The problem is that conspiracy nuts don't understand statistics, or science. They are always asking for "proof" or "certainty" and in the the face of its lack they default to their own crazy idea.

But knowledge is not certain: it can always be updated in the face of new (possibly currently-unimaginable) evidence. So science, which creates knowledge, cannot create certainty of any kind. Not even falsification is certain, and people who ask for certainty are like alchemists of old, who rejected mere chemistry because it didn't give them the impossible capability of turning base metals into gold.

All that said, I have found it useful to ask Moon-landing-deniers who say, "The only source of light is the sun", "Then how can I see the ground?" If they acknowledge they can see the ground in the photos, they are pretty much forced to admit there is reflected light from the ground so they start looking pretty foolish when they repeat (as they often do) "The only source of light is the sun". It doesn't convince them of anything, but it does shut them up and make them go away.

The spectral shape of the dust signal and the CMB signal are different, so multi-frequency analysis can be used to tease out the various effects. There are various technical reasons the BICEP2 team chose a single frequency (150 MHz) but additional measurements at different frequencies (353 MHz is mentioned in TFA) would allow untangling of the two, albeit with some loss in sensitivity. This sounds like the approach that will be taken in future by various teams investigating this.

The BICEP2 team got bitten by designing their experiment around a flawed theoretical understanding of the dust distribution in our galaxy, which is too bad, but this is the way science works: we publicly test our ideas, and let others see if what we've done can stand up to scrutiny. I've worked on experiments in the design phase where the design team has missed important backgrounds: it's easier than you'd think. And I once worked on an experiment that had been taking data for two years before it was found that there was a background process that precisely aped the signal (it had been missed originally because of a dropped sign in a calculation that caused two terms to nearly cancel when they should have added up.)

In the best case we find these sorts of things before we publish. In the worst case--such as this one--after.

Actually, we're pretty sure dark matter is relatively smooth in its distribution and does not form anything even remotely equivalent to stars and planets (we'd see the gravitational lensing if it did). Nor can we figure out any way it could self-interact in a way that would sustain imperfectly replicating objects but not end up in clumps that eventually form large agglomerations.

This is not to say that such things are impossible--the fact that we can't imagine something is not evidence for its non-existence--but likewise no one has been able to make the notion more than minimally plausible.

To repeat: the evidence is that dark matter does not clump, and so far as we know clumping (self-interaction) is required for self-replication. If you can come up with a way to make self-replication with almost no self-interaction, it would help. Such self-replicating systems would have to be extremely large (galaxy-size or larger?) as the size of a structure is generally inversely proportional to the strength of the coupling between it's parts.

With regard to the evidence for non-clumping dark matter, it comes in two forms. One is the direct observation of an absence of gravitational lensing by dark-matter clumps except on very large scales. The other is that if dark matter gets too clumpy it screws up the dynamics of normal matter on cosmic scales in such a way that stars and galaxies don't form in the way we observe them to.

Simply because we have "never seen" something does not mean we don't have evidence that can allow us to draw useful inferences.

To get a sense of the energies involved: if you're a light-year way from a supernova, the neutrinos will kill you, even though they barely interact with matter at all. One light year is pretty close, but the gamma ray flux is comparable to the neutrino flux and forty orders of magnitude more deadly, so no, deep ocean life even at great distances from the exploding star will be boiled rather than fried, but it would still be dead.

...the discipline of publicly testing ideas by systematic observation, controlled experiment, and Bayesian inference.

Science is not a "method". Feyrabend was more nearly right than he realized when he said the cardinal rule of science is, "Anything goes": we can use any clever tricks that pass the tests to change the posterior plausibility of an idea, and they do not have to adhere to some philosopher's notions of method.

Science is a discipline, and like any other discipline has to be practiced to get good at it. Methods in science are like katas in fighting disciplines: valuable training devices, but not anything like sufficient to win a real fight.

Furthermore, as a discipline, science does not explain anything and has no content: the sciences (biology, physics, geology...) do, but not the overarching discipline of science itself.

The discipline of science can be practiced by anyone, although history has shown that education can help (try inventing any fighting discipline on your own and you'll see how much better off you'd be learning from someone else.) The scope of science is unlimited, and it is the only way of creating knowledge. It is not "scientism" to practice the discipline of science when testing ideas about human behaviour or society: it is just science.

Because science is Bayesian, it does not produce certainty. Bayes' rule cannot generate a plausibility of 0 or 1 for any proposition, and it identifies anyone who assigns a plausibility of 0 or 1 as being in a state of sin... err... error.

A proposition that has 0 or 1 plausibility cannot have its plausibility changed by further applications of Bayes rule, so it is beyond correction, opaque to any further evidence, cut off from the world it claims to apply to.

The technical term for a belief held in such an erroneous fashion is "faith".

Except that after every single warm snap or hurricane the same people who were smugly reminding us that "weather is not climate" are busy pointing to the event as "evidence" of AGW, which it is not. Distributions are evidence, events are not, because the science shows that the AGW/no-AGW distributions substantially overlap in our current situation, particularly with regard to extreme weather events.

So only a person who does not understand science and statistics would ever suggest that any single event or small handful of events is worth mentioning as evidence either way. And yet Warmists are always out in force after any given extremum telling us it "proves" AGW.

Don't get me wrong: AGW is real, although there are some pretty well-proven techniques for reducing it (notably carbon taxes, which also have the benefit of reducing corporate and income taxes, so you'd have to be some anti-capitalist nut-job to oppose them). But anyone who ever opens their mouth to point at a single event as if it was somehow worth the bother of discussing is an anti-science wing-nut, and adds only heat, not light, to the discussion.

...and no one knows what to do to fix it.

In 2010 the new Web was all about "user generated content". Today, the modern mantra is: "Don't read the comments"

Reviews and review sites have almost exactly the same problems as comment sections: there is no way to filter the ignorant and/or malicious from the informed and sincere. Case in point: there are currently exactly two reviews of my book on Amazon. One from a reasonably thoughtful reader (3 stars) and one from a troll who apparently has given Charles Dickens the same rating as me (2 stars).

There was a five-star review which was from someone who had read the book and genuinely liked it, but Amazon determined it was from someone I knew (likely because I bought her a book on the site a few years ago) and removed the review. This is a ridiculous practice--it would invalidate a huge number of reviews in traditional publications--but is made necessary by authors who try to game the review system in the stupidest possible way.

If there is a solution to these problems it's likely some kind of reputation system, but as near as I can tell no one--not Amazon, not GoodReads, not TripAdvisor, not Yelp, not anyone--is even thinking along those lines, which suggests there is no money in building a site that provides honest peer-to-peer feedback. This is a shame, because the Web should be enabling us to help each other, not increasing our distrust of each other (we're plenty good enough at that already).

/. has had a basically functional reputation system for well over a decade, so it's not like there's any real mystery as to how to do this. I wonder if there might be some b2b model where users sign up with a third party reputation system that then sells reputation information (which would exist across all sites that use it, like discus does for comments) to review sites. Without something like that there seems to be very little hope of getting much long-term value from online reviews of any kind.

You would be amazed how unselfaware many startups are. In the late 90's, early 2000's time period I frequently had to remind people in companies with 2 - 200 employees selling niche products that "But Microsoft does it that way!" was an argument against doing it that way for us, because we were anything like Microsoft in terms of resources, product or market.

You'd think that no one would ever have to be told that, but the reality is that most people look at something as incredibly difficult to build as Windows (in software) or an iPhone (in hardware) and think, "Yeah, I could knock that out over a weekend and ship a few million units a year, no problem!"

I've never been able to find any reliable under-reporting data for men, so this would be extremely interesting to see.

A priori I find it fairly implausible that men failing to report sexual assault is a lot more common than women, but would love to see the data. One informal observation is that in the multi-thousand-comment threads that are spawned after every accusation leveled at a public figure like Michael Shermer, there seem to be a lot of women self-identifying as victims of sexual assault but no men. Given that rates of sexual assault on adult men are reported at 10% of women's rate, and that male children are at least as vulnerable as female children (as the data here suggest) it is more than a little odd that no man seems willing to self-identify as a survivor.

At the very least this speaks to the way in which we silence men's voices in these debates, which in my view should be understood not in terms of women vs men but citizens vs predators (most predators are men, but most men are not predators.)

Two words: "kin selection".

Humans and chimps are social primates. We live in groups that are relatively close to us, genetically, although humans practice exogamy (mating outside their immediate kin group) a lot more aggressively than any of our cousins.

So to say "fighting for mates is always one vs one" is to say "kin selection does not exist", which it manifestly does.

War is mate competition carried out by other means. There is no other rational for it (war is always economically irrational, although this is not generally understood because it "just makes sense" to so many people that war is somehow a good idea.)

No individual of any species ever under any circumstances kills another member of the same species for any reason other than mate competition, either for themselves or for close kin (this is not quite true, but it should be the starting point of any analysis of deadly interpersonal violence.) Killing has zero to do with hunting behaviour--both male and female bonobos hunt, and don't kill each other. Elk are vegetarian, and do kill each other. Only when reproduction is on the line does the risk of being killed in a potentially deadly fight make evolutionary sense, in humans as well as in other species.

In humans, war creates all kinds of mating opportunities beyond the simple-minded "conquer the enemy and rape their women" scenario. In particular, it creates opportunities on the home front of all kinds, and that is a very fundamental part of its completely irrational appeal.

Probably pretty much everything.

Matter and anti-matter are--up to a flip in charge and parity--the same thing. That is, if you take an electron (a matter particle), flip its charge and look at in a mirror you'll see a positron (an anti-matter particle).

So it is actually perfectly consistent, logically if not linguistically, for dark matter to be entirely anti-matter.

Exotic dark matter can also produce anti-matter when its particles collide with each other, which is what this report seems to be about. The significant thing is that the energy spectrum of the positrons that the AMS detector sees appear to have about the right energy spectrum for one particular type of exotic dark matter (which I personally have a pretty low prior for).

There are a whole bunch of follow-on papers from other people doing what scientists do, which is check for consistency between the exotic dark matter interpretation of this result and reality, in the sense that if this signal really is due to exotic dark matter there should be a number of different consequences (including the anti-proton signal the article mentions): http://arxiv.org/find/all/1/al...

Yup, and the solution available to any rational being is the same: since by hypothesis the two choices are indistinguishable, flip a coin to create a new situation in which one of them has a trivial weight on its side.

Starving to death (or letting everyone die) is obviously inferior to this to any rational being (which the donkey and the robot are both presumed to be) and adding randomness is a perfectly general solution to the problem.

Buridan's donkey is not in fact an example of a rational being, but rather a passive, uncreative being, who must for some unspecified reason decide without acting on the situation, as if it was living in some bizarrely unrealistic world like Plato's Cave, where it could only know the world via shadows on the wall which it cannot act on in any way.

Why anyone thinks thought-experiments about such limited beings, which are completely unlike humans in their inability to act on the world to change their situation, is beyond me.

...that the Earth is flat.

"It stands to reason", "it just makes sense", "it's common sense"... these are not just not arguments, they are anti-arguments: anyone using them is saying loudly and clearly "I have nothing to contribute to this discussion but here's some noise to dilute the signal."

Any time you find yourself offering an opinion based only on your imagination, please don't. Get some data, learn some modelling, do some statistics before you speak.

Philosophers attempted to understand the world for thousands of years based on what "just makes sense" and failed completely and utterly. After three hundred years of scientists showing us a better way--and showing that what "stands to reason" has absolutely nothing at all to do with the way the world actually is--there is really very little excuse for continuing to promulgate this erroneous and basically useless way of knowing.

All true, in some strict sense. But...

Science lacks something that gives religion a ridiculous amount of power: narrative. (shameless plug) I wrote a book exploring this subject: http://www.amazon.com/Darwins-...

The gist of my argument is--in the terms of TFA--is that "Spockism" lacks narrative hooks, while "Kirkism" is full of them. "Science fiction" is an attempt to give science narrative power, and sometimes it really works, but it needs to be continually renewed because unlike religion science moves and changes and grows, so each generation needs its new Asimov or Heinlein or Clarke.

Nope. A patent is a license to sue. Small players rarely have the resources to do so. A very small number take the risk, fewer still manage it successfully. Pointing to one or two cases where small players were successful is not an argument. You have to look at all patents held by small players, find out how many get violated and what fraction of those use the courts or plausible threat of legal action to defend themselves.

I don't have the numbers, but from an insiders perspective (I am a small patent holder and have worked for a number of small players with patents) I can tell you that the average small player is very unlikely take court action, and that the average large player is unlikely to be much bothered by a threat of patent litigation from a small player, because they know they can simply exhaust the small player's resources.