cschmidt

29 ms·

301.
by cschmidt 10y ago
Absolutely. If you live in a world of ascii, then maybe python 2 is adequate. But I always deal with unicode, and python 3 is a big benefit just for that.
302.
by cschmidt 10y ago
There is a popular science book on this topic as well: Why: A Guide to Finding and Using Causes https://www.amazon.com/Why-Guide-Finding-Using-Causes-ebook/... It is sitting on my bookshelf, but I haven't managed
303.
by cschmidt 10y ago
I have to quit the Mac app a few times a week, when it just decides it isn't going to play songs.
304.
by cschmidt 10y ago
I've worked in Waltham myself, so I'm not dissing on it. I just don't really buy that his statistic was 4 times better. In fact, doing it my way says that neither Waltham or Cambridge is statistically better than the other,
305.
by cschmidt 10y ago
This is a called a binomial process. We want to estimate the true proportion of exits for a given city. A city has x exits out of n trials, so one estimate of this proportion is just x/n. However, if you have a bunch of cities like t
306.
by cschmidt 10y ago
Yeah, while you can solve it that way, you're not getting any insight from the analysis. Logistic regression assumes all data are known exactly, while error bars are important for this case. Take a look at my comment to the parent for
307.
by cschmidt 10y ago
cont .... Washington, DC USA 20 52 0.253 0.385 0.530 Orem, UT USA 7 13 0.251 0.538 0.808 Dallas, TX USA 30 86 0.249 0.349 0.459 Mississauga, ON CAN
308.
by cschmidt 10y ago
Using logistic regression that way isn't going to do much, since each city has a single nonzero independent variable. I'd try something like this, stats wise. It is just 10 minutes of effort, but accounts for the different sample
309.
by cschmidt 10y ago
I suspect maybe you're showing (X exits)/(N funded companies). You can't just compare binomial proportions like that, between proportions of different N. You'll have lots of random fluctuation. If Waltham had a couple
310.
by cschmidt 10y ago
Source? And why is the column labeled standard deviation? Looks a bit weird that Waltham, MA is 4x Cambridge, MA.
311.
by cschmidt 10y ago
If two cultures isn't enough for you, there was an interesting blog post from a year ago called "The Three Cultures of Machine Learning": http://cs.jhu.edu/~jason/tutorials/ml-simplex.html
312.
by cschmidt 10y ago
The talk is here: http://www.securitytube.net/video/14656?utm_source=feedburne... He makes the point I was referring to at 12:53. The question starts about 11:45.
313.
by cschmidt 10y ago
I'm going by memory, but I'm pretty sure he made that point (that I'm paraphrasing, of course). Anyone else here his talk that would care to comment?
314.
by cschmidt 10y ago
Schneier made an interesting point at his talk at DEFCON last year. He suggested people use encryption from the big players, rather than more niche crypto products. He said he thought the NSA etc. monitor users of those niche software clos
315.
by cschmidt 10y ago
I think Sketch has become the default UI prototype tool for many people. Our designers use it created mock ups of web and mobile pages.
316.
The “hobbits” of human evolution are even older than we thought(economist.com)
1 points by cschmidt 10y ago|0 comments
317.
by cschmidt 10y ago
Also "Model Building for Mathematical Programming" is almost unique in that it talks about formulating models, rather than the algorithms you use to solve them. That tends to be what an OR person spends the most time doing: http:
318.
by cschmidt 10y ago
I don't know, I haven't had the same experience as you. I did a lot of integer programming in grad school. While you're right that integer programming doesn't scale well enough to solve many real world problems [1], OR a
319.
by cschmidt 10y ago
FWIW the automatic update worked for me.
320.
by cschmidt 10y ago
Like this http://www.phdcomics.com/comics/archive.php?comicid=1174
321.
by cschmidt 10y ago
I think password reuse is the big deal. Lots of people use the same passwords on more important accounts, which they would mind losing.
322.
by cschmidt 10y ago
This seems very similar to the "Fathom Neural Compute Stick" from Movidius: http://www.movidius.com/solutions/machine-vision-algorithms/... TensorFlow on a chip....
323.
by cschmidt 10y ago
Numba seems to fall back to regular slow python code when it can't handle things, without any warning. I couldn't find a way to tell if a routine was compiling to fast code, or dropping back to slow code. (As I say, I didn'
324.
by cschmidt 10y ago
In my (somewhat limited) experience, Numba works great until it doesn't. Loops on numpy arrays work great. Other stuff becomes super slow, without any real way to understand what is happening. Julia sees more understandable.
325.
by cschmidt 11y ago
I'm fairly sure that only applies to the iPhone 5, and the FBI is interested in an iPhone 4, which doesn't have the Secure Enclave.
326.
by cschmidt 11y ago
A uniform prior would be one example. That's just saying "I've got no idea what the value is".
327.
by cschmidt 11y ago
Not that I'm aware of. (I've read a fair bit on the topic.)
328.
by cschmidt 11y ago
It seems to have been often studied by Economists in the area of game theory. This very nice book is by two of them: http://www.amazon.com/Two-Sided-Matching-Game-Theoretic-Econ... To be fair, it is also studied by compute
329.
by cschmidt 11y ago
Well, 28mm is supposedly equivalent to the iPhone camera, so lots of people do manage to take photos at that angle.
330.
by cschmidt 11y ago
See also Feynman’s Derivation of the Schrödinger Equation discussed a couple of weeks ago... https://news.ycombinator.com/item?id=10974803

More ›