Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
xiaodai
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
23 ms
·
301.
▲
by
xiaodai
5y ago
One thing that's often not discussed is the time required to understand the pros and cons of another language. If you don't know it well enough, choosing it over something else is a big RISK. The investment required to learning en
302.
▲
by
xiaodai
5y ago
very nicely arguued.
303.
▲
by
xiaodai
5y ago
right move
304.
▲
Solving Sparse Linear Systems Faster Than Matrix Multiplication
(arxiv.org)
3 points
by
xiaodai
5y ago
|
0 comments
305.
▲
by
xiaodai
5y ago
Does Paul Graham know Prolog or Wolfram or SAS? If not then... isn't it ironic?
306.
▲
Cause and effect is an illusion: Quantum eraser experiment
(en.wikipedia.org)
2 points
by
xiaodai
5y ago
|
0 comments
307.
▲
by
xiaodai
5y ago
Many of the techniques discussed are documented in my Julia implementation of the fastest string sorting algorithm https://www.codementor.io/@evalparse/julia-vs-r-vs-python-st...
308.
▲
by
xiaodai
5y ago
I want to understand something, why use Spark and DataFusion? As the way to interact with them is SQL. Why not just use traditional DBMS like PostgreSQL? Are there explainer articles on this topic? See Quora: https://www.quora.co
309.
▲
by
xiaodai
5y ago
I thought only China has the scale, know-how, and talent to build technologies companies that rival the US. But let's see. M-pesa is an quite interesting tech-ecosystem to have come out of East Africa.
310.
▲
by
xiaodai
5y ago
Very interesting. Must be the first time I have seen JDF.jl talked about in a post. I guess the point about confidence being low is correct, since the format is not yet finalized and stable. Also, another point is about how fast things can
311.
▲
by
xiaodai
5y ago
Can't agree less.
312.
▲
ML for Non-Technical Readers [pdf]
(pages.dataiku.com)
2 points
by
xiaodai
5y ago
|
0 comments
313.
▲
by
xiaodai
5y ago
group CNN are drop in replacement of convolution layers.
314.
▲
by
xiaodai
5y ago
"Contemporary COBOL". At least they didn't say modern COBOL, or it will be an oxymoron.
315.
▲
Next-Generation Automatic Differentiation (Ad) in Julia
(github.com)
9 points
by
xiaodai
5y ago
|
3 comments
316.
▲
by
xiaodai
5y ago
I am so glad this post exists
317.
▲
Julia's mind-blowing ways of defining functions
(medium.com)
4 points
by
xiaodai
5y ago
|
1 comments
318.
▲
by
xiaodai
5y ago
``` $addresses = [ { 'name' => 'John Doe', 'address' => '123 Any Street', 'city' => 'Anytown', 'state' => 'TX', }, { 'name' => 'Ma
319.
▲
Julia's multiple dispatch explained with Pokemons
(moll.dev)
161 points
by
xiaodai
5y ago
|
108 comments
320.
▲
by
xiaodai
5y ago
hmm, but doesn't it just mean that ppl should extend the same `fit!` method rather the define their own? The bigger issue for production in my experience is about packaging the right model with the right version. I don't think any
321.
▲
by
xiaodai
5y ago
did you end up using package compiler as well?
322.
▲
by
xiaodai
5y ago
need better basic stuff like pca and glm.
323.
▲
by
xiaodai
5y ago
Hmm, reading your comments. I think you have understood how `fit!` works. Each develop will develop their model in their own package. But they just define this fix function. E.g. ``` # in PkgA function fit!(model::ModelTypeA, X, y) # some
324.
▲
by
xiaodai
5y ago
> one of the major grips I have is how little information Julia provides you with stack traces as opposed to Python and JVM traces? LOL Also, python has been around for alot longer. So perhaps that Python has "everything" is a
325.
▲
by
xiaodai
5y ago
Money don't do jobs, humans do. You see the logical fallacy in ur argument now?
326.
▲
by
xiaodai
5y ago
that's why SAS developed Viya so you can run Python and R code.
327.
▲
by
xiaodai
5y ago
Reminds me of a Stephen Chow movie where a nutty inventor invented a solar-powered torch that only works if you shine a torch at it.
328.
▲
by
xiaodai
5y ago
What's yellow and equivalent to the Axiom of Choice? Zorn's lemon
329.
▲
by
xiaodai
5y ago
sarcasm warning. it's demand supply. if there is supply then rent will fall so that these min wage worker can afford it? also, maybe it's a demand issue. these works just don't demand housing. so yeah.
330.
▲
Data Processing Units
(nvidia.com)
1 points
by
xiaodai
5y ago
|
0 comments
More ›