imh

89 ms·

721.
by imh 11y ago
Maybe the difference is that we all learned calculus in classes from experts, while the only way for most people to learn haskell is to pick it up themselves.
722.
by imh 11y ago
Before everyone panics, there's an important note on page 70: >The LSTM model can also successfully classify signals with an accuracy of 19% when the dataset used for training and logging are originated from two different keypads. S
723.
by imh 11y ago
Automated response questions are so disrespectful. "We won't take the time to see if we're in any way interested, and won't unless you devote your own time."
724.
by imh 11y ago
>And by extension, it means that in every interview cycle, some portion of interviewees are losing interest in joining your company just because they didn’t think they did well, despite the fact that they actually did. I hate the asserti
725.
by imh 11y ago
Math? No. Math in the form of a research community which has built up centuries of work for each other to further build on? Absolutely.
726.
by imh 11y ago
To be fair, there's not much to read in the article. There's fluff about Marcus as a person and some tiny amount of amazingly vague talk about what he's doing.
727.
by imh 11y ago
If you're really concerned and want to change that, I had great success with one of those books of practice interview questions. It sounds cheesy, but (obvious in retrospect) simply practicing the reasoning patterns tested in that kind
728.
by imh 11y ago
This is great reporting. The reporter explicitly pointed out it isn't necessarily a causal effect instead of exploiting the misunderstanding that it might be. Then she succinctly explains the findings and a few potential interpretation
729.
by imh 11y ago
>How might Gates spend its money more wisely? McGoey offers little guidance on this. This is the crux of the article to me. These problems don't have known solutions. The options are to do expensive research/experiments or just
730.
by imh 11y ago
I had a similar problem and couldn't even get them to replace the damned thing. They said it was 'normal wear and tear' or something. It was absolutely infuriating that they called a clear fire hazard normal. So then I got an
731.
by imh 11y ago
Reading this article made me realize I have no idea what "gender" is. It used to be synonymous with what parts and chromosomes you have, where there are no questions of "I identify as XYZ," but now what the hell is it re
732.
by imh 11y ago
I guess I hate the idea of a "statistical philosophy." It's math for god's sake. Given these axioms, these other things are true. The point about teaching relativistic versus newtonian mechanics is totally different beca
733.
by imh 11y ago
For a more thorough intro, I can't recommend anything better than this http://arxiv.org/abs/1506.00019
734.
by imh 11y ago
I really never understood the divide. They are all just tools in our toolbox. It feels like arguing screwdrivers versus hammers. They may not do the same thing, but both can be useful!
735.
by imh 11y ago
I think OP addresses that concern: >That does not mean you shouldn’t show code on stage. Please, by all means, put code up there on your slide and talk about it. Just don’t write it in real time. Live coding is lazy; it is a shortcut for
736.
by imh 11y ago
That's why I wouldn't worry about how it performs in general, but in edge cases. The question isn't whether it's memorizing the whole dataset, but whether it's "memorizing" any particular points it shouldn
737.
by imh 11y ago
I'm talking about a slightly different problem. I'm not suggesting that you might accidentally click to send a reply you didn't want to share, so you reviewing it is beside the point. I'm suggesting that by mining all ou
738.
by imh 11y ago
I appreciate the privacy standards they used (no humans reading your email to develop this), but am concerned that it's not enough. As I understand with language models, overfitting takes the form of returning a sequence of words seen
739.
by imh 11y ago
Pick up this book. It's fantastic. https://mitpress.mit.edu/books/probabilistic-graphical-model...
740.
by imh 11y ago
On Firefox 41.0.1 on OSX Yosemite, neither the space bar nor the arrow keys work. No extensions.
741.
by imh 11y ago
But then why do we have such a strong culture around hiring the "10x" engineer?
742.
by imh 11y ago
>You suck stupid shit. FUCK YOUUU?? Same here, but I was going for 0.
743.
by imh 11y ago
It's strange to read this story immediately after this other one https://news.ycombinator.com/item?id=10367253 I'd be afraid to see the correlation between who uses those kinds of methods and who is at the frontie
744.
by imh 11y ago
Yeah, medicine is nuts. The whole field of medical imaging is magic. fMRI still blows my mind. But everything medicine is swept together as one advancement, just as everything computing related is.
745.
by imh 11y ago
Doesn't it invalidate a patent if someone else was already doing it?
746.
by imh 11y ago
The top comment literally said "given books and free time." Not exactly moving goalposts.
747.
by imh 11y ago
I distinctly remember this one as an offender in unnecessary language. http://blogs.law.harvard.edu/pamphlet/2015/09/28/whence-func...
748.
by imh 11y ago
But that takes ages to train!
749.
by imh 11y ago
How about, "Could the U.S. Ever Adopt an Eight-Hour Workday?" I'm not talking about including off-hour emailing and such. I'm talking about a plain old 9-5. I'm probably biased by living in SF, but I don't know
750.
by imh 11y ago
I've come to the conclusion that more flexibility often leads to a race to the bottom. Like the "unlimited vacation" that's so popular now. Same with flexible hours, I'd guess. Anyone know how to combat that aspect

More ›