Taleb is... not a good source for learning statistics. Start with Wasserman. Taleb says obvious and well known things using his own invented terminology in order to cast himself as some sort of contrarian genius. It's not that he's wrong, it's that the insights he hawks are banal. That's why his readership base are insight porn book junkies not people actually trying to learn statistical methods.
Yeah, I think I first heard it in relation to Malcolm Gladwell and it's just so apt at capturing everything wrong with that category of book. I mean he's a skillful writer, and it's definitely entertaining stuff. But if you flip into critical mode and do comparative research vs authoritative sources, you start seeing how vapid it is really fast.
When I read Fooled by Randomness I found it useful. Not groundbreaking work, but it drew some nice analogies between statistical distributions and human's over-certainty.
If you're referring to "All of Statistics" by Wasserman, then there are some significantly easier textbooks to learn statistics from. Depending on the program, "All of Statistics" is a book used by senior undergrads or grad students. Are there more mathematical heavy stats books, yes, but this isn't a casual read for someone who is trying to learn statistics either.
I like "Probability and Statistics for Engineering and the Sciences" by Devore as an intro book. It covers the basics of probability distributions, maximum likelihood and method of moments estimation, ANOVA, and linear regression. Pre-requisite knowledge is probably multivariable calculus, matrix multiplication, determinants, and eigenvalues.
Devore's book is great. It's sad it gets many negative reviews. In my experience, there are two types of people:
1. Those who want a statistics book to be like a math book: Fewer words and more equations.
2. Those who want a wordy book with little math
Devore's book is in between, which is why I think both camps tend to hate it. It has a decent amount of math, and has quite a bit of text. The text is invaluable: You get information about common rules of thumb. You get insights on why the technique works. Etc.
And the examples/problems are great. So many of them are from real papers/books. You're not working on some contrived example, but on real world problems.
I have read this book and want to leave an anti-recommendation here. It's a poorly edited mess and makes at least one blatant mathematical error.
More broadly, let me leave a Taleb anti-recommendation. His entire shtick is yelling that traditional statisticians have ignored heavy-tailed random variables in their modeling and that he has special insight into the nature of tail risk (perhaps along with a few select other people, like Mandelbrot).
But this is manifestly not the case. In fact, if you go through his Amazon reviews page, you can find him leaving positive reviews several years ago on all the books written by traditional statisticians that he learned about heavy-tailed randomness from!
For a more detailed critique, see Robert Lund, Revenge of the White Swan, The American Statistician
Vol. 61, No. 3 (Aug., 2007). Accessible through your favorite Russian website.
If you want a better book on heavy-tailed randomness, I like
Didier Sornette's
Critical Phenomena in Natural Sciences (subtitled
Chaos, Fractals, Selforganization and Disorder: Concepts and Tools).
calling someone's book suggestion an "advertisement" is rude, and inaccurate. taleb wouldn't pay anyone to suggest his book when he could instead just show up here and insult everyone for free.
I didn’t mean it strictly literally. The original comment was a thoughtless namedrop of a brand new book (which means by tautology it’s irrelevant to the topic at hand) and doesn’t have a shred of reasoning behind it. So, functionally, it’s a billboard advertising Taleb’s book.
I am aware that the Cult of Taleb means that people are willing to advertise his work for free.
Quite plausible. Extreme Value theory [1] appears to have been codified by the 1960s, and one of the main theorems is credited “to Fréchet (1927), Ronald Fisher and Leonard Henry Caleb Tippett (1928), Mises (1936) and Gnedenko (1943)” [2]. ETA: And the second theorem of Extreme Value Analysis is from the mid 1970s. [3]
Rather than proposing a new technological solution to the problem ("platforms to protocols") which may work but is unproven, we could simply restart the use of existing tools that have been proven to work. This has the advantage of being a strategy that can be explained to non-technical folks.
If I have a car that worked great 50 years ago but haven't been able to turn it on since the 70s when my neighbor destroyed the engine, I would not call that a proven, working car.
The "existing tools" have been proven to have a major weakness, namely they can be neutered by regulatory capture. Unless you can get everyone to magically forget how they circumvented the rules last time, you can't simply restart using them.
"The government has been unwilling to use existing powers against powerful corporations, maybe if we give them new and more expansive powers that'll do the trick. There's no possible way that these new powers will be turned against the powerless, right?"
I think most "master" software engineers are a bit more idealistic than that... Not saying I am a master but I occasionally write something and consider it genuinely beautiful.
It seems like with Quora you either get very bad or very good answers. The good ones tend to be significantly older, i.e. before it got big and people started to gamify the system.
It's not clear who made Telegram's crypto but it was probably Nikolai Durov who is a genius mathematician and engineer, so it's not like it's some guy in a garage.
I remember he was pretty cocky here on HN about it. Half a year later it was shown to be thoroughly broken because the server could provide shit entropy to the client which it for some godforsaken reason used to generate the encryption keys.
So, yes. He (or any of the other 6 world champion coders that were bragged about) designed a broken protocol. Broken, as in "trivial access to the plain text by the server", which is pretty damn awful.
When people say that telegram rolled their own crypto that mean that they chose weird crypto primitives, weird key handling and generation and combined them in a way that made people go "whoa! That looks weird" and defended it by refering to their own world championship victories, by extension saying that people were to stupid to understand.
And 6 months later their protocol was shown to be broken.
Use telegram all you like. Don't say that their encryption is even close to being the gold standard.
Signal's crypto has also been independently audited [0], with pretty encouraging results, and as it is open source, can continually be audited.
Telegram's on the other hand, can't be continually audited, but the MTProto scheme they put together has been found to have a number of flaws [1], and that hasn't changed. They also haven't really allowed third-party audit of their actual code, so there may or may not be extra bugs waiting to bite you.
This is true, but thankfully the security of Signal's double ratchet scheme means that you only need to verify your client to ensure the encryption remains intact. The server can't peek inside.
I found it interesting that direct calls (just calling one of your contacts) can be peer-to-peer if you enable it in the settings. They even have a unique code at the top of the screen (like 5 emojis) that you can verify match on the other participant's phone.
