polm23

The official fastText repo was archived on March 19th. There doesn't seem to have been any official statement on it.

Why not just use symlinks in your home dir? Like: cd ~/blog

Author here. Nobody has ever asked about the art before. It depicts Maruyama Oukyo, a famous painter of ghosts (and other things), where one of his pieces comes to life and frightens him.

https://en.wikipedia.org/wiki/Maruyama_%C5%8Ckyo

Besides the base rate effect did you not see the article already posted in this thread with the many problems with that paper?

https://www.theregister.com/2019/03/05/ai_gaydar/

I also can't disprove the existence of unicorns, but I can cite a preponderance of the evidence.

Why is the shape of your face different than the lumps on your head? Even if you find a correlation in the data why would there be a causative relation? If I'm innocent one day but steal a loaf of bread do you expect the shape of my face to change? The idea makes no sense.

When a computer can accurately predict (90%~) sexuality

Oh you mean the thing where it still "works" with faces blurred out because it's not picking up face shape?

https://www.theregister.com/2019/03/05/ai_gaydar/

criminal proclivity

You mean the smile detector?

https://www.callingbullshit.org/case_studies/case_study_crim...

This stuff is bullshit and it can't work. The idea that it could work is magical thinking with no basis in reality.

I was a little surprised that wasn't the default usage model, but I followed the instructions and it showed up as expected.

https://github.com/polm/fugashi

Sounds like the large buddhist numbers like the 不可説不可説転 (unexplainable unexplainable turn).

https://en.wikipedia.org/wiki/History_of_large_numbers?wprov...

I saw this article shortly before a trip to Berlin in 2019 and went by the button shop. It definitely felt like the guy was some kind of button philosopher, or lord of a button empire.

I hadn't brought a picture, but I told him I was looking for some square buttons and he seemed upset - "buttons should be round, you might as well have a square wheel!"

I didn't find the buttons I was looking for but it an amusing experience.

There are already plenty of 3rd party services to auto-delete tweets after any condition you can want to specify.

https://carbon.now.sh/

I used it in ecommerce to find duplicate product titles.

Hey, this is a book announcement, which I realize is a little thin on content for HN, but if anyone has any thoughts about what they'd like to see in the book we'd love to hear about it.

Masato and I are writing this to address what we see as a gap in Japanese-specific introductory NLP resources. There are a lot of great tools for Japanese NLP, but long time practictioners basically pick up knowhow along the way at some point, and beginners are confused by out of date documentation and tools that offer choices without explaining them.

Our goal with this book is to lay out why things are the way they are and explain why you need tools, what problems they solve (or don't), and how to make good use of the available tools, whether classics or recent innovations.

Especially for the English version of the book, we know there are people who will be interested, but we're not sure that they form a coherent community, so we'd be interested in hearing about what you'd want from a book like this, and what your position is that makes you interested. Are you an NLP researcher who wants to know more about Asian languages? Are you an engineer at a company with a multilingual site who needs to add Japanese support? Are you a CS student studying Japanese? There's definitely something we haven't imagined yet, and the more we know the better we can make the book.

I'm skeptical this will actually see wide adoption.

I don't mean to suggest this is a universal trend, but I remember being surprised to see that Tora no Ana, a comic book shop with tech company aspirations, advertised "super engineer" positions that let you work four days a week. But it turned out this was a position with higher qualifications than a normal engineer that just gave you the option of working four ten hour days instead of five eight hour ones.

https://www.toranoana.jp/recruit/super-engineer/

While the suggested four day work week here sounds like a real four days and not longer hours, I do think that managing company expectations will be difficult.

That sounds like my experience with Agricola. The person who introduced it to me was very enthusiastic and friendly but I found it just a miserable experience.

That said, I play and enjoy a lot of games. Many of them are more social and lighter, and all are more directly competitive/interactive. It sounds like, while of course board games just may not be for you, it's also possible your tastes didn't match with the group you were with.

That said a lot of people tend to like games with no direct competition. These are derisively called "multiplayer solitaire", and a few times I've played some games like this by mistake and been mystified as to what's interesting about them. Recently, partly due to the pandemic, the "roll-and-write" genre, which is basically all like this, has been experiencing a burst of popularity. So a lot of people seem to enjoy it.

It's not as solitary, but Gloomhaven, the number one ranked game on BoardGameGeek, is a very expensive game with a typical session taking four hours, and by the admission of its fans a large chunk of the gameplay time is setup, teardown, and bookkeeping. I played it once and I can't imagine what anyone sees in it, but there's demand so more games like it end up on Kickstarter or whatever.

It's frustrating to see something I can't enjoy at all be the top of the industry in some ways, but the nice thing about games is they're quite diverse, so it's not that hard to find something more to my taste. I just have to keep in mind that some people have very difference preferences and that's not always communicated clearly.

I had been using a chair that was past its useful life so I asked friends for recommendations recently. Besides the chairs you see recommended elsewhere in this very thread, one friend recommended a "gamer chair". I was dubious, but it had some nice points:

- has all the features of expensive office chairs

- not mesh (I hate mesh chairs)

- half the price of fancy office chairs (~500USD)

I got a Secretlab Omega specifically and after a week or two I'm happy with it.

I can't speak to long term back effects, but it has more options with height and angle adjustment than any chair I've ever had and comes with a lumbar support pillow.

PyPI will not host large models, those have to be hosted externally. I had to deal with that before and wrote about that here.

https://www.dampfkraft.com/code/distributing-large-files-wit...

As another reply mentioned, the main things PyPI hosts are compiled extension libs. Sometimes these result in giant 300MB binary files, but more often what happens is you have many versions of a ~20MB file due to the combinations of OS, Python version, and API version.

PyPI packages don't include the base CUDA install (not sure if you meant to imply that, but it can read that way).

But yes, this is it. If you have a compiled library based on CUDA that results in a pretty large binary, and you need one for each version of the CUDA API * Python version * OS, that adds up quickly.

I used to use Unite and Shougo's plugin manager, but his lack of communication or documentation, as well as just kind of abandoning projects when he felt like it, makes me no longer comfortable using his stuff.

My Famicase Exhibition is an annual exhibit of imaginary Famicom games. Anyone can submit a label, title, and description via the Internet, and the labels are printed and pasted on to real cartridges and exhibited in Tokyo.

This year the live exhibit is delayed, but the website is updated with all the games as usual. Click on a game to see it up close and with the description. The event attracts many overseas participants, so about half the cartridge descriptions are in English.

Last year's exhibition on HN:

https://news.ycombinator.com/item?id=23210921

FMS Cat also shared a list of some of his inspirations for this demo on Twitter.

https://twitter.com/FMS_Cat/status/1379090489488470018

There's actually Unicode and CSS support for it, but it just isn't widely integrated / used.

https://ja.wikipedia.org/wiki/%E5%9C%8F%E7%82%B9

Thank you for writing IceCream! I've used it off and on over the years and it's been a great help.

Previously:

(2019) https://news.ycombinator.com/item?id=19867460

(2018) https://news.ycombinator.com/item?id=16377714

Thanks, I'll stick those in!

Ah, right you are - I didn't realize. I'll have to add that.

I had heard of 電子辞書 having this feature, but I've never actually used one. Glad to see that 部品 was the right word, I had been kind of uncertain about that and gotten feedback about it.

https://jp.sharp/sc/eihon/pa660/text/kangi.html

SKIP is cool but you have to learn it, whereas this just uses knowledge you already have.

Thanks! For the second piece of feedback, I actually already do this for a few characters, like ム. ル is another good candidate. I am not sure if ウ is a bit of a stretch... Note this is also done for radicals with obvious correspondences, like 心草水獣, so you can search 水木 for 淋.

For the first piece of feedback, I have had other people request that, but I am not sure how to balance it with having detailed kanji pages. One thing to note is that while it is less convenient, on the detailed kanji page there is a link to "search this kanji" - I use that when I know a character that contains the part I want, but can't remember the name of the part itself.

I'm not aware of any methods like that in widespread use for Japanese.

I think in some IMEs you can search kanji by radicals, but you have to type the radicals phonetically first.

I made this years ago, still update it sometimes.

https://github.com/polm/awesome-digital-collections