henrydark

8 ms·

31.
by henrydark 3y ago
Similarly, in 2003, they had some code-under-caps promotion. I wrote a script to submit thousands of random codes to the website, and subsequently someone from Coca-Cola NZ called my home. They calmed down when my dad said I wasn't hom
32.
by henrydark 3y ago
How do you do pca on dimension 2M?
33.
by henrydark 3y ago
> Grothendieck was, to be clear, a strong mathematician compared to most anyone, but these peers were the most talented young mathematicians in France, and unlike Grothendieck, who had spent the war in an internment camp at Rieucros, nea
34.
by henrydark 3y ago
Personally, no. I did go through it, but didn't get much from it. Beyond the original paper, I'd recommend David Khourshid's talks, which you can find on YouTube. For C++ I used Kris Jusiak's library [boost].sml extensiv
35.
by henrydark 3y ago
My idea of being explicit and clear changed dramatically after I was exposed to D. Harel's "Statecharts: a visual formalism for complex systems". Ironically, I think the paper presents more than just the idea and examples of
36.
by henrydark 3y ago
Who else was disappointed to find it's Apache Kafka and not Franz Kafka?
37.
by henrydark 3y ago
Don't think so, doesn't computer program equivalence require solving the halting problem and undecidable problems? For example, consider the empty program, and the program that print "hello, world!" if an undecidable con
38.
by henrydark 3y ago
I wouldn't trust anyone that is against cake
39.
by henrydark 3y ago
I agree with everything but the last statement. This all comes down to: do you consider memoization to be state. I predict people's answers to this question will come from experience with memoization. Here's mine: I kept trying to
40.
by henrydark 3y ago
Looks cool. This caught my eye > C++ and Java developers expect to be able to run a program at full speed (or very close to it) under a debugger. I haven't worked in Java too much, but in C++ I don't remember having this expect
41.
by henrydark 3y ago
I used to use micromamba, recently moved to mach-nix with nix. It's also really fast, but with the reproducibility of nix.
42.
by henrydark 3y ago
Also "any matrix can be perturbed" assumes a normed field, though those equations work over other fields as well
43.
by henrydark 3y ago
It sounds like you're assuming diagonalizability, which is stronger than "eigen values defined over the field". Even over the complex numbers there are matrices that aren't diagonalizable, like [[2,1],[0,2]].
44.
by henrydark 3y ago
I run ml algorithms like boosted trees (i.e xgboost) on data sets with 30k-1m rows and 200-2k columns. Sorting is the bottleneck, it's what the algorithm does. I doubt I'm special, and I'm sure these size are common
45.
by henrydark 3y ago
> Today's developers didn't learn binary before learning Python I'm sure it's rare, but we had C in 10th grade, and then assembly and electronics in 11th grade at my high school. I remember my first assembly class 21
46.
by henrydark 3y ago
Sorry, auto correct malfunction, I meant "isn't strictly necessary"
47.
by henrydark 3y ago
I've heard the connection between the dotcom crash and terrorism once before, but never got it. The crash definitely happened before 9/11. What terrorists are connected with the dotcom crash?
48.
by henrydark 3y ago
The example in the reference in the piece is to have a header and a footer, where header would be open and footer closed. As for linear types, I mean that somewhere the type system won't allow progress if you have an open type that was
49.
by henrydark 3y ago
Hot take: it's possible to have a strict language where tags don't have to start and end in the same template. For example, make linear types Html, HtmlOpen, HtmlBody, and HtmlClosed, with the addition rules (which could be functi
50.
by henrydark 3y ago
I would be surprised if hammers didn't succumb to exactly the same type of competitions. Here are a few hammer review pieces, pitting metal against metal https://www.nytimes.com/wirecutter/reviews/best-hammer&
51.
by henrydark 3y ago
100 use cases for 1k cores
52.
by henrydark 3y ago
1. I git commit with a token 2. Master has the deployed models, other experiments are in branches - just like other experimental code
53.
by henrydark 3y ago
I wonder what happens if instead of dark matter halos, mond is used for galaxies that do the lensing. If the cmb estimate gives the correct hubble constant, and mond doesn't do enough lensing for this result to make sense (less mass, l
54.
by henrydark 3y ago
Even though the opposite is recited more frequently, in some cases good can be the enemy of perfect
55.
by henrydark 3y ago
Well, he could mean that some notes are connected (where I think we all mean the 8 notes are of equal measure). So between every two consecutive notes you can add a line, multiplying the space by 2^7
56.
by henrydark 3y ago
A different, also very explicit way to go about this type of problem, that also generalizes fully, is to use a Bayesian hierarchical model of a dirichlet process and sub-isotonic regressions. Gelman et al have written a lot about this, and
57.
by henrydark 4y ago
What controversial theory did Galileo propose?
58.
by henrydark 4y ago
Sqlite?? MS Access for the win!
59.
by henrydark 4y ago
Say what you will about sql, polars, pyspark, or whatever else. But nothing beats pandas' df[col].value_counts().value_counts()
60.
by henrydark 4y ago
The flip side is a company interviewing me, and didn't open any of my relevant social profiles. Instead, I'm getting complexity questions whose solution amounts to "do a cumulative sum?"

More ›