mruts

79 ms·

391.
by mruts 7y ago
I’ve wanted to make a keyboard, but I’m not sure it’s possible to make something materially better than the kinesis advantage. I switched to it 10 years after my hands would feel numb after typing. I think I might have some problem other pe
392.
by mruts 7y ago
I always thought it related to continuous variables and the highly useful concept of the infinitesimal. But maybe that definition works as well.
393.
by mruts 7y ago
Lambda calculus nor regular calculus are going to help you write a compiler. And certainly regular calculus isn’t going to help you design a language...
394.
by mruts 7y ago
I dunno, the people in that picture don't exactly look... cool. I love D&D but have had a hard time finding a group of people who play who I would actually want to be around. Maybe that's a little harsh though..
395.
by mruts 7y ago
Would you prefer book to market? That's what Fama and French used. All of these ratios are related to the value factor, I didn't say you need/should use on or another.
396.
by mruts 7y ago
Sorry, I mean Big minus Small. My bad! I guess I started associating high with big and low with small!
397.
by mruts 7y ago
Certainly the proliferation of alternative data (online news sentiment, satellite imaginary, GPS location data, etc) has made finding those cheap harder to find because they are probably more fairly priced. Buffet has said that he switched
398.
by mruts 7y ago
Yeah that's a pretty good guess. Growth generally means a company is expensive relative to it's earnings, profits, or book value. Ratios like P/E (price to earnings), and P/B (price to book) can be used to identify/
399.
by mruts 7y ago
Current research from AQR seems to suggest that HmL isn't actually a persistent factor, just a behavioral bias without any economic basis (anymore?). Also current research by AQR suggests that that value is more structural, and is a re
400.
by mruts 7y ago
Well, you would expect the markets to reward the people who didn't have these biases, and hurt the people who still had them. The people with the bias will get poorer and poorer, and the people without it will get richer and richer. Th
401.
by mruts 7y ago
There are two rational justification for the (seemingly) persistent value factor: 1. It’s just a risk premium. The efficient market hypothesis says that all risk adjusted returns are the same. i.e there is only one Sharpe ratio. Or put anot
402.
by mruts 7y ago
I don’t think static typing is really related to a repl. There are many statically types languages that provide great repls: Haskell, OCaml, Scala, etc.
403.
by mruts 7y ago
It’s true that they both work, but I think the floating nature of the 1-indexing has some very desirable properties. If we wanted to sample from a distribution, the stride would relate to 1/n x 100 percent sampling. With 0-indexing we
404.
by mruts 7y ago
Founder’s effects are really interesting. One might hypothesis that this led to the descendants of these colonists to have stronger bones. You see this effect in many populations that originated from few individuals. A good example is Polyn
405.
by mruts 7y ago
Check out the book “Range: Why Generalists Triumph in a Specialized World” by David Epstein for why generalists actually have better outcomes than specialists, even at specialized tasks. Intuitively, I think this makes sense. Imagine skill
406.
by mruts 7y ago
The classic example is getting the last element of an array. With 1-based indexing the length of the array is the index of the last element. It has a nice symmetry to it. Also I find it elegant that for 1-indexing that the start and end val
407.
by mruts 7y ago
Julia has static typing. I wouldn’t go that far and say Python is suitable for large programs. It’s clearly not. Working on a large python code base is hell.
408.
by mruts 7y ago
I’m not sure what Python’s goals are to be honest. It seems to me that the language is outclassed in every way by better, more consistent, more powerful, and more performant languages. Python programmers seem content implementing the same t
409.
by mruts 7y ago
Julia is everything python could have been, and much more. I'm stuck with python right now as a lot of people in the data science/ML community are, but it's becoming increasingly viable to use Julia for "real" work.
410.
by mruts 7y ago
I mean, we can assume the downtime variance follows a normal distribution. It should pretty easy to calculate P<.05 with just a little bit of data.
411.
by mruts 7y ago
The cost of defecting is low. Why do they think anyone will actually change their future decisions because of this “resolution”.
412.
by mruts 7y ago
Do they provide in and out services?
413.
by mruts 7y ago
I mean, is this even true? I love programming for example but love it a lot more when I’m being paid a lot of money.
414.
by mruts 7y ago
Not to be harsh, but, like, that is your problem. I, for example, used to be addicted to heroin. If opiates had never been discovered, I guess I probably would have never had a problem with it. But that's a weak excuse. I'm glad I
415.
by mruts 7y ago
Has anyone considered that the ability for people to be easily distracted reveals a weakness in themselves? i.e we all wish we were more bored so we could discover our "true" selves. If you're distracted all the time by twitt
416.
by mruts 7y ago
Instead of creating a half-assed extension language, a better solution is to have one language do both. Like Common Lisp for example.
417.
by mruts 7y ago
Things like this are already extremely rare. The number of kids that die from cancer has to be a very very large sigma event.
418.
by mruts 7y ago
People don’t want their food to come out fast. They want to sit, drink a beer, talm, etc. Unless it’s fast food, people go to restuarants to have a 1-3 hour experience, not to get calories as quickly ad possible.
419.
by mruts 7y ago
This is why I switched to Arch. Debian is just too conservative.
420.
by mruts 7y ago
word2vec simply is a neural network with a hidden layer of n perceptrons. The number they use in the paper is 300 I believe. Thus, this is the dimensionality of each word. To train the model, we generate feature-label pairs by taking a word

More ›