gsteinb88

6 ms·

1.
by gsteinb88 5y ago
Just FYI, Amazon and most other providers offer instances that go up to many TB of ram — the largest listed on EC2 right now is 24TB. Depending on your workload, you may get far better throughput not worrying about distributing the work and
2.
by gsteinb88 5y ago
If you can, working with the logarithms of the intermediate large or small values is one way around the issue One example talking about this here: http://aosabook.org/en/500L/a-rejection-sampler.html#the-mul...
3.
by gsteinb88 5y ago
Context for point 5: - MIT sued Gehry's firm: https://thetech.com/2007/11/09/lawsuit-v127-n53 - Eventual settlement: https://thetech.com/2010/03/19/statasuit-v130-n14
4.
by gsteinb88 6y ago
Carrier freezout actually makes this non-workable -- there's a limit to how cold you can make CMOS devices before they stop functioning. To say nothing of the specific heat of liquid helium, which is miniscule compared to LN2
5.
by gsteinb88 7y ago
That is not the free file page, and the fact that it tricks you into thinking it is is exactly the point: https://www.propublica.org/article/turbotax-just-tricked-you...
6.
by gsteinb88 7y ago
Looks like matplotlib to me (assuming you mean the top window, not the terminal or what looks like a browser behind)
7.
by gsteinb88 8y ago
oops: [1] https://www.analog.com/en/products/ad9361.html
8.
by gsteinb88 8y ago
Yeah, the BladeRF uses an AD9361[1] as the transceiver, which features a pair of ADCs per channel, of which there are two (so, a total of 4 ADCs and 4 DACs in the package). It's got a bunch of other convenient features like built-in PL
9.
by gsteinb88 8y ago
C-o for me, for exactly the same reason
10.
by gsteinb88 8y ago
The latter can be accomplished with inverted flame graphs (sometimes called icicle graphs) which show the call stack inverted
11.
by gsteinb88 8y ago
Most of the rocket scientists I know also exercise
12.
by gsteinb88 8y ago
Surface mount device: https://en.wikipedia.org/wiki/Surface-mount_technology
13.
by gsteinb88 9y ago
How about pointing to several hundred years of systemic slavery and oppression without any meaningful reparation efforts?
14.
by gsteinb88 9y ago
https://www.apple.com/business/docs/iOS_Security_Guide.pdf https://images.apple.com/business/docs/FaceID_Security_Guide... Seems pretty likely there will be a whitepaper and some confere
15.
by gsteinb88 9y ago
So long as all frequencies you're trying to represent are below the Nyquist frequency, there's still only one band-limited reconstruction of the samples. So yes, it's still possible to reproduce everything accurately so long
16.
by gsteinb88 9y ago
If you're going to do any random processes, the graduate course (and corresponding textbook) from EECS at MIT is great: https://ocw.mit.edu/courses/electrical-engineering-and-compu... Textbook: https://
17.
by gsteinb88 9y ago
https://en.wiktionary.org/wiki/bury_the_lede OP is implying the "real story" (i.e. the most important part of the article) is not what the article begins with
18.
by gsteinb88 10y ago
I paid $12 on Amazon for 6 Travalo lightning cables on Amazon and I couldn't be happier with them. Lightning interface is legit (no notifications about unsupported lightning connector) and they're more robust than the official App
19.
by gsteinb88 10y ago
FFTW is one of the seminal examples of the importance of cache awareness, if I remember my history correctly. As I understand it, the reason the fftw authors (Matteo Frigo and Steven G. Johnson) won the Wilkinson prize was essentially becau
20.
by gsteinb88 10y ago
Same goes for me when doing CAD and having to constantly convert between standard & metric -- 254, 508, etc show up in numbers a lot and it gives you a nice hint that it's a "clean" value in the other system. Super useful
21.
by gsteinb88 10y ago
[1] turns out to be more subtle than people originally thought. Even assuming you can make loophole free bell tests (hey, it's been done!), the best anyone has been able to show this does -- even theoretically -- is randomness /am
22.
by gsteinb88 10y ago
"Utterly bizarre" captures a lot of my feelings about science funding and the graduate student research mill, yeah
23.
by gsteinb88 10y ago
Not universally true -- I know students on NDSEG fellowships (i.e. DoD) receive a 1099. I can't speak to the other fellowships though, and I don't get any deduction/withholding done for me, so I have to pay estimated quarterl
24.
by gsteinb88 10y ago
Wow, I've literally never heard of a PhD where you were only expected to work 20 hours/week. I'd say 50-60 is closer to the norm, especially once you're done with your coursework (usually a year or two into a 5-7 year lo
25.
by gsteinb88 10y ago
Well, the entire human race is inadequate to run a marathon in less than 2h so far...
26.
by gsteinb88 11y ago
Previous discussion (~5 years ago): https://news.ycombinator.com/item?id=2477667
27.
by gsteinb88 11y ago
Just worked for me
28.
by gsteinb88 11y ago
That was a fantastic post, thank you for sharing. Ironically, while affirming the view I held above, it also makes me think I should -- maybe -- give them another chance now that I'm well along in graduate school as opposed to a strugg
29.
by gsteinb88 11y ago
I'm not sure that Feynman is the best example here. In fact, in my experience, he's actually more of an example of the problem than the solution. While Feynman does a great job of explaining concepts if you never need to use them,
30.
by gsteinb88 11y ago
Nifty, thanks. I've been lucky enough to be able to default to microcontrollers with usb-serial built in recently, but that's lookkng to be short lived...

More ›