xscott

47 ms·

601.
by xscott 6y ago
Let's take an example in your favor and say you've got 16 integer local variables per stack frame and you're recursing 1000 (!!!) times over some really unbalanced tree or such. You saved 64 kilobytes... Ok, now let's s
602.
by xscott 6y ago
I wish new platforms would embrace 64 bit as the default. C and C++ for all common platforms use 32 bit literals and prefer 32 bit operations (char * short => int). Rust made a similar mistake. Java arrays use 32 bit subscripts, etc..
603.
by xscott 6y ago
Definitely. It's also a really good (if painful) way to decide what things are actually important to the delivery.
604.
by xscott 6y ago
Fear. Nothing got me motivated to finish something like the fear of failing and the shame that I'd feel as a result. Not my favorite technique, and probably not what you're looking for, but I'm guessing half of my good acco
605.
by xscott 6y ago
Somewhere around 1994, I used to dial into the university's terminal server using a Wyse 370 and a 9600 baud modem. (Note, this was not PPP or SLIP providing TCP/IP. You had to type the atdt modem commands directly...) This par
606.
by xscott 6y ago
The frightening one for me was R0 == 1: Basically, we'd need a positive or negative infinite percentage of people to be infected depending on which direction you approach it from.
607.
by xscott 6y ago
I was watching this video earlier tonight: https://youtu.be/Vat-PKdQ13A?t=30 That looked like some really sloppy lab technique, and I hope that's not how it is really done in practice. If real tests are done by hand l
608.
by xscott 6y ago
No worries - it was clearly just a typo :-)
609.
by xscott 6y ago
> First, `1/1-R_0`. This is exactly what Wikipedia says: https://en.wikipedia.org/wiki/Herd_immunity#Mechanics I have no expertise in this area, but Wikipedia does not say: 1/(1 - R_0) It's:
610.
by xscott 6y ago
Nelder Mead is sometimes effective, and it is pretty to visualize, but it gets stuck in some trivial cases. It can fail to find the bottom of a low dimensional parabola, for instance. CMA-ES is worth looking at, and has a similar idea but
611.
by xscott 6y ago
There are "limited memory" algorithms which use low rank approximations. https://en.wikipedia.org/wiki/Limited-memory_BFGS
612.
by xscott 6y ago
And yet, as a human, that's probably where you should put your retirement savings.
613.
by xscott 6y ago
I'm no expert, but this is the explanation I heard, and it makes sense to me: You're judging the market by looking at the indexes. The DOW is 30 large companies. The S&P 500 and Russel 3000 are weighted proportionally to how
614.
by xscott 7y ago
Do you have any links for the prior art? I'm sincerely interested, thank you if you do. I'd also be interested in any articles (or even blog posts) that describe Rust's process for that from a compiler writer's point o
615.
by xscott 7y ago
Spot on, and the difference between a tracing JIT vs a method JIT can be night and day too.
616.
by xscott 7y ago
> millions of engineer-hours, and the result was better Unicode support and a handful of minor features Yes, and I wish the better Unicode support had been implemented similarly to how Go does it - one string type, and you use it to hold
617.
by xscott 7y ago
As soon as you do any operations (addition, subtraction, etc...) on those views, you're going to get temporary arrays. For instance, your Mandelbrot example doesn't even use views and it creates two temporaries the size of the ent
618.
by xscott 7y ago
Yeah, that's the thing with silly examples (like my Mandelbrot program). For real code, I frequently need to write low level loops which aren't easily expressed as parallel operations. If numpy doesn't have it, or if you ca
619.
by xscott 7y ago
Honestly, I think Python's C API exposing so much of the internals of the implementation is the real problem. You can basically see every pointer in every struct, including tons of things you shouldn't need. Large packages inevi
620.
by xscott 7y ago
For those with Python 2 baselines that aren't (or will never be) ported, PyPy will support version 2 basically forever [0]. PyPy supports a huge number (but not all) of Python packages, including numpy [1]. Moreover, PyPy is signific
621.
by xscott 7y ago
I suspect your sentiment is why so many people accept confident predictions from blowhards who don't actually know/care how complicated the real world is. Maybe it's more comforting to listen to people who pretend to have th
622.
by xscott 7y ago
> Truly negative sides become more noticeable much later when it’s too late. I slept poorly for years, so I suspect I can guess some of these, but I'm curious which side effects you meant. It's hard for me to know what negativ
623.
by xscott 7y ago
I've looked at your other posts, and you seem like a reasonable enough person. I don't have a horrible beef with you, but in general I really despise it when people do the chicken-little doom and gloom thing and project gazillion
624.
by xscott 7y ago
> For clarity, I'm not predicting. Why are you stating this now, and not when you wrote the 100,000 number above? If that wasn't a prediction, what was it?
625.
by xscott 7y ago
And how long do you predict a 5.5 day doubling period to continue?
626.
by xscott 7y ago
Not sure if you're still checking back, but at day 15 the US deaths are at 16k where his prediction says 25k.
627.
by xscott 7y ago
> without studying the problem extremely carefully, you'll never know[...] It seems likely to me that this problem will get studied extremely carefully.
628.
by xscott 7y ago
I really appreciate this answer, and as a lay person, it helps me put some pieces together in my mind. I suspect there are at least four reasons for testing, and they all benefit from better results, but two of these need good quality unbi
629.
by xscott 7y ago
The fact he had to make a rule for her not to belittle him in front of the children makes me think I wouldn't want to be married to his wife either. We're only seeing one half of the picture here, but she doesn't sound like
630.
by xscott 7y ago
I really don't understand your intent. Are you trying to say more people have died so far from Covid-19 than die from a bad flu or pneumonia season?

More ›