brendano

9 ms·

31.
by brendano 13y ago
Congrats, folks!
32.
by brendano 14y ago
eh, it's just binning :)
33.
by brendano 14y ago
I don't understand why this is "wildly unprincipled." Doesn't it have all the guarantees that Weinberger and Shi et al papers describe?
34.
by brendano 14y ago
But anybody that writes off an IDE without fully understanding just how powerful the two features I mentioned above can be when dealing with multi-million loc codebases is being narrow minded. Yes! This is the crux of the confusion. Wor
35.
by brendano 14y ago
The article makes it sound like a "whiz kid loner" sort of company, but the summly.com website suggests there's a lot more to it than is implied by this article. The About page shows a pretty senior, experienced team (decades of engineerin
36.
by brendano 14y ago
legit's "git branches" reminds me of git-wtf, which has saved me many times: http://git-wt-commit.rubyforge.org/
37.
by brendano 14y ago
Yes, i did policy debate in high school in Minnesota. It was really influential in how I think about things, especially analytic/critical thinking.
38.
by brendano 14y ago
Thanks in part to the popularity of his books, movie, and speeches, Kurzweil now knows pretty much every AI researcher in the planet Um... More important question: how many AI researchers respect the last 20 years of his work?
39.
by brendano 14y ago
> > His contributions to science are negligible if not entirely detrimental. > I'm not a linguistics expert, but from my vantage point as a computer scientist, it would be hard to conclude that. Chomsky essentially founded the f
40.
by brendano 14y ago
For the questions about relative ranking of contributions, sure. But the entropy results seem reasonable, since it might average over a large number per-developer preferences.
41.
by brendano 14y ago
Following up on blahedo's comment and the questions about what the heck their p-values mean -- This is a nice example that you can get statistical significance for small effects, if your sample is big enough. Their p-values are explained v
42.
by brendano 14y ago
This part was glossed over a lot, I had no idea what they were doing either. On your #1, you could do a comparison of (% Agree) for one font versus another. For #2, the weights on the scale don't matter, you can just compare the means betw
43.
by brendano 14y ago
Standard null hypothesis testing -- including what they're talking about here -- focuses on P(results | random noise is active) You're talking about P(the thing we care about | results) The former is more like a simple sanity check. If ran
44.
by brendano 14y ago
I love posting speaker notes too. I did it for this one a bit ago; I think the notes helped people who were reading it. http://brenocon.com/IE%20lecture.pdf
45.
by brendano 14y ago
I've been told (by people from Inktomi who then worked at Yahoo), that yes, the Inktomi technology became the main Yahoo search engine. (How much it changed or not in the process, I don't know)
46.
by brendano 14y ago
For NLP applications, which I think Aria's article is mostly concerned with, non-linear kernelized classifiers are often little better than linear ones. I think that's one part of the recent interest in SGD-style training algorithms (they
47.
by brendano 14y ago
Careful with the Collobert ICML-2008 paper. It has a very negative reputation among NLP researchers who actually know the area, just for its setup/evaluation. If you're interested in the methods (which I think are interesting), that group
48.
by brendano 15y ago
PHP has no app servers to go bonkers.
49.
by brendano 15y ago
PHP is much better than CGI, because it doesn't have to load an interpreter and all libraries per request -- they were loaded at server startup time. This is why all serious web frameworks don't use CGI. But PHP avoids the administrative/
50.
by brendano 15y ago
...that's not really inherent in PHP. You can configure Python to be just as easy, and some webhosts do this... It's very rare to find webhosts that configure Python, Ruby, Node or the rest to be as easy as "upload a file, click reload."
51.
by brendano 15y ago
Python - first commit, 22 years ago https://github.com/python-git/python/commit/463d099321a4572e... (Current first page: https://github.com/python-git/python/commits/master?page=117... ) Perl - first commit, 24 years ago Amazing title:
52.
by brendano 15y ago
You can turn his argument around. Why should the needs of a business search be more important than historical search or journalism search? Maybe, on average, users more often want to learn about the historical accident than the resort. O
53.
by brendano 15y ago
Yes! -- translation into a strictly delimited format is key.
54.
by brendano 15y ago
Very true. I spent quite a while trying to learn subprocess, then gave up and just use os.popen() now. It's a shame -- there are certain subprocess features I really would like to have, but it's too hard to remember how to use it.
55.
by brendano 15y ago
Yes. For example, the Facebook numbers are always number of people who've logged in at least once in the last 28 or 30 days (or something) -- I'm sure the G+ people are computing this internally. The fact they don't report it implies it m
56.
by brendano 15y ago
if you don't have "sort -h" (mine doesn't do it), you can also use "du -m", which gives file sizes in megabytes -- pretty reasonable. (then "sort -n")
57.
by brendano 15y ago
awk (and to a lesser extent sed) is cleaner and easier to remember.
58.
by brendano 15y ago
Matlab, R, NumPy: http://mathesaurus.sourceforge.net/
59.
by brendano 15y ago
I had to type it in, not copy-and-paste, and got it.
60.
by brendano 15y ago
There used to be an earlier era of machine learning that wasn't as statistical. Ng, and most other current ML researchers, now heavily draw on mainstream statistics. It really does make sense to do logistic regression as the foundation fo

More ›