Frotag

Toontown Online used to be a fairly popular mmo in the 2000s. The game itself is gone but the game engine was open sourced as Panda3d (C++ with Python bindings). The project is still active too, with modern features like PBR still being added. I personally like how it's code-first (vs editor-centric like Godot), but the GUI options are really fugly and clunky.

Bleh, took like 2 hours to scan every file. Anyways, overall stats are here

https://gist.github.com/anon962/26bedc907e1f85123ae92df194de...

Sorry I should clarify, the 250 shows + movies estimate comes from number of shows / movies listed on MAL for this season times 4.

But since I also auto-download "related" releases for each thing airing (like seasons / movies from previous years), this comes out to ~24000 video files on disk (mp4 / mkv, excluding the million bluray raw segments). This tally is probably inflated by the bluray extras / op / ed files and a tiny bit by dupes (tracker is poorly organized)

I'll dump the actual stats in a few minutes but point is the 30TB has a lot more than the year's worth I mentioned ("250 seasons / movies"), because myself and the people I share the library with like to catch up on stuff before watching the newest release.

But yeah looks like ~15TB are from bluray remuxes which is definitely overkill (40~100 GiB per season).

> Can you explain what mpv offers over vlc ?

People have complained about VLC's decoding quality in the past (no specific source but lots of google results). I'm sure there's valid bugs that have been patched over and some still there, but I'd take it with a grain of salt since VLC is super popular and therefore more noise when it drops the ball.

That said, I personally prefer MPV since I can display multiple subtitles at once (for learning foreign languages) and also drop in upscaling / sharpening filters like [1]. Oh and VLC is kinda buggy on my android 14 pixel when streaming from ftp / samba sources (video freezes when seeking too often).

[1] https://github.com/bloc97/Anime4K

Personally anime uses around 30TB for me. For the best quality (mostly 1080p + bluray) of everything released in the last year (~250 shows / movies), but also including older releases for the same franchise (eg past seasons / related spin offs).

For western media it'll really depend on how much quality you're willing to forgo (4k? hdr? blu-ray rips?) but 20GB for a 1080p season pack (~10 eps) and 5GB for a movie is pretty normal. And a cursory glance at certain sites suggests 250 seasons and 850 movies were released in 2023, so ~15TB seems like a good estimate for a years worth. Releases for asian stuff should be pretty similar in size, except less hdr / bluray options.

Storage wise I just have a raid 1 (zfs) setup with a bunch of 14tb drives (at 8~10usd per TB [1]). I did need to buy a hba doodad to convert a pcie slot to more sata slots though.

[1] https://diskprices.com/

Oh wow, I just binged Paranoid Mage and it was really good (especially books 1-3). Thanks for the recommendation!

Some people call this genre rational fiction. Personally I really liked Harry Potter: Methods of Rationality. The MC was super narcissistic and nitpicky, a true engineer. For example he was often looking for ways to create perpetual motion machines / infinite energy with magic.

Sadly that's another author I'm already caught up with. His SCP series was really memorable though.

Anyone have recommendations for similar authors?

I love how his stories grow in scope (one discovery plausibly leads to another and eventually interdimensional travel). I know Adrian Tchaikovsky is a popular recommendation around here (especially Children of Time and rightfully so), but I'm looking for something with less drama / less character-driven. The 3 body problem is a good example.

Also tangential but I recommend trying Last and First Idol for anyone looking for ridiculous, gory, self-consistent sci-fi.

Lenovo laptops tend to be best bang for buck but ime poor build quality and support. I've sent each of the three I've owned in for warranty repair at some point. It took at least a month for all of them and two of them still died a month or two later from repair-related issues (eg glue around new lcd housing was too weak and it eventually warped / cracked).

I personally prefer more meat-y plots and stay away from series with scenes that are cute (or stylish) for the sake of being cute (or cool).

That said, here are some of my favorites:

https://myanimelist.net/anime/13125/Shinsekai_yori (2012)

https://myanimelist.net/anime/42310/Cyberpunk__Edgerunners (2022)

https://myanimelist.net/anime/23283/Zankyou_no_Terror (2014)

https://myanimelist.net/anime/2581/Kidou_Senshi_Gundam_00 (2007)

https://myanimelist.net/anime/50709/Lycoris_Recoil (2022)

Also an honorable mention for Pantheon. It's a western animation but season 2's finale was basically one of greg egan's books come to life.

https://www.imdb.com/title/tt11680642/

> meaning 20% accuracy is equal to 80% accuracy,

Only if your model is outputting a yes/no answer right? And that your definition of accuracy is "class with highest probability" (and not "N classes with highest prob")

If your dataset has more than 2 classes like MNIST, a super low accuracy only tells you to ignore the class the model guesses. It doesn't tell you which of the remaining classes is correct

I've run into this once. IIRC the workaround was to add a n=n arg to the lambda

If you order everyone alive from youngest to oldest on that date, you'll be older than 5 billion people.

It's pretty common when wrapping a function that has a large number of config options.

The wrapper is usually some shorthand for building a handful of those args or adding some side-effect, while still allowing the caller access to the remaining config options via kwargs.

Here's one example of that in the wild https://matplotlib.org/stable/api/_as_gen/matplotlib.pyplot....

Theres ~10^31 possible links.

Assuming theres 100 billion valid links and someones scanning 1mil links per second, they'd still be averaging 1 discovered link per million years.

On android, some video ads will even pause if you pull down the notification bar.

I suspect that even if an official release is available (with free samples) potential buyers tend to check out the scanlation first. Mostly since checking for updates on websites that aggregate fan scans are more convenient than checking 3ish publisher sites.

But anyways, I'm playing devil's advocate but I'd imagine a sloppy scanlation of chapter 1 has very real consequences on interest in the later / better official release. By sloppy I mean distracting grammar mistakes, poor image quality, etc. I know I personally wouldn't give the official a chance if I mistakenly thought the plot / art was meh from the scanlation. And that's bad because it doesn't support the authors, who tend be super overworked.

These days it's not uncommon for scanlation groups to be in it for the donation money. Well I say "group" but it's usually just the leader pocketing the money and recruiting bored / enthusiastic teens.

The excuse is that they need money for buying books to scan or digitals, but imo unless they're working on many series, spending 1~3usd per chapter (at 4~8 manhours / chapter), isn't bad enough to warrant begging for donations. Plus the digitals tend to be free if they know where to look.

On a lighter note, not every group is like this, like you mentioned, some people just genuinely think some stories are worth sharing. Others just wanted to practice reading / writing a foreign language and figured they might as well write the translation down and share that.

Someone stirred up a bit of drama doing something similar a few years ago. The biggest difference is that they were selling it and harvesting it without any consent from users or server admins.

https://web.archive.org/web/20221205181242/https://www.reddi...

I like the author (read through a few of his other series) but it gets a little tiring how often he switches between perspectives. Each perspective is a different plotline and they all tie together in the finale but it feels like a 50-100 page story is being stretched into 300.

Especially since the dialog can be a bit boring imo (each character has some quirk and all their lines / actions tend to be predictable when taking this into account). IIRC, the Children of Time books aren't so bad in this regard (less drama, more reverse-engineering alien tech), but I definitely think the newest book for The Final Architecture kinda suffers from this.

But yeah nonetheless I really enjoy how creative the world building / overall plots are and would definitely recommend his stuff to anyone who likes sci-fi.

The specs are nice on paper but even at 60Hz and minimal brightness, the battery on my brand new one barely lasts 2-3 hours when just using the browser. It'll even drain completely if I leave it on standby for a few days.

There's also a serious red tint to the screen [0].

[0] https://old.reddit.com/r/Lenovo/comments/zmht0j/lenovo_p11_p...

My goto example for RxJS are event handlers. Heres an example that implements drag-and-drop (for which the vanillajs / d3js versions aren't nearly as pretty)

It basically translates to

1. Wait for a mousedown event

2. Start listening for mousemove events

2a. On each mousemove event, do something

2b. When a mouseup event occurs, stop listening for mousemoves and return to 1.

    import { fromEvent } from 'rxjs';
    import { exhaustMap, takeUntil, tap } from 'rxjs/operators';

    const el = document.querySelector('body');
    const mousedown$ = fromEvent<MouseEvent>(el, 'mousedown');
    const mousemove$ = fromEvent<MouseEvent>(document, 'mousemove');
    const mouseup$ = fromEvent<MouseEvent>(document, 'mouseup');

    mousedown$.pipe(
      exhaustMap(mousedownEvent => mousemove$.pipe(takeUntil(mouseup$))),
      tap(mousemoveEvent => {
        console.log('mousemove')
      })
    ).subscribe();

https://stackblitz.com/edit/drag-and-drop-rx-rvdkzv?file=ind...

---

More specifically I've found RxJS makes simple one-off calls (eg API fetches) ceremoniously painful (do I need to unsubscribe? do I get the value sync-ly or async?), but anything that involves multiple values (eg websockets, polling, event handlers) gets much cleaner.

My biggest complaint with RxJS is probably the dev experience though. It makes call stacks / the debugger useless so you have to get comfortable with console.log's everywhere. Which is one of the reasons I avoid chaining observables and filter()s in general.

Zero effort solution is to have permanent group chats so background info like discussions on requirements are searchable.

I settled for reading the docs and searching github issues like a caveman. I wonder if this'll eventually popularize low-key advertisements in bug reports. Like "I scanned the repo using this (my company's) tool and found mixed usage of ' and ""

The grammar is really similar to english (subject-verb-object). Words only have one form (they do not change with tense / plurality / etc).

If you translated each word indivdually you'd probably get an almost correct english sentence.

So yeah, I'd like to say it's not super difficult. Especially compared to other east-asian languages like cn, jp, kr. But I'm probably biased since I grew up around it (in the States, so it isn't my primary language).

I've been using nginx for caching when the response isn't user-specific. Pretty painless to set up and doesn't bloat the codebase.

(But it's mostly for small personal projects, so grain of salt and all.)

https://www.nginx.com/blog/nginx-caching-guide/

stdlib is where things go to die. Backwards compatibility, limits on dependencies, release cycles, etc all make development a pain and discourage potential contributions. Or so I've read.

Here's one discussion about it in the requests repo https://github.com/psf/requests/issues/2424

Not conda, but using the built-ins should work for most use cases.

---

1. Create a virtual environment named venv

  cd $PROJECT_FOLDER
  python -m venv venv

2. Activate it.

  source ./venv/bin/activate

(In vscode, Ctrl+P > Select: Python Interpreter)

3. Define and install dependencies

  nano requirements.txt
  pip install -r requirements.txt

(or just "pip install ___")

Some common Angular Material gotchas may include...

  - Checkbox text doesn't wrap
  - There isn't a datetime picker (only date)
  - Styling library components / child components is a pain in the ass. Their escape hatch for this (ng-deep) has been deprecated for years without a replacement.