sfpotter

16 ms·

181.
by sfpotter 1y ago
Not that it matters, but O(n log n) is often referred to as "quasilinear", but O(n^(1+eps)) is regarded as "superlinear" (and in fact grows faster than O(n log n) for any eps > 0).
182.
by sfpotter 1y ago
The article and its title are abundantly clear: this is NASA Stennis's first open source software. It's obviously not NASA's first piece of open source software. Are you saying that it's not Stennis's first open
183.
by sfpotter 1y ago
So they can work even better...?
184.
by sfpotter 1y ago
Have you used Kagi or Google recently? Kagi works way better.
185.
by sfpotter 1y ago
Expand your view of what constitutes excellence.
186.
by sfpotter 1y ago
I haven't seen these before. Thanks for the reference.
187.
by sfpotter 1y ago
Probably relevant: https://prng.di.unimi.it/ I use the PRNGs here for my own needs and they work great... at least as far as I can tell. :-)
188.
by sfpotter 1y ago
Sorry, let me clarify: In D, you can write Python-style slop code and/or C#-style slop code. Two very different styles of slop code, both of which can be written in D. Or even mixed and matched.
189.
by sfpotter 1y ago
D is really, really good. I hope it gets more love soon. D's focus on just getting shit done, lightning builds, QOL improvements all over the place, actually good modules, templates and metaprogramming that work, simpler more regular s
190.
by sfpotter 2y ago
Author probably just unaware of Binet's formula. But I'd agree with their assessment that there's unlikely to be a closed form for the (indeed, much more complicated!) quantity that they're interested in.
191.
by sfpotter 2y ago
This is the kind of thing I don't want to have to think about as a programmer. The compiler should just make it work.
192.
by sfpotter 2y ago
One thing I just can't understand is proactively using the :: syntax. It's sooo ugly with so much unnecessary line noise. Just use a single period! I think one of the best decisions D made was to get of -> and :: and just use
193.
by sfpotter 2y ago
And I'm just a simple applied mathematician. For me, the gradient is the vector that points in the direction of steepest increase of a scalar field, and the Jacobian (or indeed, "differential") is the linear map in the Taylor
194.
by sfpotter 2y ago
I find that with Paul Graham's writing, once you stop and think about what he's saying, almost nothing he says holds up under scrutiny. He's basically just a propagandist for VC-backed startups and a certain concomitant world
195.
by sfpotter 2y ago
I'm not a big Riemannian geometry buff, but I took a look at the definition in Do Carmo's book and it appears that "grad f" actually lies in TM, consistent with what I said above. Would love to learn more if I've go
196.
by sfpotter 2y ago
If you mean for how to use Taylor expansions and linear algebra, here's one I just made up. Let's say I want to differentiate tr(X^T X), tr is the trace, X is a matrix, and X^T is its transpose. Expand: tr((X + dX)^T (X + dX
197.
by sfpotter 2y ago
For m = 1, the gradient is a "vector" (a column vector). The Jacobian is a functional/a linear map (a row vector, dual to a column vector). They're transposes of one another. For m > 1, I would normally just define th
198.
by sfpotter 2y ago
If you want to get handy with matrix calculus, the real prerequisite is being comfortable with Taylor expansions and linear algebra. In a graduate numerical optimization class I took over a decade ago, the professor spent 10 minutes on the
199.
by sfpotter 2y ago
Have met lots of professors who are glorified managers who do not actual research taking sabbaticals for a fat paycheck. I doubt very much they do any real work during these sabbaticals either. If I had to guess, I would bet that these sabb
200.
by sfpotter 2y ago
The problem is the "desire to be relevant in an important technological shift". There's loads of worthwhile research to do that has nothing to do with LLMs. A lot of it will not or cannot be done in an industrial environment
201.
by sfpotter 2y ago
No need for any of this if you use D!
202.
by sfpotter 2y ago
People have tried applying Newton's method and other quasi-Newton methods to ML problems, and my understanding is that it isn't worth it for large problems. In ML, you care about getting a few digits for a very large, very nonline
203.
by sfpotter 2y ago
I'm in the middle of pursuing an autism diagnosis as an adult, and despite the obvious manifold benefits it would bring to my family's life, judgment from people like you is the #1 reason I'm hesitating going through with it.
204.
by sfpotter 2y ago
A lot of places will allow you to transfer an associates to a 4 year. Do the general education requirements at the CC, which is practically free and will be easier to work into your schedule. Then transfer directly into your major of choice
205.
by sfpotter 2y ago
Not burning your employees out is good. Totally unclear that working "80 hour weeks" is any more productive than a standard 40 hour week. Keeping employees around by slowing things down and making things more humane means you can
206.
by sfpotter 2y ago
Couldn't disagree more about what the author says about the style of the book. The book is conversational, but "tech bro"? Huh??? I've read Ousterhout's book cover to cover and don't feel like it was a waste of
207.
by sfpotter 2y ago
In my experience, there are usually enough differences in the downstream applications that it can be difficult to adapt these solvers to different use cases. These libraries often make subtle choices about the way data is laid out that can
208.
by sfpotter 2y ago
I think this varies a lot depending on what you're doing. I agree 100% that matplotlib is really slow and should be made to run as fast as humanly possible. I would add a (3) to my list above: optimize matplotlib! OTOH, at least for wh
209.
by sfpotter 2y ago
Every time I've needed to scrub through something in time like that, dumping a ton of frames to disk using imshow has been good enough. Usually, the limiting factor is how quickly I can generate a single frame. It's hard for me to
210.
by sfpotter 2y ago
Yeah, I agree with the general sentiment of what you're saying. Re: wtallis, I think my original complaint about EDA per se is indeed off the mark. Certainly creating a 20x20 grid of live-updating GPU plots and visualizations is a form

More ›