knzhou

54 ms·

511.
by knzhou 7y ago
From the perspective of a non-biologist reader only seeing popular articles, the past ten years have been completely bewildering. Nutrition has seemingly been boiled down to a couple of magic totems that are always simplistically either goo
512.
by knzhou 7y ago
Yeah, half the elegance of the linked expressions just comes from setting A = 1. This is definitely well-known. I learned this under the name of "Vieta's formulas" back in middle school. Vieta lived in the 1500s.
513.
by knzhou 7y ago
This kind of approach is familiar to me from competition math. Back in middle school math club we were taught this exact approach (under the name of "Vieta's formulas"), i.e. that thinking about the sum and product of the roo
514.
by knzhou 7y ago
I'm a physicist that uses frequentist statistics, and it works fine. However, it can't be denied that some fields misuse it, though precisely the failure modes I pointed out.
515.
by knzhou 7y ago
> No, the p-value is defined as the likelihood of a result at least as extreme as the one we obtained, under the null hypothesis. Define for me in an objective way what "at least as extreme" is. Let's say I think the strin
516.
by knzhou 7y ago
The exact point I am making is that all of this is totally up to the researcher. This is the standard methodology in social science: yes, in theory a low p-value does nothing but support the complement of a fairly bland null hypothesis. But
517.
by knzhou 7y ago
> Wait, in frequentist statistics getting, say, a p-value of 1 is not a bad thing--unless you erroneously assume that value is evidence for your null hypothesis. That's exactly what I'm saying. Suppose you get HHTHT. Then you r
518.
by knzhou 7y ago
Adding to the other comments, you still have prior-dependence on a more subtle level, because it depends on what hypotheses are allowed. Here's an extreme example. Consider flipping an apparently fair coin and getting "THHT".
519.
by knzhou 7y ago
But this isn't actually a criticism of Bayes at all. Yes, the result depends on your prior. But the result always depends on your preconceptions -- even in frequentist statistics, where it determines which statistical tests you use a
520.
by knzhou 7y ago
That's already been happening. Basically ever since it started, there has been the same disclosure that you get in a standard print publication, i.e. it says 'advertisement' at the top. Doesn't stop anyone from complaini
521.
by knzhou 7y ago
In these comments, and in the press at large, we have another great display of "damned if you do, damned if you don't". It is constantly argued that Facebook, Google, etc. must stop political ads in America, because allowing
522.
by knzhou 7y ago
Ah, I see. I was under the impression the paper was new.
523.
by knzhou 7y ago
For almost all people, name and profile picture are set public. It's an incredibly tiny fraction of all the data, which is why I'm amused at all the uproar. There really is very little you can do with that information.
524.
by knzhou 7y ago
Libgen is one of the greatest contributors to scientific productivity worldwide, possibly beaten only by Sci-Hub. Just about everybody in academia knows about it. If it ever vanished, some of us could probably still get by trading files fro
525.
by knzhou 7y ago
This is nice, but I could've sworn I saw the exact same diagrams in all the old-school information theory books I originally learned it from. I've drawn the I-diagrams the author gives in my notes quite a few times. I'm guess
526.
by knzhou 7y ago
Sure, but you always need roughly that amount, or more, to get a deep understanding of new material. MIT operates on a 1:3 ratio of lecture to review, Cambridge is even higher.
527.
by knzhou 7y ago
You mean the digital equivalent of a big book full of people's names and faces? That's an exact description of Facebook.
528.
by knzhou 7y ago
You’re completely correct: new modes of explanation can be extremely useful, and you can’t judge from the Wikipedia article alone because it can’t get to the meat! That’s why I had to read the papers. Turns out, the meat just isn’t there, i
529.
by knzhou 7y ago
I'm always ticked off when I hear the hype around constructor theory. While information theory has had profound impacts in physics, constructor theory has always seemed to me to accomplish nothing but add fancy words. Yet it has done v
530.
by knzhou 7y ago
It's a literal angle, like you learned in geometry class. After the decay, two particles fly off in different directions. They mean the angle between those directions.
531.
by knzhou 7y ago
You need to look at the source. Most Forbes content is bad, but this is Ethan Siegel's blog, which is usually pretty good. The CERN article is also good, and the two don't actually contradict each other -- nobody is rebutting anyb
532.
by knzhou 7y ago
This would only detect the crudest type of bias —- it does absolutely nothing against selective reporting or editorial slant. People read sensational news precisely because they think “respectable” outlets are selectively and disingenuous
533.
by knzhou 7y ago
This simply doesn’t work because “information”, in the form of news articles and blog posts, consists of 10% actual facts and 90% narrative framing. As such, any distributed rating of factuality will instantly devolve into a rating of popul
534.
by knzhou 7y ago
That's true for almost any nontrivial transcendental algebraic or differential equation. Special functions to solve special cases are the norm. We could repeat the same discussion at a lower level. Suppose you tell a bright middle scho
535.
by knzhou 7y ago
Reserving any judgment on this specific action, I just note that it's all consistent with how Singapore has treated this in the past. From Lee Kuan Yew's autobiography: > My early experiences in Singapore and Malaya shaped my v
536.
by knzhou 7y ago
So when you're at the beach, hiking, or driving on the highway, what are you looking at instead? The blue sky is going to be most of your field of view unless you just stare at the ground.
537.
by knzhou 7y ago
Yes it does! A giant spike in the blue region is literally what makes the sky blue in the first place. If there weren't such a spike the sky would be white. When you look up at the sky on a typical clear day, you're getting more b
538.
by knzhou 7y ago
It's not about first impression intuition. I'll bet you'll find relativistic quantum field theory less intuitive, at the popular level, than either string theory or dark energy -- but it has more empirical support than almost
539.
by knzhou 7y ago
It was an application of Occam's razor. Dark energy arises automatically in the Standard Model. That is, you don't put it in by hand -- it emerges no matter what you do. In fact, we often get too much of it; this is the cosmolog
540.
by knzhou 7y ago
A paper containing only things we already knew for sure is not a real scientific paper at all. It's easy to sit back and feel smug because physicists might be wrong -- but it's impossible to do anything worthwhile without taking o

More ›