clircle

49 ms·

601.
by clircle 6y ago
The first chart is kind of shocking to me. Look at the 95th percentile line. Is it really the case that in 2018, 1 in 20 Americans was making > ~190k? Are these household incomes?
602.
by clircle 6y ago
I think Emacs is nearly perfect. I use it everyday at work for different tasks. But I can't seriously recommend it to anyone. I'm looking at my configuration file now. It's 1300 lines of elisp. It took me 4 years to make it b
603.
by clircle 6y ago
One of the top three reasons I use Ubuntu: no fiddling with fonts. They look good out of the box. There's enough things to fiddle with on linux.
604.
by clircle 6y ago
I don't think I've played a game as good as XCOM 2 since it came out.
605.
by clircle 6y ago
Dipshitize.com Awesome.
606.
Vim Keybindings for MS Word(github.com)
2 points by clircle 6y ago|1 comments
607.
by clircle 6y ago
Nope. been using Google Reader and NewsBlur for 15 years to avoid this foolery.
608.
by clircle 6y ago
I was somewhat disappointed that the movie didn't much discuss the downsides of regulating social media, e.g. censorship, slippery slopes. like other commenters say, I didn't learn anything new, but I'm glad this kind of docu
609.
by clircle 6y ago
Would be interesting to put a confidence "interval" on the predicted face. Perhaps show a range of faces that are consistent with the model. That might help to alleviate some concerns.
610.
by clircle 6y ago
> Needing to download a million random packages from grad students of the internet is a recipe for disaster. This is baloney. You don't need to rely on a million R packages. You really only need a good plotting library (the built-in
611.
by clircle 6y ago
I think there is something left out of the post, but it's hard to put my finger on it. In a vacuum, 71% does sound like a relatively low amount of evidence in favor of Biden, but when you put it next to most elections, which are only w
612.
by clircle 6y ago
Emacs isn't better than Rstudio, Pycharm, TexMaker, and the other programs I would be using otherwise, but it is more comfortable and keeps me happy.
613.
by clircle 6y ago
I was a freshman CS student in 2007 and this was the language that I learned in CS 101. I wonder how common that was. Frankly, that was my first exposure to programming, and I thought the language was strange enough and the prof wasn't
614.
by clircle 6y ago
Looking through the repo, it's a handful of elisp files, so you can pick-and-choose by deleting the stuff you don't want.
615.
by clircle 6y ago
I can get by pretty well by only using org-capture and org-agenda and ignoring everything else about Org mode.
616.
by clircle 6y ago
Compared to the actual problem, public health and the economy, this request looks extremely childish
617.
by clircle 6y ago
Awesome! Great list of updates. But what is early init?
618.
by clircle 6y ago
I'm been an Emacs+Evil user for many years, but I would like to switch to Vim-- mainly because I use Windows 10 most of the time now and Emacs can have some input latency. But I'm old now and can't really imagine life without
619.
by clircle 6y ago
I don't quite follow the mouse hating. And I'm a heavy Emacs/vim user! The mouse is great. Sometimes it's the best way to select text or follow links. I find it useful to bind some common hotkeys to mouse buttons: I bind
620.
by clircle 6y ago
MGFs do not determine distributions (The Student's t distribution does not have an MGF). CDFs and characteristic functions determine distributions.
621.
by clircle 6y ago
Looks like a very nice list of topics. Glad to see GLMs in there -- I think if more folks had exposure to GLMs, then we could dispel the myth that statistics is a bag of tricks. GLMs are a very awesome expansion of regression that connect m
622.
by clircle 6y ago
Long live Thunderbird.
623.
by clircle 6y ago
Machine learning
624.
by clircle 6y ago
Data driven priors are more common now. Sometimes this is called empirical Bayes. Also priors can change after model checking. These are data driven priors, but not EB. The prior is only negligible in infinite sample sizes. It could still h
625.
by clircle 6y ago
The source of the prior distribution affects analysis and whether the Likelihood Principle (LP) is obeyed. If the prior is pure/subjective, LP applies. If the source is not subjective LP does not apply, and with it goes some of the nic
626.
by clircle 6y ago
Another issue: in some machine learning models, like image classifiers, race is not an a priori covariate or group in the data, but is only scrutinized a posteriori. I'm not sure what you mean when you said the model defaults to white
627.
by clircle 6y ago
What does it mean that an AI is unbiased? Would it mean that the AI classifies race correctly on average? Or something else -- e.g. zero misclassification error for a specific race? With bias, it's helpful to define what property is de
628.
by clircle 6y ago
Ah, the old "regression from scratch" post that is mandatory for all blogs
629.
by clircle 6y ago
On the last page. > If taking the geometric mean of all 172 tests on this AMD Ryzen Threadripper 3970X workstation, using WSL2 yielded about 87% of the performance of running bare metal Ubuntu 20.04 LTS on the same hardware. So many ques
630.
by clircle 6y ago
It's also fun to define the cantor function as the CDF of a distribution. The distribution has neither a pdf nor a pmf (it is singular) https://en.wikipedia.org/wiki/Cantor_distribution

More ›