or ketchup, but injecting any of them directly into your bloodstream isn't the same as eating them.
Yes it is. What do you think your intestines do? (Pepsin doesn't digest proteins all the way into their constituent amino acids, after all. And your intestinal cells are positively bathed in whatever peptides you just ate, and they're just as easily inflamed as any other epithelial tissue in the body.)
Richard Lenski could have saved himself a lot of time if he had asked himself "was any new information created when it mutated" . The answer of course is NO!
Even though you're a troll, I'm feeling generous today. This is completely and utterly wrong, and if you understood what information was, you'd agree with me — and all of biology — that evolution occurs within a species. (Speciation will have to wait for another day.)
Claude Shannon, of Bell Labs fame, invented Information Theory in the late 1940s for the utterly practical purpose of cramming more data onto copper wires. What he discovered, with a bit of a shock quite soon after, was that the equations were identical to those describing thermodynamic entropy. In fact, thermodynamic entropy turned out to be a special case of Information Theory. After discovering this, Shannon took to calling his discovery "information entropy".
Fundamentally, thermodynamic entropy is the unpredictableness of a physical system. The more unpredictable a physical system is, the more information it takes to describe the system. This was the link between the two.
About 10 years later in computer science, two researchers named Kolmogorov and Chaitin independently invented a hypothetical measure for the complexity of any arbitrary data: measure the length of the shortest possible computer program that can produce that data. Again, random data has the highest complexity: if the data has a pattern, then a short program can compute the pattern starting from a tiny piece of the data; but if there is no pattern in the data, the program must be large enough to duplicate a full copy of the data.
Getting back to biology: mutations add randomness to DNA. Therefore, they make the DNA less predictable, and therefore they add information and complexity to the DNA. After that, natural selection acts on that mutation: if the mutation was harmful for the cell, the cell makes fewer copies of itself; if the mutation was beneficial for the cell, the cell makes more copies of itself.
(Aside: It helps that DNA duplication is a fairly common event, especially in kingdoms like the animal kingdom where virus-like transposons infect all of our genomes. If you're a cell, and you have two copies of a gene, and one copy is mutated into something useless by a mutation, then nothing bad happens to you. In fact, if having the extra copy was a bad thing, making the copy shut up or do something else is a good thing. A lot of new proteins arose because the gene coding for them was copied then modified until it finally did something useful again, like the mammal blood clotting cascade or the photoreceptor pigments for color vision.)
Natural selection provides a filter: it layers meaning on top of the information in the DNA, in much the same way that "English" is a filter that layers meaning on top of "light-emitting screen that displays funny squiggly marks". If information "A" means "cell lives", and information "B" means "cell dies", then natural selection is the process that distinguishes between "A" and "B" by giving them meaning. Information is complexity. Information is unpredictability. Information is randomness. Information is not meaning. Meaning is something you do with information, not something the information inherently has.
Once you understand the difference between "information" and "meaning", you necessarily realize that DNA was the final, unequivocal proof that microevolution logically must exist in biology and that it's silly to argue otherwise. Your beliefs are contradicted by reality itself.
No, modern agriculture provides food as a function of how much energy and petrochemicals are put into each unit area of land. The oil price spike caused a global food crisis because not only did it increase the costs of fuel (for farm operation as well as transport) it also increased the cost of the oil derived chemical feedstocks used in fertilizer.
If you're going to repeat this myth, please take the time to come up with at least one fertilizer component made from a petroleum-derived feedstock before repeating it.
Mineral fertilizers contain nitrogen, phosphorous, and potassium (the "NPK" trio), sometimes plus other trace minerals like calcium. Oil contains none of these, so oil is worthless as a fertilizer feedstock. Only for nitrogen does this come even within spitting distance of the truth: the Haber process consumes hydrogen and atmospheric nitrogen to create ammonia, and the most common source for the hydrogen is natural gas — NOT petroleum, which is extremely carbon-rich and hydrogen-poor.
100% of petroleum in farming goes to the use of diesel: operating diesel-powered farming equipment and hauling the harvest in diesel-powered trucks. From a slightly broader perspective of all fossil fuels used in farming, the Haber process barely scratches the surface, as it's tiny compared to the need for diesel. And from an even broader net-carbon perspective, even the diesel use is massively outstripped by the conversion of natural carbon sinks (e.g. old-growth forests, peat bogs, wetlands) into farmland, which can hold far less carbon than the systems it replaced. The Haber process is too tiny to pick on, and calling fertilizers "petrochemicals" just serves as a factually incorrect "buzzword bingo" game to distract people from the real issues.
I think we're basically on the same page now.
Essentially, *if* we knew the laws of physics in sufficient detail to build an accurate simulation — i.e. had a working quantum gravity theory — *and* we could probe all the entropy of the black hole to insert into the simulation — an feat that makes quantum gravity look like high school algebra — *then* we could build the simulation and run it backward (yay time-reversible physics) to figure out how the black hole was created, including the fact that one of the formative objects was indeed a shoe. Likewise, if we were to study *all* the outflowing Hawking radiation, we could model the emission of the radiation, and from that a model of the black hole that emitted it, and from that a model of the objects that formed it, shoe included.
This is of course a "spherical cow" thought experiment because of the immense amounts of entropy involved in a star-sized black hole: just the RAM in the simulating computer would necessarily occupy a volume greater than star-sized. But for a sufficiently small and short-lived black hole — for instance, one formed by colliding a proton and antiproton with the correct kinetic energy — the numbers suddenly become much more reasonable. It's conceivable that we could perform the experiment on a microscopic black hole and arbitrarily approach that limit of understanding in our models, even if we could never truly reach it.
(At this point, the idea reminds me a bit of X-ray crystallography: in crystallography, the image generated is a fuzzy, imperfect view of the spatial Fourier transform of a physical crystalline structure, with sharp lines indicating the distance between repeating units in the lattice (i.e. the spatial frequency). I imagine that interpreting the Hawking radiation from a microscopic lab-created black hole would be a roughly similar experience — with the exception that humans can, with practice, glean meaning from a spatial Fourier transform just by looking at it, whereas I doubt a similar thing would be possible with a pictoral representation of entropy in Hawking radiation.)
That said, even if we never did achieve such an experiment with microscopic black holes, the entropy continues to be information even when we're not looking at it, or even when we have no ability to derive meaning from it — which was my original point in my first reply. Though I've definitely found this to be a useful and thought-provoking discussion.
With all due respect, I still think the mistake is on your end, in that you seem to equate entropy, information, and informational entropy.
Entropy refers to the logarithm of the number of possible states a system could be in, based on what you can observe about it. (And of course to fully generalize it, you actually sum over the probability times log of the inverse of probability, which is Shannon entropy, to account for states not being equiprobable.)
Correct.
Information, however, is the opposite of entropy: the more you know about a system (i.e. the more information you have about it), the less entropy it has (because you've reduced the number of states it could possibly be in -- that's the point Less Wrong makes about how knowing more about particles makes them colder.)
Correct, with the caveat that you're only hinting at the underlying idea. Entropy is measured relative to a frame of reference (a closed system), just as position and velocity are measured relative to a frame of reference (an inertial/non-accelerating frame in Newtonian physics and SR).
The Maxwell's Demon refutation shows that *entropy* cannot be destroyed (summed over all the universe), including informational entropy. But *information* can be destroyed -- by its conversion into entropy, or increase in the number of possible states it can be in. (In the example, after using the knowledge the demon had, the mutual *information* between the demon and the system is indeed destroyed.) What physicists warn cannot happen is destruction of entropy: collapsing mutliple states into one state, which would violate Liouville's Theorem.
Wrong emphasis: *mutual* information, not mutual *information*. "Mutual information" is a special case of the broader category of "information". Some instances of *mutual* information even belong to the restricted and rarefied subcategory of *meaningful* mutual information, i.e. mutual information that human mental algorithms are capable of making direct use of. But, again, this has nothing to do with the Shannon definition of information, which makes no claims as to whether or not the information is meaningful [1].
Why am I emphasizing the word "mutual" in all of this? Because the mutual information is a property of the closed system itself, not of either component! If two components of a closed system are correlated, it doesn't mean the equality of entropy and information is broken, it just means that they cannot be fully examined in isolation from each other.
If you were to examine each component in isolation, as if each one existed inside a closed system that excluded the other, you would dutifully arrive at an entropy figure for each one. If you then incorrectly summed the apparent entropy of these two components, as if the fictitious isolation actually existed, the sum would double-count that portion of the entropy which is correlated between the two parts. The entropy figures you calculated were only apparent entropy, not true entropy. Instead, you have to subtract out the degree to which their entropies are correlated — that is, their "negentropy" or "mutual information" — to arrive at the correct figure for the closed system as a whole. But the negentropy does not physically exist: it is a mathematical artifact of your attempt to consider each component in isolation, just as Newtonian centrifugal force is an artifact of a non-inertial reference frame. The negentropy "exists" only because it cancels out your first mathematical artifact: when you tried to consider the parts independently of their correlation, you invented "apparent entropy" figures that had nothing to do with the "true entropy" figure (relative to the larger system in which the two components are correlated).
It's enlightening to think about entropy as being the physical representation of "history". If an irreversible action happens, entropy is the historical record that we exist in
I don't know what the Maxwell's Demon refutation has to do with what I posted. I agree with all of that, and am familiar with it. In fact, I've helped to promote that very link you gave about Engines of Cognition.
My apologies, but have you followed through on the consequences? Information entropy and thermodynamic entropy are the same thing -- rather, the latter is a special case of the former -- and if black holes could permanently destroy information, then one could build a Maxwell's Demon by using a black hole to destroy the thermodynamic entropy gained in learning the particle velocities. Because Maxwell's Demon is impossible, black holes must preserve information. QED.
But anyway, I agree that perhaps I went too far in saying that no information escapes. Rather, the minimum possible information escapes. You can infer the total mass-energy that went into a black hole based on what comes out, but that's it. A black hole still removes all constraints on the degrees of freedom of a system that keep it from having the entropy that a black hole would. The degrees of freedom are constrained when e.g. the object has shape, a specific pattern of motion, remains a solid rather than a liquid, etc. We recognize such constraints as information.
Again, you're missing my point. All the information entropy is preserved, and every bit of it is present in the resulting Hawking Radiation: for each bit of information entropy that you observed falling into the black hole, there is one bit of information entropy trapped in the event horizon that you can predict better than chance. The information has been smeared and jumbled with all the other bits, but it is still in there, and it still correlates with the circumstances that led to the black hole's formation.
A human confusion is the idea that "information" has anything to do with "meaning". One bit of information allows you to correctly predict the answer to one yes/no question with equal odds. Nothing more. It requires more information to describe hot water than it does to describe cold water, but that doesn't mean the hot water is more meaningful. Bulk properties like shape, state of matter, etc. are very tiny bits of information, vastly outnumbered by the bits of information that exist closer to the atomic level. That means, if you smear and jumble the bits together, it's hard to find the ones that a human is interested in. That doesn't mean the bits ceased to exist -- any of them.
And, yes, the dust is still settling on this interpretation of black holes, but it's increasingly the dominant position among physicists. This is why I mentioned Stephen Hawking: as the original proponent of the "black holes destroy information" hypothesis, the fact that he was convinced otherwise by the evidence and recanted his old hypothesis in 2004 is extremely indicative of the mood of the field. There are still some unresolved questions, yes, but it's not obvious how to proceed: first we need to discover a coherent description of what "entropy" means (either thermodynamic or informational) in a Quantum Mechanical context, because classical entropy doesn't have a sane interpretation in QM.
Interestingly, this is wrong, and Stephen Hawking lost a bet over it.
The Second Law is actually very tightly coupled to Shannon-style Information Entropy: if you knew enough information (as a fait accompli) about the current state of a system at equilibrium, you could successfully build a Maxwell's Demon that used that information to separate the system into hot and cold reservoirs, allowing you to cancel out the entropy of the system with your information's "negentropy" (as it's called). Learning the information in the first place causes your information entropy (i.e. correlation with the system, negentropy) to increase, which by physical necessity also causes your thermodynamic entropy to increase in tandem. This is why a Maxwell's Demon doesn't work: the entropy undone in using the information is always less than (or, in a perfect system, equal to) the entropy done while learning it. (If blind faith provided non-tautological and accurate information about the universe with better-than-random chance, then you could build a Maxwell's Demon that broke the Second Law -- and since the Second Law is inviolable, it must be the case that blind faith tells you nothing... except possibly tautologies if your brain uses reversible computing. If you think hard enough about it, it also disproves substance dualism.)
In the specific case of information entropy and black holes, it turns out that the information never crosses the event horizon, and thus never has to break the speed of light limits when leaving it. As modern physics and Stephen Hawking have both discovered, all the entropy of a black hole's formative mass/energy is encoded in two dimensions as ripples in the event horizon of the black hole. Black holes have also been discovered to be maximum-entropy objects in modern physics, containing the largest amount of entropy physically possible for the volume of space enclosed by the event horizon. (This has interesting implications on the nature of reality -- look up the anti-de Sitter/CFT correspondence for all sorts of 2D/3D weirdness, like the universe being equivalent to a 2D hologram.) When a quantum of Hawking Radiation emerges from the event horizon, it carries off precisely the amount of entropy equal to the entropy carried by the change in the surface area of the event horizon when expelling the quantum, thus maintaining the invariant that the black hole is a maximum-entropy object. Because "information entropy" is another way of saying "too random to predict ahead of time" -- that's what information is, by Shannon's definition of it -- the radiation looks quite random indeed. But that doesn't mean it's uncorrelated with the history of the black hole.
Isn't Nominum that company that was formed about ten years ago for the purpose of developing the open source BIND and DHCP for ISC?
Yeah, these guys.
And now they're turning around and saying "Don't use that open source BIND because it's crap. We should know, we wrote it!"
Even more beautifully, try digging the version numbers from their nameservers:
$ dig +short @ns1.nominum.net CH TXT version.bind.
"Nominum ANS 3.0.1.0"
$ dig +short @ns2.nominum.net CH TXT version.bind.
"9.3.5-P2"
$ dig +short @ns3.nominum.net CH TXT version.bind.
"Nominum ANSPremier 4.1.0.0"
One of the 3 nameservers for their own domain is running BIND, and a fairly old version of it at that!
... Plus, I honestly don't understand why receiving calls is considered so dangerous while on the road. Text messaging is a different beast entirely, one that requires your full attention, especially when they are being sent. But unless you are one of those people who can't drive and listen to a conversation in the car or the sound system, then I don't see why receiving a phone call is any more dangerous. ...
The problem, as you grasped with texting, is attention. Driving is filled with critical moments: brief events that are capable of leading to an accident if not dealt with swiftly. They happen a lot more than we realize in retrospect: if they are dealt with well, we don't associate them with strong emotion and thus we forget them; and because we estimate the probability of an event based on how many examples of it we can call to mind, we don't realize how frequently they happen.
Having music playing in the background requires no attention, allowing you to drop attention entirely in order to focus on a critical moment. Having a conversation with a person in the car is slightly more dangerous, because your attention is more likely to be diverted when a critical moment is approaching, but once the critical moment is noticed, the conversation will gracefully pause as you deal with it, because your passenger can see your body cues as well as the road ahead of you. Having a conversation with a person on the phone, in contrast, is orders of magnitude more dangerous: the person on the other end of the phone will converse as if he or she has your undivided attention, because neither your body language nor the road is visible to him/her.
This is greatly compounded by our social expectations surrounding telephone use, which arose in the context of landlines: it's considered extremely rude to stop paying attention to the person on the other end of the line. (This is why text messaging is taking over: it's more polite, because there's no expectation of undivided attention.) Accordingly, our brains innately rate this social risk of rudeness as justifying a higher level of attention than a conversation with a passenger in your car. Your brain thinks of being rude as a danger, just as it thinks of getting into a car accident as a danger, but doesn't recognize that the difference is several orders of magnitude — it severely discounts the risk of getting into an accident, because the brain isn't innately good at understanding low-probability events (because it has a hard time calling up examples). Therefore, the innate attention systems devote undue attention to the phone conversation, increasing the risk of an accident: both from failures to recognize critical moments early enough (a risk that was present with a passenger, but now greatly amplified), and from failures to devote enough attention on the critical moments before they evolve into near-accidents or accidents (a new risk entirely, because the brain still believes the accident is unlikely).
AFAIK, a classic electromechanical meter, which is overwhelmingly typical in the United States, requires 100% of the load to flow through an inductive coil. Since an inductive coil acts as a low-pass filter, radio frequencies are stopped at the meter itself. More modern solid-state meters are more likely to use a current transformer, which does not require the load to pass through a particularly strong inductance.
... For addiction science to progress AA has got to be publicly debunked as a religious cult-like group that stagnates science ...
Ohfuckyes. AA and similar organizations have inflicted an amazing amount of damage through their cultish, quasi-religious tactics. AA's tactics almost always have 1:1 parallels with the evangelical Christian groups that they spawned from.
A prominent phenomenon in evangelical Christianity is the testimony, where the person in question describes (in gruesome detail) every wrong they did (or wish they did, or imagine themselves as wishing they did) prior to their most recent Born Again experience, baptism, or whatever ritual happens to be their flavor-of-the-month. Inevitably, they fall off the wagon in secret, get caught in public, beg for forgiveness, then start the cycle over again with another testimonial (possibly at a new church). The sin/confession cycle is deeply psychologically addictive because of the euphoric rush provided during the testimony phase. Combine that with a taboo against judging the current actions of current group members and a stubborn insistence that people are not responsible for their own actions, and they create the very helplessness that they preach, trapping people in the religion and keeping them from developing emotional stability in their lives as they lurch from one destructive behavior to the next, each time convinced that the church/God/Jesus has once again saved them from destruction.
Modulo a few changed words, this whole cycle plays itself out almost identically with AA and similar groups, except with the added disturbing component that their victims are people who have already proven themselves to be drawn to addictive patterns, making them perfect prey for the groups' emotional predation. Given that nine-tenths of any drug addiction is psychological, the actual chemical addictiveness of a given drug is almost irrelevant to the equation.
The SRY gene is responsible for only a small proportion of direct masculinization: it causes the gonads to develop as testicles, it causes production of anti-Müllerian hormone, it causes production of androgens (i.e. testosterone), and it causes some tiny changes in the brain. All the remaining primary and secondary characteristics are the result of either AMH or testosterone.
Supposing a 46-XY subject with CAIS (Complete Androgen Insensitivity Syndrome), note that AMH is still effective at de-feminizing the internal genitalia by triggering the reabsorption of the Müllerian ducts: no fallopian tubes, uterus, cervix, or upper vagina. However, because the androgen signal cannot be received, Wolffian ducts do not form: no efferent ducts, epididymis, vas deferens, or seminal vesicle. Also, the brain is not fully masculinized, as the bulk of brain masculinization is androgen-induced, not SRY-induced.
The final CAIS result is external female genitalia, a shallow vagina, active testes producing male gametes, male-normal androgen levels, and no internal genitalia (male or female) beyond the testes and vagina. The testes may remain undescended, or may descend into the labia majora (homologous to the scrotum) where they are easily palpable. Gender identity is most frequently female, due to largely absent masculinization of the brain. Secondary sexual characteristics are frequently hyper-feminine: in contrast to a CAIS individual, a typical woman express some androgen-induced characteristics (albeit vastly reduced from the same in a typical man). In particular, CAIS individuals frequently have little to no body/pubic hair, and highly feminine facial bone structure. For this reason, there is a disproportionate positive correlation between CAIS and supermodels.
PAIS (Partial Androgen Insensitivity Syndrome) is much less sharply defined, as it runs the gamut from nearly-CAIS-woman to man-with-micropenis, including everything between the two (depending on how badly impaired the androgen receptors are).
Note that the athlete in the article clearly does not have androgen insensitivity, either complete or partial, so this line of discussion is irrelevant to that topic. My best guess (as a well-educated layperson) is 46-XX, absence of SRY, with abnormally high testosterone production. The facial bone structure suggests high androgen exposure during puberty, which for reasons of exposure consistency would tend to rule out steroids or an androgenic endocrine tumor, and the female gender identity and (presumed) presence of external female genitalia rules out fetal androgen exposure from testicular tissue (e.g. chimeric absorption of a 46-XY twin brother). Perhaps some genetic disruption of the regulatory regions involved in testosterone production (more likely), or an abnormally strong response to female-typical androgen levels (less likely). My conclusion: in essence, an unusually butch woman.
A predictable response, but consider that if a person is clinically addicted to certain drugs, then not having those drugs regularly is highly likely to cause death. In that sense, there is some truth to the reasoning that "the drugs made me do it" - the drugs are necessary to avoid death, and if the person were not addicted, then the drugs would not be necessary.
Happily, even among highly chemically-addictive drugs, there are very few that can kill you if you discontinue them abruptly. Sadly, the worst of the lot (alcohol) is the most widely abused. Because alcohol is a powerful GABA agonist (i.e. amplifies the effects of GABA, an inhibitory neurotransmitter), the body responds to long-term alcohol use by decreasing sensitivity to GABA. During alcohol withdrawal, the reduced sensitivity means the body's natural GABA signals are no longer effective, causing glutamate and other excitatory transmitters to operate unchecked. Untreated withdrawal in a heavy alcoholic causes elevated heart rate, tremors, convulsions, permanent brain damage, then death.
Other highly addictive drugs with absolutely dreadful withdrawals (e.g. heroin) are rarely even remotely life threatening (however miserable the person feels while it's happening).
The concept underlying "gaydar" applies much more broadly than to just gay people. Humans naturally broadcast extremely subtle, ambiguous social signals indicating their membership in whichever subculture(s) they belong to, often through the use of word choice or references to common subcultural touchstones. Because each signal is ambiguous, they individually mean nothing and are normally tuned out as noise by non-members. However, a person who shares membership in one or more subcultures will spot the ambiguous signals and wonder, "Huh, I wonder if X is also a member of Y group". The received signal will prime them to look for more signals of the same kind, and cause them to semi-subconsciously broadcast return signals. If they see more signals in response, a feedback loop forms as they become increasingly certain, and eventually there's an unspoken knowledge by both parties that each knows the other knows that both are members of the subculture. At that point, they start a conversation.
This is an extremely broad human behavior, and applies to almost all subcultures, no matter how trivial: from ones as secretive as illegal drug users and 19th-century gay men, to ones as openly-declarable as churchgoers and swing music enthusiasts. The point at which a conversation is held depends on the risk and consequences of reading the signals incorrectly: members of more mainstream, well-known subcultures will start a conversation with each other more quickly, because the odds are good and the consequences are few. But all subcultures do it; even an enthusiast of something completely non-offensive, like crochet, won't start a conversation with a random acquaintance about crochet without first seeing a hint or two that the acquaintance actually has a crochet interest. The minimum consequence of being wrong is wasting the other person's time, which is rude and thus a social negative.
Seriously folks, is there no already existing file system that can already meet these needs? If not, then what are Google's competitors using?
Is that no one else has yet to face up to this issues properly and this is a huge competitive advantage for Google, or is it simply NIH?
HDFS (the Hadoop File System) is based on fairly similar concepts, AFAIK. I haven't seen anything that directly competes with Bigtable, but the article goes into detail about SSTables and compaction, so I don't see any barrier to writing an open source competitor to that, either.
(I've heard substantial rumors that there are already closed-source clones of both GFS and Bigtable, in addition to the significant and often well-publicized internal adoptions of Hadoop, e.g. at Yahoo. The GFS and Bigtable white papers are rather surprisingly complete.)
It's not a debate. Doing this turns those power lines into big antennas. You can't debate the laws of physics.
Note that the article specifically mentions that it's not about broadband-over-power-lines (BPL). The author is restricting himself to discussing home replacements for Ethernet, which are blocked at the meter and cannot leave the house. Yes, the wires inside the house become antennae, but it doesn't cause the widespread disruption that BPL would.
You do realize that this was in reference to Ogg Theora, right, and not Ogg Vorbis?
Theora the video codec is formerly known as VP3, and it was commercially sold by On2 Technologies before On2 open-sourced it. On2 owned (and owns) the patents on VP3/Theora, and they were confident enough that there were no third-party patents pertaining to it that they felt safe selling it commercially. If On2 says they've granted everyone a perpetual license to use the VP3 patents, then I trust On2's judgement that everyone has permission to use VP3, especially since commercial infringement opens you up to much greater patent liability than open source infringement. In a world where MJPEG probably has two or three submarine patents still floating around, it's not worth worrying about VP3/Theora.