Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
blt
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
12 ms
·
121.
▲
by
blt
2y ago
Nothing happens when I click anywhere in Firefox 115 on MacOS.
122.
▲
by
blt
2y ago
yes, in this case I'd say a good crowd funding platform is the IRS.
123.
▲
by
blt
2y ago
> It's more important to go deeper on a topic than what the topic is. 100% agree. Both calculus and probability can be used successfully as a vehicle to teach rigorous though, which is the real point of math education.
124.
▲
by
blt
2y ago
Not sure why you are downvoted for this. Your points have merit. I agree that stats classes can have a "cookbook" flavor and do not generally lead to a deep understanding of probability. But I would rather fix the stats classes th
125.
▲
by
blt
2y ago
been using PyPI a lot recently for non-Python stuff such as FFmpeg and Eigen. Part of the reason why I have been able to ditch Homebrew entirely!
126.
▲
by
blt
2y ago
@srush some feature requests after finishing these: - It would be nice if we could use Boolean arrays as 0/1 values in arithmetic without multiplying by 1 first, as Torch and NumPy allow. - The coloration of the output diagrams is hard
127.
▲
by
blt
2y ago
I suggest to implement LaTeX math in comments if you want to support math-heavy fields. Plain text comments make certain ideas almost impossible to express.
128.
▲
by
blt
2y ago
This is cool! It might sound silly, but figuring out `None` and `...` indexing in numpy was one of the more delightful moments of learning as a programmer for me. Tangent: I wish we had not settled on the word "tensor" for multidi
129.
▲
by
blt
2y ago
> 3D CAD apps are among the most complex UIs that exist. I agree. > If you want to make a good one and you only have a small team, the framework had better do a lot of heavy lifting! I am skeptical that general-purpose UI framewor
130.
▲
by
blt
2y ago
Scrabble is interesting because average people think about it like a casual game, but it's nearly as strategic as chess despite significant randomness. Most people play with a completely greedy approach. Thinking just one move ahead ab
131.
▲
by
blt
2y ago
gosh, I wonder if there was anything else about Texas that changed in recent years that might make well-educated people from diverse cultures want to leave?
132.
▲
by
blt
2y ago
the simplicity of the transformer is quite refreshing. especially in vision where the Vision Transformer with linear patch encodings replaces complex intertwined decisions about filter size, striding, pooling, #filters, depth, etc., with th
133.
▲
by
blt
2y ago
yes, that is what I meant.
134.
▲
by
blt
2y ago
This is great. It would be a good C/C++ interview question to compare these. Of course you can't expect a Raymond Chen level performance, but it should give some insight into experience with low level programming.
135.
▲
by
blt
2y ago
[1] Was an empirical paper that inspired much theoretical follow-up. [2] Is one such follow-up, and the references therein should point to many of the other key works in the years between. [3] Introduces the neural tangent kernel (NTK), a t
136.
▲
by
blt
2y ago
That is what people thought until around 2018, but it was wrong. It turns out that deep learning optimization problems have many global optima. In fact, when the #parameters exceeds the #data, SGD reliably finds parameters that interpolate
137.
▲
by
blt
2y ago
Not really, you can always change the indexing to account for it. For example, the GEMM matrix multiplication subroutines from BLAS can transpose their arguments [1]. So if you have A (m x n) and B (n x p) stored row-major, but you want to
138.
▲
by
blt
2y ago
most serious fonts have separate "display" variants for such applications, whereas the version accounting for ink spread is called "text" or "book".
139.
▲
by
blt
2y ago
As a computer scientist, the "differentiable hash table" interpretation worked for me. The AIAYN paper alludes to it by using the query/key/value names, but doesn't explicitly say the words "hash table". I
140.
▲
by
blt
2y ago
Never seen a truss where the diagonal members are all one direction like that. Also bit weird that the strongest part of the bridge is not the middle of the longest span, but it does not appear to be cantilever. Disclaimer: I am a mere Poly
141.
▲
by
blt
3y ago
This is pretty subtle. I wish more languages explicitly distinguished references and values like C++ and Rust.
142.
▲
by
blt
3y ago
There is one big improvement - now ATMs force you to remove your card before they dispense cash. The problem of people forgetting cards in ATMs was completely eliminated due to improved UX design.
143.
▲
by
blt
3y ago
I suppose implementing effects like this in CSS instead of GPU shaders is the modern version of implementing a 3D spinning cube on an 8-bit home computer. Pushing the hardware to its limits is now too open-ended, so we push specifications t
144.
▲
by
blt
3y ago
agree, just a good "sanity check" for first-order optimization algos.
145.
▲
by
blt
3y ago
+1, curious to see if the paper has a convergence rate proof for convex objectives.
146.
▲
by
blt
3y ago
Yeah, the ideal-ness of electronics is much better than other systems like mechanisms, chemical plants, etc. In systems with substantial unmodeled effects the principled methods for tuning don't work as well. But I agree that knowing t
147.
▲
by
blt
3y ago
(2023)
148.
▲
by
blt
3y ago
In my Cuckoo rice cooker it's better to let the rice sit in "keep warm" for a few minutes anyway. When the cooker detects "done" there is no water left on the bottom but the rice grains still aren't finished ab
149.
▲
by
blt
3y ago
Curious if anyone can see the difference between #DD2 and #DD3. It was impossible on my 2013 Retina Macbook.
150.
▲
by
blt
3y ago
Agree, that game is great for anyone who enjoys a color perception challenge! I just bought my first OLED phone, maybe it's time to play again.
More ›