Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
rhdunn
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
19 ms
·
331.
▲
by
rhdunn
2y ago
When profiling slow queries/code I often collect the elapsed time of a test where I take 5-10 runs and calculate the mean/average, standard deiviation, min, and max. As well as using line charts on the average, I've used a bo
332.
▲
by
rhdunn
2y ago
The unrealistic squares issue could be resolved by using legal moves for the piece. You could then evaluate the best for 1 move, 2 moves, up to e.g. 4 moves. You could also eliminate moves that visit a square already reached by a previous
333.
▲
by
rhdunn
2y ago
That was my point, though. Don't worry about minor possible changes to the code where the performance doesn't matter. For example, if the ++i/i++ is only ever executed at most 10 times in a loop, is on an integer (where the c
334.
▲
by
rhdunn
2y ago
Premature optimization is about not making a micro-implementation change (e.g. `++i` vs `i++`) for the sake of percieved performance. You should always measure to identify slow points in expected workloads, profile to identify the actual sl
335.
▲
by
rhdunn
2y ago
I recall some Minecraft tests being saved worlds with redstone logic that will light a beacon green if it is working or red if not. That's usefull for games like that. For games like Starcraft 2 with replay functionality, you could pro
336.
▲
by
rhdunn
2y ago
January -- The item for January is ‘0’ = 0 February, March -- The item ... for February or March (the 3rd month), ‘3’ = 3 December -- The item ... for December (the 12th month), ’12.’ So these all use fixed values. The `(10-place)+days mod
337.
▲
by
rhdunn
2y ago
From the TinyStories dataset card [1] the dataset is generated by GPT-3.5 and GPT-4. Reading the discussions in the community tab [2] it looks like there are a lot of incomplete or misspelled words, incorrect grammar, and even Chinese chara
338.
▲
by
rhdunn
2y ago
Starting with a 0 dimensional (0D) point: 1. create a copy of the point and move it a fixed distance along a specific direction (which we will call the X-axis) and join the two points together -- this is a 1D line; 2. create a copy of the l
339.
▲
by
rhdunn
2y ago
Yes, it would return "Russe" when the current locale is French. -- It is designed for displaying the locale to the user on a UI, console output, etc.
340.
▲
by
rhdunn
2y ago
For language: 1. `getLanguage` should return the 2-letter ISO 3166-1 language code, or an empty string if missing. -- e.g. en, ru, fr, etc. 2. `getISO3Language` should return the 3-letter ISO 3166-1 language code, or an empty string if miss
341.
▲
by
rhdunn
2y ago
I signed up for a trial for WoW back in the day. It was about 3 seasons in, and the way the install worked was that you had to download and install each update separately, each of which was 2-3 GB each. I was using a slow dial-up at the tim
342.
▲
by
rhdunn
2y ago
I've experienced LLMs forgetting details (no memory). This is expecially a problem when the information is out of the context window, but I've seen it in other cases as well. I've experienced LLMs lacking spatial awareness, s
343.
▲
by
rhdunn
2y ago
IIUC, it's trained on LibriVox audio mainly, along with a few other sources. I'm not sure how it is handling spelling as the spelling will depend on the source content being read, unless the source text has been processed/edi
344.
▲
by
rhdunn
2y ago
Isn't the issue more that traditional ASR systems use US (General American) phonetic transcriptions so then struggle with accents that have different splits and mergers? My understanding on whisper is that it is using a model trained o
345.
▲
by
rhdunn
2y ago
The default voice sounds robotic for several reasons. It has a low sample rate to conserve space. It is built using a mix of techniques that make it difficult to reconstruct the original waveform exactly. And it uses things like artificial
346.
▲
by
rhdunn
2y ago
That's effectively what eSpeak-ng is doing. For the CMU flite voices they represent the data as LPC (linear predictive coding) data with residual remainder (residual excited LPC). The HTS models use simple neural networks to predict th
347.
▲
by
rhdunn
2y ago
The parent mentioned avoiding the dispatch of a virtual function call used to evaluate the AST/bytecode so I was assuming it was a minimal JIT instead of running the resulting bytecode in an interprative loop. But yes, if you are inter
348.
▲
by
rhdunn
2y ago
Which is why JavaScript engines (and JIT compilers for other languages) are typically designed to: 1. start interpreting the code once it has been parsed, and start jitting a function being called; 2. generate naive bytecode for the functi
349.
▲
by
rhdunn
2y ago
Another advantage to that is that it avoids the branching of the while loop in the interpreter that iterates over the AST, providing better instruction pipelining with having all the run code next to each other. The downside -- especially f
350.
▲
by
rhdunn
2y ago
Ah, ok. I wasn't able to get the last row working right with flex, but didn't experiment too much as the CSS grid approach worked. You could look through the other answers in that StackOverflow question.
351.
▲
by
rhdunn
2y ago
No, just: main.books { display: grid; grid-template-columns: repeat(auto-fill, 180px); grid-gap: 1rem; justify-content: space-between; } That will make the child items 180px wide, repeat as many
352.
▲
by
rhdunn
2y ago
CSS grid makes it possible to implement wrapped content that is uniformally aligned with the last row keeping the same number of columns as the others, and have that scale correctly to different page widths. This is useful in things like bo
353.
▲
by
rhdunn
3y ago
Searching e.g. "xq31 curlyarrayconstructor does not use enclosedexpr": 1. In google, I get 9 results, none of which are a reference to the W3C XQuery/XPath issue. This is because google is actually "helpfully" searc
354.
▲
by
rhdunn
3y ago
Thanks. At the moment forgejo does not have features I'm interested in as I'm not making my internal gitea instance public. However, I'd be interested in it if it provides support for things like hierarchical project grouping
355.
▲
by
rhdunn
3y ago
I'm using Gitea locally. I haven't come across forgejo before. It seems to be forked on the basis of providing a fully open source solution (it states that some Gitea enterprise features are not present). Do you know what the main
356.
▲
by
rhdunn
3y ago
You can say the same thing about Patreon, donations, etc. Sure, you don't have to pay to watch them but its a way of giving back to them. And on platforms like twitch, being subscribed to a channel does things like remove ads and acces
357.
▲
by
rhdunn
3y ago
Like spending money on other forms of entertainment? Going to the cinema, theatre, a concert, renting a movie, etc. will either cost about the same or more.
358.
▲
by
rhdunn
3y ago
I'm using a local audiobookshelf server to manage the podcasts I listen to.
359.
▲
by
rhdunn
3y ago
Yeah. That's something you would expect databases to be good at. Expecially as normalization and many-to-many relationships make multiple JOINs common, which was my point.
360.
▲
by
rhdunn
3y ago
Doesn't that happen when you use a highly normalized database table design, though where you create tables to avoid repeating data? Or if you have a lot of 1/many to many relationships in mapping tables. I.e. you'll end up wi
More ›