cbkeller

27 ms·

181.
by cbkeller 6y ago
Here is the full text [1] of Whewell's review of this book, by the way, in which the term scientist appears to have first appeared in print [2,3] [1] https://blogs.kent.ac.uk/sciencecomma/files/2015/04&
182.
by cbkeller 6y ago
The word does indeed seem to be attributable to Whewell, with the OED attributing the first printed use of the word to an 1834 article by Whewell in the Quarterly Review, in which the discussion you allude to regarding "philosophy&quot
183.
by cbkeller 6y ago
Her most famous book On the Connexion of the Physical Sciences , which is mentioned in the article, is available in full on Archive.org and looks to be a rather interesting read [1]. [1] https://archive.org/details/onc
184.
by cbkeller 6y ago
Haven't found anything yet myself, but would love to know -- that looks quite interesting!
185.
by cbkeller 6y ago
Hah, you were not kidding about the names being unintuitive. For anyone else interested there is a large list starting here [1]. The in-place versions will all end with `!`. [1] https://docs.julialang.org/en/v1/std
186.
by cbkeller 6y ago
I understand you are frustrated, however, please remember > Please don't comment on whether someone read an article. "Did you even read the article? It mentions that" can be shortened to "The article mentions that.&qu
187.
by cbkeller 6y ago
This is fairly far from my field, but as I understand it, that would be a hidden variable interpretation of QM [1], and specifically in that analogy a local hidden variable theory. That's what Einstein himself wanted. There is a fa
188.
by cbkeller 6y ago
They're soils, but they're special soils [1] that can harden to a brick-like consistency (called a "duricrust") upon being dried out and exposed to the air [2,3]. The unusual durability of such soils has made them a ta
189.
by cbkeller 6y ago
> precompiled and even distributed There is work on this. This can be done with https://julialang.github.io/PackageCompiler.jl/dev/ , but so far only to a rather large binary / “bundle” IIUC, In principle i
190.
by cbkeller 6y ago
That’s why Pandas read_csv is what they benchmarked
191.
by cbkeller 6y ago
Geologist here! A fun fact that has received some previous discussion [1] here: Some of these same missing megafauna (likely giant ground sloths) dug an impressive network of several-meter-diameter tunnels through the unusually stable, iron
192.
by cbkeller 6y ago
I was wondering about this latter point as well, and apparently this is exactly the goal if one could implement reversible computing at the hardware level [1]: > One of the earliest proposals to reduce heat from lost bits was to make com
193.
by cbkeller 6y ago
I think it's a variant of the latter -- IIUC no information can be "lost" in reversible computing, so the output of a reversible hash function might be both the hash and some number of other outputs, all of which you would
194.
by cbkeller 6y ago
I'm probably just missing something, but did you mean to say "fast derivatives of scalar functions" by any chance? I.e., that you can use the free inverses to get at the derivatives in some way?
195.
by cbkeller 6y ago
For someone who's been following this only tangentially, how similar is the interface and implementation to what's used in Grant's https://github.com/3b1b/manim ?
196.
by cbkeller 6y ago
Well, the promise of overhead-free multiple dispatch is that you don't have to choose between a general solution with poor performance and a specific solution with good performance. You can have all the specific solutions you need in o
197.
by cbkeller 6y ago
Doesn't that validate the applicability of the benchmark then? GP said: > A typical user will rarely ever use the built in CSV library in Python as a criticism of the benchmark, but if a typical user does use stdlib’s csv after all,
198.
by cbkeller 6y ago
I mean, A lot of people did! HPC, DiffEq, etc.* And to some degree it's these communities who are now adopting Julia, interestingly enough. *Elaborated in my other comment on GP https://news.ycombinator.com/item?id=2474
199.
by cbkeller 6y ago
As far as I understand it (which is not very far): Zygote calculates derivatives using source-to-source automatic differentiation. This calculates function inverses (so to stretch the analogy a bit, it's kinda like "source-to-
200.
by cbkeller 6y ago
I see what you mean, but from my perspective what this argument is missing is the context that a large and important subset of the scientific community has never moved to Python to begin with . In other words, there is an even crankier and
201.
by cbkeller 6y ago
Geologist here! Earth's internal heat budget [1] is a really cool subject since it is surprisingly hard to constrain, yet influences the operation of pretty much the entire solid Earth system (plate tectonics is entirely driven by secu
202.
by cbkeller 6y ago
Yeah, I think type instability is definitely the biggest culprit when someone's ported over some Python or Matlab code for the first time and isn't getting as much of a speedup as expected. It isn't hard to fix, you just have
203.
by cbkeller 6y ago
I had an experience slightly like that at first where I was surprised to only be going 1-2x faster than my old MATLAB, but then I realized my code was full of trivially avoidable type instabilities and got another 100x speedup in Julia.
204.
by cbkeller 6y ago
Tkf has so many cool packages! Less to do with transducers or parallelism, but in terms of just cool and useful stuff, Maybe [1] and BangBang [2] definitely come to mind as well. [1] https://github.com/tkf/Maybe.jl [2]
205.
by cbkeller 6y ago
Thank you for building this and keeping it running!
206.
by cbkeller 6y ago
"Favor" is the wrong word, but only because that implies a lack of payment. Quoting from https://news.ycombinator.com/item?id=18411779 : > 2009 - NSA offering 'billions' for Skype eavesdrop solution [
207.
by cbkeller 6y ago
It exists! `@which fn(a,b,c)` will do exactly that. There's also `@less fn(a,b,c)` if you want to take quick look at the source code for the relevant method, or `@edit fn(a,b,c)` to open the relevant source file in an editor.
208.
by cbkeller 6y ago
Yeah, absolutely
209.
by cbkeller 6y ago
Looking into Amazon's power bill might be a useful start: Fugaku is listed as drawing 28 MW in OP. It's more power efficient than most, but to an order of magnitude that's a number we can work with. Amazon's power usage
210.
by cbkeller 6y ago
Since Fugaku has been in the works for a while, I wonder if Apple just tried to choose the date of their announcement to coincide with the Top500 ranking

More ›