Half the Facts You Know Are Probably Wrong
reason.com
reason.com
One of the difficult parts of being a teenager and young adult is that you learn the rhythm and gist of how things operate without actually learning how they operate. This is one of the reasons "being cool" is such a big deal at that age -- the idea is to learn to mimic and fit in with the rest of how society works, not actually to be able to manipulate it. Faking it is much more important than actually doing it. That's what everybody else is doing anyway.
Somewhere in your 30s or 40s (perhaps sooner or later) you realize that much of life is like that: you don't really understand how a car works, only that most of the time you can get in it and go places. When it doesn't work, you take it to another person. She may also not understand what's wrong. But most all of the time she does. If she doesn't, she takes it to another person, and so on. You keep getting closer and closer to total understanding but you never reach 100%. That same pattern applies to medicine, advanced physics, and most everything else. When there's a billion areas of knowledge, even 99% knowledge leaves a helluva lot of holes. There's this surface layer of generalities and half-truths that work so often that it's just not worth diving down into the details on everything else.
On top of that, other people are very happy to share their own generalizations and heuristics, so we're kind of operating in a rumor-of-a-rumor mode. The vast majority of the time it doesn't matter, but sometimes, like when you ask questions about how startups work and most of your surrounding culture gives you bad answers, it does. Most of life isn't a geometric proof built on the linear progress of science; it's a bunch of heuristics strewn together to make something practical that you can use.
What we "know" is a product of generalizations and communications about half-truths from the environment we're in. This is one reason why it's so critical to spend a lot of time challenging the things you think you know, especially in areas that make a big difference to you personally. It also means that you must accept a deep and abiding peacefulness in not knowing very much at all about most of the stuff you actually work with in your day-to-day life.
Especially the stuff you really think you know. And the ease of doing so is inversely correlated with the amount of stuff you think you know about topic X.
That's how 90% percent of the people think (the ones I only care about using and manipulating to achieve my purposes - not necessarily egotistical purposes, mind it, I might actually ask charity donation from them), but not how I and the 10% I care to hang around with do. For example, you don't kneed to know the details of how a car works, but knowing what an internal combustion engine is and that most breaks work by rubbing two things together and basic stuff like this is what everyone should know - my grandma knows this and she can reason heuristically about this even if she has never been to school and can't read or write. Same about computers - if you are an IT manager or even a sales or marketing guy in something computer related, if, for example, you don't know the difference between compiled and interpreted code and what this has to do with portability and, in consequence, cost of multi-platform development, then you're a mindless tool (probably a bad "tool") and I'll have no respect for you and your rights.
You can't separate how it works type of knowledge from how to use it one, no matter how good the abstractions and UIs are. If it's something you're working with, then you should know the principles of how it works!
At this point, forget, how the pads operate, I'd be happy if people just knew how to spell "brakes"...
I'm not picking on you in particular- I swear it is one of the most frequently misspelled words I read online.
These things are very frequently closely related in at least one direction. Useful as a filter.
I've known people way smarter than I am that have way worse spelling, grammar and even basic communication skills than I have.
It isn't really the rule, though, is it?
Why should you know how an internal combustion engine works in order to drive a car? What's the point in knowing the effects of compiled and interpreted code on portability when so many other factors could make that basic knowledge irrelevant?
You save your actual argument for your last paragraph (you can't use it if you don't know how it works) but you never give any examples or explain why that'd even be plausible.
That's how I believe you should frame it. "Explore a little bit more of our world, it's awesome" and not "you're an idiot".
And think about the fact that people that don't know something you do, might also know all sorts of things that you don't.
> Why should you know how an internal combustion engine works in order to drive a car?
Easy—so when you go to a mechanic with an exhaust blockage and he tries to sell you a new transmission, you know enough to tell him where to stick his transmission.
Yes, this actually happened to me. Thankfully, I knew enough about cars to know I needed another mechanic.
Obviously not every driver needs to be an automotive engineer or even a technician, but it's a little silly to throw up your hands and give up on knowing anything when the basics aren't that hard to understand.
> What's the point in knowing the effects of compiled and interpreted code on portability when so many other factors could make that basic knowledge irrelevant?
At first, I thought "the effects of compiled and interpreted code on portability" was an absurd thing to talk about. Compiled vs. interpreted is a pretty fuzzy distinction (Java is compiled, but only to bytecode that's then interpreted, but that interpreter just just-in-time compiles to native code anyway), and a compiled language doesn't inherently mean you're stuck on one CPU or OS until you rewrite your whole application—sometimes you just need to recompile it. But then I have the basic knowledge that "compiled" often means "compiled to and shipped as native binaries," which matters a lot if I'm trying to run Jim Bob's Digital Accountant on my SPARC workstation running LunarOS 2.7. If that software's only distributed as a Windows binary, I'm pretty much out of luck (unless I happen to have an emulator handy that interprets Windows binaries, but there's that fuzzy distinction again), but if it's a Java application there's a chance it'll just work, as long as Java's been ported to LunarOS on SPARC.
The moral of this story: you should know this stuff so when people talk about it it's not all nonsensical babbling and you don't have to just smile and nod. And, so you can reason about things and know when someone's feeding you a load of BS.
Or, so you don't go off and, say, write an article on how Apple is going to switch to ARM so they don't have to maintain two separate versions of their applications for phones and tablets vs. laptops because you don't know code doesn't have to be rewritten from scratch for each, but that you do need a different UI on a phone than you do on a laptop.
In general, know more things so the world makes more sense, and so you make more sense interacting with the world.
No disagreement there. It's good to know things. But it's petty to label people as stupid because they don't know <random fact, skill or theory that you happen to know about> and that's what I was railing against.
I don't understand how you got from point A to point B. Where did you actually show that it's critical to be unsatisfied with the system that is worth 99% of the time?
This is due to cheap, low quality media, which broadcast and publish any crap imaginable. Pseudo-psychology, pseudo-medicine, pseudo-economics, pseudo-sociology, pseudo-everything.
Just memes made out of other memes. "Sugar is bad, eat fiber". "Paper money is bad, buy gold". "Apple will expand infinitely, buy stocks". "Java is the platform of choice", you name it. And everyone is ready to approve or argue, to express an opinion, based on his own internalized memes.
Uncontrolled consumption of lowest quality media leads to losing ability to see the world as it is, not as some talking heads around describe it.
And all those few standard responses, ready to be displayed whether appropriate or not - not this, then it must be that. lol wut?
Oh, please! "She"? Let's all ruin the flow of our prose in an attempt at social engineering that probably isn't even worthwhile, shall we? I suppose you didn't bat an eyelash at Denzel Washington as savant pilot Whip Whitaker either.
I'm growing increasingly convinced that this is wrong. Knowledge has value, but it is a strictly secondary value. It's value comes from the things that depend on it. For example, if you're an artisan and/or business person, then your primary driving force is "make stuff people want to buy". In that case, knowledge is indeed very useful for improving your outcomes (from how to build things to what people want to how to how to reach people, etc). If you are a scientist then knowledge is a critical component to building new knowledge--and subsequently defending it. Even that most basic satisfaction, the satisfaction of curiosity, is only a pleasant, transient jolt. (And that pleasure is, alas, so disconnected from the real world that even false knowledge is sufficient to trigger it, much to the chagrin of the habitually skeptical.)
So yes, I think your model is fundamentally correct, that knowledge is an extraordinarily practical thing, designed primarily to know it's limits and find a specialist when that limit is reached. But you missed the outer context that gives meaning to this conversation - the fact that knowledge is merely the handmaiden to another driving force, and a handmaiden only need be "good enough" to reach a minimum acceptable thresh-hold.
In 2005, the physician and statistician John Ioannides published “Why Most Published Research Findings Are False” in the journal PLoS Medicine. Ioannides cataloged the flaws of much biomedical research, pointing out that reported studies are less likely to be true when they are small, the postulated effect is likely to be weak, research designs and endpoints are flexible, financial and nonfinancial conflicts of interest are common, and competition in the field is fierce. Ioannides concluded that “for many current scientific fields, claimed research findings may often be simply accurate measures of the prevailing bias.”
If Ioannidis' findings apply to areas of science other than biomedical research (where there is also competition, pressure to publish, conflict of interest, etc.), then most scientific knowledge is already false at the time that it's published; you don't have to wait 45 years for it to become false like this article claims.
Here are some references to Ioannidis' work:
- The cited PLoS paper (2005): http://www.plosmedicine.org/article/info:doi/10.1371/journal...
- A less technical article in The Atlantic: "Lies, Damned Lies, and Medical Science" (2010): http://www.theatlantic.com/magazine/archive/2010/11/lies-dam...
Edit: Corrected spelling of Ioannidis' name; it's misspelled in the article.
The front line of science is a messy place; a mostly wrong messy place, as any of us who have spent time there know. A recent study claimed massive error rates across all scientific papers - which is not a surprise to scientists. The closer to the edge of knowledge you come, the more wrong you'll find - a great frothing sea of wrong, enthusiastically generated by scientists in search of nuggets of right. It's all part of the process, and you have to step back from the details in order to see where the process is taking you. In any complex field, and biotechnology and medicine are about as complex is it gets outside astrophysics, validating truth takes time. Scratch any unanswered question and it'll bleed papers and reviews, a dozen for any given position on the topic.
Reading the article, the most interesting aspect was the concept of a "half-life" of knowledge. This, of course, applies not just to the front lines, but also to the rear guard.
By definition, a new idea that proves to work well obsoletes an old one that didn't work quite as well.
It can take a lot longer than a half-century to catch up.
For example, it's surprising how often I hear people talk about "exponential" population growth, which was the going theory in 1798 [0] but known to be inaccurate by 1838 [1]. I hear similarly out-of-date comments on topics ranging from evolution to Bible manuscripts with shocking regularity.
People sometimes learn "facts" from trusted friends or preferred authority figures, who learned them the same way, going back generations without anybody thinking to double-check. I suspect the emotional cost of finding out a friend was wrong provides just enough friction to keep people from looking things up, even in a world with wikipedia and snopes at our fingertips.
[0] http://en.wikipedia.org/wiki/An_Essay_on_the_Principle_of_Po...
[1] http://en.wikipedia.org/wiki/Logistic_function#In_ecology:_m...
I also feel like its a pretty good analogue for modern life. People seem less willing to discover the truth than they are to use facts that support the point they want to make while ignoring the inconvenient parts of the facts.
therefore, it is a more lonesome endeavour, and harder than subscribing to the pap fed to the masses.
see my [other](http://news.ycombinator.com/item?id=4968390) comment.
(edit: link, grammar &c.)
One scientist might come up with a hypothesis, never test it, but nevertheless believe it devoutly. But science progresses because other scientists are sceptical and test the hypothesis.
It's in the first scientist's self-interest that the hypothesis is proven true, but it's generally more in other scientists' self-interest to prove the hypothesis false.
Some of the greatest scientists of the 20th century closed their eyes to overwhelming evidence that their pet hypothesis was wrong. One of the classic examples was Geoffrey Burbidge (known best for showing that the heavy elements were created in stars and released into the universe in supernovae) who believed until the day he died that quasars were galactic objects.
This topic is still very controversial and I don't think the article should be citing it in this context.
It was only a few months ago that the ENCODE project released its first results claiming that 80% of the genome is "functional", and many scientists have noted that their definition of "functional" is much too broad. It includes any piece of DNA that becomes bound to protein at some time, while it is known that many proteins bind nonspecifically.
No publishing on DNA until we know everything, etc.
I'm cautious about children. I've seen plenty of boys learn a long list of (for example) dinosaur names. Is this something I should try to avoid with my son, or should I encourage it, or should I just remain neutral? I have no idea.
I do know that I want to encourage him to think about relationships between things, rather than just learning a list of names. (For example; "That tree is evergreen, which means it keeps its leaves all year round. It does that because [...]" rather than "That's a Leylandii".
I don't suppose there's any harm in loading your children up with trivial knowledge, but I would certainly put more effort into teaching them thinking skills than specific knowledge.
These days, if you have the necessary thinking skills, the sum of human knowledge is within reaching distance. If you know how to research a subject, how to corroborate and verify the veracity of positive claims and theories presented to you, you can find out pretty much everything about anything in a matter of minutes.
I suppose it's the old "give a man a fish" adage applied to knowledge.
I find that kids are like sponges for details like names. They'll soak them up whether you're teaching them intentionally or not. Or if you're trying to get your kid to learn animals or stars or whatever, they'll be memorizing scores of baseball statistics, or Pokemon, or airplanes.
Teach thinking skills; the trivia will take care of itself.
Sorry for the diversion: it was a question in today's Trivial Pursuit.
Avoid.
I would avoid spending time learning the names of things. Learn the thing itself, but the name tells you very little.
Watch this: http://www.youtube.com/watch?v=05WS0WN7zMQ
Dinosaur names change or get dropped, only because of more information becoming available. Extant species' names are more settled, but mostly because we have so much more information. I expect that the dinosaur species with the most fossil evidence aren't going to have their names change any time soon, unless some sort of new, general approach to naming sweeps the field leading to rethinking the whole thing.
What should be avoided is unintentionally creating the kind of mind-set that could have a hard time dealing with corrections later on in life.
Worst-case scenario, years later he finds out a favorite dinosaur of his youth was a misnamed individual of a different type, and ends up deciding that scientists are frauds and becomes a young-earth creationist.
Basically, as long as there's understanding that scientists are generally doing the best they can with the available information; that scientists in the early years of a new field may have made more mistakes due to the limited knowledge framework (early misconception of the iguanadon's thumb spike as a nose spike, etc); and that scientific errors made in good faith are a sign of science's strength, because they can be corrected by sufficient contrary evidence.
Basically information is power. Due to informational asymmetry[1] being exploitable via economic, political and other forms of power, it is almost always in somebody else's interest that you be misinformed.
Therefore the person with the most to lose (and the person with the best interest in being informed) is you. And instead of seeking out the forms of information that empower us, we fritter away our attentions on trivialities and pop culture (yes, including programming pop culture). A good example of a prophet who tries to steer us programmers the right way is Kalzumeus, imho.
Part of the reason we fritter away our time is that it is simply easier to consume the mass-delusion, because the search for truth can be fairly lonesome (since you are the only interested party).
Under the circumstances outlined above, it is no wonder that most of what we believe may be false, and part of a larger (self-bootstrapping) conspiracy to keep us in the dark (cue ominous music, break out the tin-foil hats).
http://xkcd.com/843/ (Misconceptions)
http://xkcd.com/1053/ (Ten Thousand)
http://en.wikipedia.org/wiki/List_of_common_misconceptions (The link referred to in 'Misconceptions')
EDIT: Darnit.
Is that a typo or humor about facts being wrong?
Consider this "fact" from the article, which has no citation whatsoever:
"Increased K-12 spending and lower pupil/teacher ratios boost public school student outcomes."
Which the author claims is "wrong."
Actually the evidence suggests that it is right, although the effect of class size alone may be small. And it is not just a matter of how much an effect reducing class size has on learning, but also in what circumstances and to what extent. And even with these "facts", it doesn't mean that big classes are inherently always "bad" and useless - MOOCs are still quite useful even though 90% of students don't finish them. And large class lectures actually could be made more effective if given after (not before) a lab or other interactive learning activity. But a wealth of research has showed how one on one tutoring is generally the most effective method of teaching (which also directly contradicts the author's assertion that reducing class size has no effect).
http://www.districtadministration.com/article/does-class-siz... http://www.ppta.org.nz/index.php/-issues-in-education/class-...
However, in a recent conversation with my mother, I asked her about a talking donkey in the bible. she confirmed the reference saying "the lord works in mysterious ways" and "donkeys could talk back then". Maybe there is some truth that 1/2 of our facts are wrong... If you skip religion, I view my mother as a very clever woman who is very astute with business. She was indoctrinated at a young age.
We're hardwired for it.
Like all of the best religious ideas, it is one that came about by a man faced with his own mortality.
On a side note, and this is opening a huge philosophical can of worms, I'm not convinced of the utility in merging our intelligence to machine intelligence in such a scenario, either from the point of view of overall utility (what do they gain from our meeker intelligence?) or from the point of view of our own "existence" actually being near-immortalized.
Stated simply, if you could make a perfect copy of my brain into a 20 year old clone of myself (I'm 39) and that clone were created such that it didn't have to worry about issues like telomere shortening or any of the other processes theorized to cause issues for clones, but after the process was completed the original copy of me would be terminated, I wouldn't take that deal and I doubt many other people would either regardless of how they feel about the theoretical ability to duplicate an exact copy of their own consciousness. Ultimately, the copy of me that is me wants to continue on but cannot.
I will die.
We are most definitely not hard wired for religion. It's a mythology that preceded the scientific era. In a thousand years or less it will be laughed off like the talking donkey nonsense. We're not hard wired at all...
I agree there's little that's religious about the Singularity itself or most Singularitarians (http://www.acceleratingfuture.com/steven/?p=21), and the above poster is probably mislead about it. (I suspect he's mislead because he brought up the 2045 date and because he's confused about identity. That date I believe is only professed strongly by Kurzweil in the public mediasphere, who is at most with respect to the Singularity a predictor with some past successes and equipped with a strong argument based on accelerated returns, but he isn't working directly toward a positive (or negative) Singularity while those who are that I'm aware of hesitate to estimate date ranges.) Personally I think barring mass human extinction the stage is set for a Singularity event within this century, and I do have many justifications for it, not a religious "I have faith and faith is good" 'justification' or an apologetic one trying to extract justification by means of some shaky philosophy and an a priori sacred book (Singularity is Near? Never read it) containing no significant errors.
Now, Marx couldn't think of politics as independent from the interests of the capitalist class. That's why he failed, as Popper described: the New Deal, Social Security, public and subsidized education, progressive taxation... all those things came and made his prophecy fail. As in hindsight was to be expected, if only because prophecies rarely become true.
Today, Kurzweil can't think of biology as the vast, incredibly complex area that it is, and so his prophecy will fail when it's clear how difficult it is to achieve immortality, or full cybernetisation of the human mind, or general AI, or whatever. I am 90% confident in that outcome.
Yeah, we are not hard wired for religion, but still it's very easy to fall for it. Scientific reasoning is not and probably never will be the norm, if simply for the fact that we're just poor dumb chimps. After all, religion didn't stop appearing after 1900. The Scientologists are proof of that.
[1] https://en.wikipedia.org/wiki/Millenarianism : a Great Event will come during which Humanity will be Judged and Reborn. Singularity just takes out the Judgement part.
[2] LOAR forgets how Moore's law, for example, was very specific and probably short-term. If Kurzweil had been born in the 19th century, he would have probably been talking about the coming Steam Engine Singularity. In the 60s, he'd be talking about scientific bases on Mars by 2000...
I agree that predicting when it would happen is somewhat silly. Also, I had no idea that "merging" humans into machines had anything to do with the singularity.
If you just take the notion that eventually AI will surpass human intelligence, I think it is a very reasonable idea.
And on top of that I'm claiming that the singularity will never happen.
We have incredibly fast computers with enormous amounts of memory, yet are no closer to AI.
And even given infinite computing power, we still wouldn't have AI, so the race for computing power is a red herring. (It's not like we know how to do it, but the computers are not powerful enough - we don't have the slightest clue how make AI.)
(I define AI as the ability to learn any topic at all with unstructured instruction - like reading a book, or being told verbally. Additionally after learning the material an AI must be able to invent and synthesize new material from the old.)
And on top of that computers aren't getting any faster, that's basically stopped. (Although if AI is parallelizable it might not matter.)
Supercomputers today are the equivalent of a tractor - they do simple things that people can do but at a greater scale. Nobody ever looked at the first tractor and claimed it was a step towards better AI, and nor should they when they look at supercomputers.
More and more functions that were once human-only are now a machine capability. Intelligently playing games, voice recognition, image recognition, and some (though this is still undergoing intensive research) working with natural language.
I really really don't see how you can believe that humans will forever be more intelligent than machines. What makes humans so special? Why would you believe that?
You'll get a lot of negative replies about this, after all HN is very pro-Singularity or however you want to call it, but my gut feeling is that you're right.
I can see Moore's law being applied successfully for the next 100 years in one form or another (more processing cores, quantum computing finally taking off etc.), enough for us to reach a state where we could say "Mission accomplished" and let the machines do our job, but I cannot imagine, not even in the remotest of ways, how AI humor would sound like, meaning how would the machines make jokes. In fact, that would be my "personal" Turing-test for deciding that the Singularity has truly arrived: can the machines understand and perpetuate humor?
Everyone has a worldview/belief system, no matter how scientific they are.
An belief systems are non-rational. These are the default nodes in your cognitive framework. They aren't evaluated, they are just assumed to be facts. They aren't easily changeable. For anyone, no matter how good they are at reasoning.
Except some worldviews aren't based on beliefs, if you define 'belief' as 'blind faith'.
But leaving donkeys aside, we are not usually thinking with extreme scientific rigor, because that would be too expensive and, frankly, wasteful. Frequently heuristics - also known as cognitive biases, prejudices, etc. - serve as well enough for everyday life. Of course, sometimes you need much better precision and then you have to know how to do it.
If 'half of our facts were wrong' is taken at face value, then our truth-determining process is no better than flipping a coin. Do you think we could engineer bridges, build computers, get to the moon, sequence the human genome, or write software if every time we didn't know something we just asked it as a yes-no question and flipped a coin?
So as much as I want to support skepticism and free-thinking or whatever this article is aiming at, this article is wrong and bad. I have the means to go out and determine true facts and I do this routinely.
Some of the examples point this out nicely, let's take the 10 biggest cities in the US. Isn't it obvious that this "fact" only works with a timestamp attached to it? The 10 biggest cities in 1990?
Maybe this is more about the language in science, textbooks and science journalism. A little more precision would make the real facts stand out better.
I don't much like the depiction of scientific knowledge as "manufactured facts".
Its amazing to me how much effort we as human beings put into memorizing and retaining 'factual' knowledge like this. Can you imagine how far we could go if all this knowledge could literally be loaded into our brains at an early age so that instead of basic arithmetic (for example) kids in elementary were already doing quantum physics and things of that nature.
I know its far far away, but it really really fascinates me to think of the implications of just doing away with having to relearn "learned" things so that we could devote energies to far more advanced topics.
Now to which half does this fact belong?
Hey, you never know, maybe in ten years we'll find out tobacco use doesn't cause cancer. Better remove all the restrictions that have been placed on tobacco until we're certain.
What?
In 7.5 years, perhaps it'll be up to 192 neurons, if the connected computers and processors can keep up with the data flow.
7.5 years after that, maybe it'll be up to 384. Etc.
>A scientific theory is a well-substantiated explanation of some aspect of the natural world, based on a body of facts that have been repeatedly confirmed through observation and experiment. Such fact-supported theories are not "guesses" but reliable accounts of the real world.[2]
[1]: https://en.wikipedia.org/wiki/Scientific_theory
[2]: http://www.aaas.org/news/press_room/evolution/qanda.shtml
But you could falsify the theory by finding characteristics incompatible with being exothermic. Finding the dinosaurs with feathers and other characteristics of the (warm-blooded) birds probably helped, too.
Some of the "incorrect facts" are really theories that were substantiated to the best degree possible at the time, and no falsifying evidence existed at the time.
I suspect some, like "non-functional junk DNA", is a case of scientific qualifiers being ignored in the mainstream media and forgotten.
Humans like "facts", but we inevitably drop the error bars that originally accompanied each new "fact".
Where this gets interesting is that you can say, what if this axiom was not true? Let's assume it's not. And then possibly come up with a whole new system of mathematics (like non-euclidean geometry) which may turn out to have actual use in previously unsolvable problems.
Issac Newton's physics, while not applying to all possible contexts, still holds as correct over a vast range, and that's because it's properly identifying an aspect of Nature's behavior, and it did so due to Newton's extreme methodological diligence. Some of these other "facts" that get overturned likely did not adhere to Newton's Rules of Reasoning, or had some other flaw that led people to jump to conclusions prematurely.
Of course, it's exceedingly unpopular to point out that we should learn why we make mistakes and then strive to not make them by scrutinizing the source of the error and adopting better methods. Most people are amenable to having their mistakes pointed out, but are extremely offended if you try to point out that the way they are arriving at conclusions is mistaken. It's as if most people don't want to believe that yes, there is a difference between wisdom and foolishness, and instead want to believe that the highest form of foolishness is in fact to believe in wisdom.