HNHacker News
TopNewBestAskShowJobs

Nadya

4,139 karma · joined April 7, 2015

nadyanay.me - U+26A5

Democracy is totalitarianism. Freedom is oppression. Censorship is freedom.

[ my public key: https://keybase.io/nadya; my proof: https://keybase.io/nadya/si gs/lklP5B4B7JREUYCzdbw60w0u3esviheCxECPU0D9TgU ]

submissionscomments
Nadya··on Japan Begins Experiment of Opening to Immigration
>Strange that Japan isn't adopting a similar points-based system

They do use a points-based system for highly-skilled workers: http://www.immi-moj.go.jp/newimmiact_3/en/evaluate/index.htm...

>One would presume a Japanese immigration policy and pathway of citizenship would mandate learning the language and other ways to facilitate assimilation

It strongly does and is incredibly difficult otherwise. When applying for citizenship if you struggle at all when asked questions in Japanese you will likely, if not certainly, be rejected. There are countless anecdotal stories out there confirming the difficulties in attaining citizenship from people who already have permanent resident status.

Nadya··on Japan Begins Experiment of Opening to Immigration
There are only so many ways to make cheese and most of the process is going to remain the same. Being a different or unique culture doesn't fundamentally change how cheese is manufactured.

I fail to see your point. There are only so many ways to do set immigration policy and currently Japanese policy is different from ones used by Western countries. Much to the chagrin of the Western countries criticizing it as their own policies are breaking down and giving rise to nationalism and increasing social and racial tensions.

Nadya··on Japan Begins Experiment of Opening to Immigration
We've seen the same results in every country that has attempted it, so why would one more country be any different? How many repeating occurrences of something occurring do you need before you can look at the data and go "There's a strong correlation there." without being called or implied to be a racist?
Nadya··on Japan Begins Experiment of Opening to Immigration
They're only "hard to tell" if you don't spend a lot of time around people of Asian ethnicity. It is very rare I cannot tell a Japanese person from a Chinese person from a Korean person.

Reading comments on this thread below yours, there are a surprisingly high number of people on HN who think all Asians look alike. Spend a few months in an Asian country and white people begin to all look alike while Asians become quite distinct.

Also, it is quite easy to tell most tourists apart from foreign immigrants. The backpacks of gear, maps, constantly getting lost, or participating in things mostly designed for tourism gives it away.

Nadya··on Ask HN: Should I teach my kids to type with QWERTY or Dvorak?
>Not in my case. I first learned to type properly on Qwerty.

I'll take you at your word, but most people I see claim this have extremely poor typing habits. Similar to a self-taught musician who ingrained poor habits because they never had anyone around to correct them.

>I'm going to need to see some proof for that assertion. Just going off the top of my head, I'd say I type far more with my left hand than my right using Dvorak.

Sure thing. Summing the frequencies in the English language and applying it to the keyboard layouts gives:

Dvorak: Left 43.2%, Right 56.8%

QWERTY: Left 58.7%, Right 41.3%

https://en.wikipedia.org/wiki/Letter_frequency#Relative_freq...

Nadya··on Ask HN: Should I teach my kids to type with QWERTY or Dvorak?
As someone who learned to type in QWERTY, Dvorak, and Colemak (in that order) my belief is that most people attribute ergonomics and improved typing ability to the new alternative layout they learned instead of the fact that they finally learned how to type properly, seeing as very few people I know seem to have learned how to properly type - even ones who can type much faster-than-the-average (80-120wpm where average is well under 60wpm)

I found Dvorak to be significantly worse than QWERTY in both speed and comfort, seeing as Dvorak is more heavily biased towards the right hand and I am left hand dominant (although mostly ambidextrous). I then gave Colemak a try and found it to be a bit more comfortable and eventually trained to reach my previous QWERTY speed of 165-170wpm. I use QWERTY at work but Colemak at home.

Note: by "mostly ambidextrous" I mean that most tasks I can use either arm/hand. I write with my right but throw and punch with my left.

Nadya··on White House launches tool to report censorship on social media
As we all know, algorithmic classification has zero issues [0] and is 100% accurate. Even Twitter's anti-ISIS classification banned a bunch of ISIS watchdog accounts but it was deemed that the collateral damage was worth it. In this case, the possible collateral damage of banning non-racist, non-white supremacists is too damaging to Twitter's reputation among a large portion of their users.

The non-politically charged answer to this is that the algorithm isn't (and likely could never) be perfect and so the statement that it would be politically bad due to collateral damage, exactly as the employee from Twitter said, is accurate. Without all the "nudge nudge, wink wink Racists=Right amiright?" rhetoric provided by the media outlets that reported on it.

Going back to the algorithm, how would one even train it? At one end it would likely be useless and at the other it would be far too inaccurate.

Would it ban for use of the OK hand emoji? How about for saying "it's okay to be white?" Would it judge someone as a white supremacist for pro-white speech or only for anti-POC speech? How about people who tweet about or quote Hitler and how would it know the context? Since "retweets aren't necessarily endorsement" would it ban for retweeting white supremacist speech to draw attention to it/call it out? Would it attempt to do classification and then ban anyone who's mostly followed by or tweets towards/is tweeted at by people classified as white supremacists? Would it ban for racial rhetoric? Being against immigration?

Even assuming it could accurately classify - it quickly becomes a measure of "who's definition are you using" and "how far is far enough and how far is too far?"

[0] https://www.theverge.com/2015/7/1/8880363/google-apologizes-...

Nadya··on Tech companies and governments sign up to Christchurch Call agreement
Incitement of violence is actually mostly legal under U.S law except when it calls for imminent action that is likely to happen, which is a surprisingly high bar to meet. See Brandenburg v. Ohio, 395 U.S. 444 (1969) [0]. Take note that this ruling was in defense of the KKK, but that it was later cited to protect the speech of the former NAACP Leader, Charles Evers [1].

>The conditions that must be met to impose criminal liability for speech that incites others to illegal actions are imminent harm, a likelihood that the incited illegal action will occur, and an intent by the speaker to cause imminent illegal actions. This precedent remains the principal standard in this area of First Amendment law, since the Supreme Court has not revisited it.

[0] https://supreme.justia.com/cases/federal/us/395/444/

[1] https://www.aclu.org/blog/free-speech/free-speech-can-be-mes...

Nadya··on Tesla's solar factory is exporting most of its cells
It began in 2001 and the effects become very obvious when you watch country GDP's. Watch the of countries, but especially China, in this animation of country GDP: https://gfycat.com/energeticrigidalbertosaurus
Nadya··on A final proposal for Rust await syntax
>I don't really know why modern programming languages bother with reserved words

Preventing potential footguns prevents a large number of bugs. In fact, preventing a specific kind of footgun is one of the primary reasons that Rust exists.

Fewer footguns means easier to work with code. Easy to work with code means fewer bugs. If having reserved words removes more footguns than it creates than they are a good thing to have - and IMO - they do indeed prevent confusion and footguns.

As long as these footguns can be removed without giving up much in exchange, I say remove all footguns!

Nadya··on Show HN: Check if a scientific paper has been supported or contradicted
The guidelines linked to in your Terms of Service fails to load / doesn't exist: https://scite.ai/guidelines You also may want to update your error message as it reports the Nginx version and Operating System being used.

How are "experts" classified? People misread studies all the time so trusting users to classify if a study supports or contradicts another study gives me even less faith than the text analysis method. Also as a suggestion, have the "Example Report" link to a random report or several different studies.

The presentation of this is really clean and it seems like an amazingly useful tool - as long as classifications remain accurate (even just the "mentioning" metric is extremely useful!)

I'm looking forward to opportunities to use the site more, this is totally something I am willing to get behind and support. Not sure how I feel about the indemnity clause, however unlikely it is to ever come up. Not even sure if it would be upheld by courts but I dislike how common it is to see nowadays.

Nadya··on A Conspiracy to Kill IE6
Who would expect the browser user-agent to actually call itself "EDG" and not "EDGE" though? If you edit the user-agent to include "EDGE" in place of "EDG" everything works, someone thought "EDG" was surely a typo. Malice and stupidity and all that.

Now why they're still sniffing user agents instead of doing feature detection is a good question for Google, who themselves, to the best of my knowledge, push for feature detection instead of sniffing user agents as best practice. Do as we say - not as we do?

Nadya··on It’s Hard to Learn French in Middle Age
You can, and I have, make your own Memrise courses as well. I'm better at remembering to visit a website than I am at remembering to open a program. It's a silly difference, but I almost always have a browser open and every new tab I open serves as a reminder. I always fail to remember to open/use software. The times I've tried converting to Anki rather than relying that Memrise will continue to be available, my study time per week dropped by about 95%~.
Nadya··on Samurai Helmets
It's a self-hosting compiler written in Scheme:

https://ideone.com/dvhScJ

Nadya··on Ikonate – fully customisable and accessible vector icons
Would need data about whether or not people are more or less likely to

1) Add text when it is missing but the attribute is there to serve as a reminder

OR

2) Change text when it is wrong

Neither one is "wrong" but one might be better than the other.

Nadya··on Ikonate – fully customisable and accessible vector icons
I handle ADA issues at work and deal with updating sites to adhere to WCAG 2.0. So perhaps I can offer my $0.02 and clear this up.

It is actually impossible for anything to be accessible out-of-the-box, but they have been designed to be accessible and have sane defaults. Those statements do not contradict one another.

For example, if you use a magnifying glass icon to mean "zoom" but their default title text for a magnifying glass is "search" then it is not out-of-the-box accessible even though it was designed to be fully accessible. They even mention this scenario (generically, not specifically) in their disclaimer.

>E.g. all the titles and descriptions given to the icons should be treated as placeholders and changed in implementation depending on an actual role of each icon. In many cases, you may not need them altogether

Having the placeholders there is questionably good - it would be better to have the attributes there but blank in my opinion. It will do more harm than good being wrong than being blank, but that's a difference in preference for implementation.

Nadya··on In China, tattoos border on illegal – and they're his life's work
Places will enforce it - I'd say most do not - but my sample size is only large enough to say confidentially that the places my friends went to do not, with one exception that enforced it.

Source: Friends with tattoos who have visited Japan as I don't have any myself.

E: This is one of very, very few shows I'd ever recommend and it covers this topic: https://www.viceland.com/en_us/show/needles-and-pins , specifically the ones in Japan and South Korea which have very similar culture surrounding tattoos.

Nadya··on Mastodon and Keybase
I solved that problem that way too: https://nadyanay.me/identities.html

It comes with some issues, namely that I suck at keeping it up to date and that not all identities I would like to list there have a way for me to provide proof beyond my word alone. For most use cases and attack vectors I consider this sufficient enough. Now this is outside most peoples' threat models, but Keybase also provides some mitigation against some other scenarios.

1) If nadyanay.me becomes compromised the imposter could update /identities.html with a new and fake list and I would need to update my link everywhere it is used or I would be pointing people to the imposter list. I have more faith in both (a) Keybase is less likely to be compromised and (b) in the event Keybase has become compromised someone will notice. Nobody would notice if my personal site was compromised, as even my closest friends don't regularly browse my website. It could honestly take weeks or even months to discover the file had been changed.

2) A person who compromises my account(s) must also have access to my private key in order to sign messages in my name. This is important because even if any of my accounts is compromised they're still unable to prove they are me if asked. This is something I actively practice with a few online friends of mine. We pretty regularly lend large sums of (virtual) game cash to one another worth in the range of $10,000-$15,000 USD if RWT'd. The last thing either of us would want is an imposter asking to borrow some money in-game from them and selling it off and so anytime we ask to borrow some in-game cash we ask to see a signed message. I admit that's the primary reason behind most of my signed messages...

3) Any attempts at creating a new key will allow users to see that my key has been revoked and replaced. Users who had signed my old key would need to re-verify with me that my new key is valid. Social engineering and people's casual use cases means the imposter would just claim to be me and most people would believe them. Few would bother verifying but it at least provides an additional opportunity for the imposter to be outed.

Nadya··on Mastodon and Keybase
Task: Send me a tweet on Twitter. Careful not to send it to any imposters.

Challenge: Finding me on Twitter. For example, I am not @Nadya

Extra Credit Challenge: Let's say I'm e-famous enough to have imposter accounts but not have a Twitter "verified" badge. Which Twitter account is the real me? And how do you know?

Where Keybase comes in: On my HN profile itself you can find my signatures on Keybase. Keybase is not necessary for these signatures but becomes a convenient place to look. You also do not need to trust Keybase; although in practice many people will. Don't lie to me and tell me you'd verify the keys. :)

Now you can go directly from my HN profile to my Twitter profile and tweet at me knowing that I am who I say I am. Or at least the individual posing as me has access to three of my accounts (HN, Keybase, and Twitter) and that you'd at least be talking to the same person.

The social proof and web of trust bit is where Keybase falls down but that's an inherit flaw of the web of trust (key exchange parties aren't as popular as they used to be and people will sign/trust keys of people they've never met IRL). Ultimately you'll have to trust that the people who follow me on Keybase are certain beyond a reasonable doubt that I am who I say I am. From there, you can trust the social proofs.

I personally use it so that people can find me on other services more easily and know that they are speaking to me.

Nadya··on Adblock Plus filter lists may execute arbitrary code in web pages
>like that it considers requests to subdomains third-party)

As it should. Sometimes subdomains are third party. If I whitelist example.org I do not want any subdomains to be whitelisted without be explicitely whitelisting *.example.org as a wildcard or any specific subdomains.

Proof of concept: I constantly host what I could consider 3rd party resources on subdomains for clients. eg: billing.example.com where billing. isn't owned or operated by example.com and could be running who knows what in terms of Javascript and Ads that I may or may not trust without whitelisting billing.example.com. The most common of these are specific marketing/tracking forms that they wish to run through some third party marketing agency and not our in-house tracking systems.

Nadya··on Sorting in Japanese – An Unsolved Problem (2011)
>That's polemical and I would not tell it that upfront to my Japanese friends but I feel like kanjis are a vastly inferior writing system compared to the alphabets they have.

The reason Kanji are used instead of Kana are due to the vast number of homophones from a language limited in the number of phonetic sounds it has. It is how you can tell はし「橋」 from はし「端」 from はし「箸」. Sure, context would almost always let you know which word is used but that requires some additional parsing. Other languages have more sounds, thus fewer homophones, and this becomes less of a problem to determine which witch is which. Or as just shown - "witch" and "which" can be spelled differently even when they sound the same, something less possible with kana. Japanese has an insane number of homophones and reading kana-only sentences quickly becomes a headache because of it.

I don't "read" sentences that contain kanji. I glance over the sentence and understand what it is saying. It's like watching someone act out a scene vs reading the scene's script. Although I am still learning Japanese, for sentences where I know all of the kanji, I can understand the Japanese much faster than the English translation (my native language).

Nadya··on 58 Bytes of CSS to look great nearly everywhere
What is the username in this image? [ https://i.vgy.me/faxIj4.png ]

Many people select fonts where there isn't ambiguity between similar characters like o/O/0 and I/l/1. Reading the home page for the font [ https://concoursefont.com/ ] gives me a headache personally. The lowercase 't' and 'f' fight one another. The 'f' with its flowing curve and the 't' in its rigid straightness makes the text fail to flow well.

My favorite font is guilty of having ambiguous characters as well - and while I'm not personally very bothered by it - I can understand and empathize with the annoyance other people have towards issues like this.

I use Stylus to inject my own stylesheets and force every website to use my chosen fonts - their design be damned. But I also eventually write my own CSS for the site myself if I use it frequently so my opinion is likely invalid as 99.9999999% of users would never do that.

Nadya··on 58 Bytes of CSS to look great nearly everywhere
I recognize some of those scrots. Funny to see another /g/ent with a minimalist setup and website.

I have about twice the amount of CSS, but most of it was to match my personal setup of colors/fonts: https://nadyanay.me/assets/css/style.css

361 characters (yours) vs 615 characters (mine)

Remove the font family and .fav classes and some of the unnecessary border overrides and mine comes out to a more equivalent 319 characters, while still retaining the spacing and color scheme.

Nadya··on Ask HN: Are there any decent image hosts left?
It may shock you that scandalous names such as "Microsoft Azure", "Imgur", "Photobucket", "Imageshack" also have to deal with large amounts of child pornography and that the name of the file host has little to do with it, the mere existence of a file host is all it takes.

I also incur zero administrative costs from dealing with anything because I don't allow anyone to upload to my website (and never have). If you have a free file host, you will be dealing with child pornography within a few weeks, almost guaranteed. There are systems made to automatically deal with known/"common" images [0] [1]. Also to clear up a few things regarding the name, the domain of my site is "kimiwo.aishitei.ru". The name of my host is "Lovehost". The domain is Japanese and translates to "I love you". The software it runs is what is called "lolisafe".

[0] https://news.ycombinator.com/item?id=9868352

[1] https://www.microsoft.com/en-us/PhotoDNA/

Nadya··on Ask HN: Are there any decent image hosts left?
This is exactly what I do: https://kimiwo.aishitei.ru

It's running an instance of lolisafe, here are some other sites (many which do allow for registration): https://github.com/WeebDev/lolisafe/wiki/Sites-using-lolisaf...

I don't allow sign ups. It's for me and a few friends only. I pay $5/mo and host my own images - after I grew sick of moving from pomf.se clone to pomf.se clone as they'd all eventually shut down.

Image/File hosting isn't profitable, Sir_Cmpwn [1] has a good write up on why it simply isn't a profitable/good business to get into. In fact, if you read the replies to Sir_Cmpwn's post you'll see I replied... and was using mixtape.moe at the time. And guess what? safe.moe, my other listed option, has shut down too. So it has been 2~ years before both of the image hosts I was using have shut down. safe.moe shutting down is actually why I began hosting my own images with my own domain.

The problem is they don't scale. Once you stop being niche and become popular - the cost to host everything starts growing exponentially while the people willing to donate does not. You'll eventually be spending more on the server(s) than you earn from it. You tank. You shut down. Then someone else thinks they can do better and the cycle continues. This is why there are very few free [0] file hosts that are capable of surviving for long periods of time. I've ran the math once and assuming I grew to 1,000~ regular users a month, I'd have about a years' runway before I'd have to begin charging users or shut down; assuming no donations and none of the 1,000 users are particularly abusive of the service. With donations it might last two or even three years, but it would be an inevitable shutdown as I wouldn't be willing to pay to provide the service past that point. 2-4 years is how long I could use a given host before it went to shit (suffered from the "poison" of needing to monetize to survive) or shut down (couldn't monetize well enough), so it seems like my math isn't too unique to myself either. And this is just the financial side of things - there's the whole administration side of banning abusive users, dealing with DMCA's, dealing with child pornography, and the tech side of scaling, upgrading servers when needed, etc. It's a whole lot to take on.

Sir_Cmpwn, if you happen to read this, I suppose you can add mixtape.moe to your list in a 2019 update. :)

[0] Typically ads somewhere, but many are donation driven by supporters willing to pitch in.

[1] https://news.ycombinator.com/item?id=13810889 and their blog post is here https://drewdevault.com/2014/10/10/The-profitability-of-onli...

Nadya··on Making Anime Faces With StyleGAN
Thank you! Sounds like I need give it another read. :)
Nadya··on Making Anime Faces With StyleGAN
gwern, first off I'd like to mention I absolutely love your work! It's always a pleasure to read your longform writeups on StyleGAN, even if half of it admittingly goes over my head. :)

Can I ask what was modified recently about the write up? I remember reading it a few days back but I see at the top that it was recently modified.

As an aside, I wonder when we'll see the first "fake artist" using TWDNE and Waifu2x then photoshopping/cleaning up some of the irregularities of the generations to output a bunch of "well drawn art". Someone who's decently talented could also use it as a shortcut to avoid a lot of time spent drawing faces/hairstyles and use the generated faces as a starting point.

Nadya··on 4chan, 8chan blocked by Australian and NZ ISPs for hosting shooting video
>I don't see the vast amount of "good things" being censored everywhere as a result of the slippery slope of censorship.

I would like to ask that you take a moment to reflect on this statement ... ... ... Okay now moving on.

Who's to say they need to censor "good things"? They only need to censor things that you don't agree with censoring. I won't pretend to know your political standings but I can guarantee you that your opinion and the governments opinion on what should be censored will eventually reach a disagreement.

People tend to give governments powers during a time of need or when they are frightened often in exchange for "increased safety". A perceived increase of safety that may - or may not - result in more safety but with 100% certainty results in giving up certain rights or freedoms.

For the easiest example of exchanging things for a perceived increase in safety. What do you think of the TSA? Is the intrusiveness and racism in the 'random selections' actually resulting in increased safety and security? Or is it mostly security theater? It tramples over peoples' privacy and sometimes violates their rights for a perceived increase of safety but the failure rate is so high that it is meaningless. The government isn't going to get rid of that power (the TSA) so easily. After all, we gave it to them and they'll gladly keep up the security theater in exchange.

The U.S signed away large portions of our privacy and removed many checks & balances placed on the government after the 9/11 attacks. The USA-PATRIOT Act was forced through in a time of need where people were frightened in exchange for safety and security.

By the time you realize you've given the government too much power it's far too late to take any of it back. History is full of examples and history may not repeat itself, but it sure does rhyme.

Nadya··on 4chan, 8chan blocked by Australian and NZ ISPs for hosting shooting video
The "it's a slippery slope fallacy to assume that censorship of bad things will lead to censoring good things" would be a more compelling argument if we didn't have thousands of years of human history proving that it happens again and again.

"Oh - don't worry, this time in human history it won't be a slippery slope!" is not a very compelling argument when we have historical precedent of the opposite being true time and time and time and time and time and time and time and time and time and time and time and time again.

Nadya··on 4chan, 8chan blocked by Australian and NZ ISPs for hosting shooting video
There was never a /news/ "back then", it was created in 2015 and is a text board. There was only /new/ and /new/ and /pol/ never co-existed. /pol/ was created after /new/ was removed because moot was sick of /new/ going constantly off-topic with racism and bigotry. The decision was made to more or less cease moderation of /new/, rename it to /pol/, and use it as a containment board.

Also people need to realize chan culture. Each board may as well be its own website. 4chan hosts /pol/, it also hosts /lgbt/. 4chan's moderation (rather, lack of) is being seen as an issue because of how heavily moderated the internet has become in the past 20 years.

Reddit, the "bastion of free speech", bans every subreddit that gets in the way of advertisement money, giving Voat more popularity. 4chan, "a bastion of free speech", banned #Gamergate discussion which gave 8chan its rise to popularity.

I'd rather have extremists on Reddit and 4chan where they can be debated and talked down to by moderates than on Voat and 8chan where they can form echo chambers. Echo chambers are a lot better at recruitment. Pushing "bad people" away where they are out of sight and out of mind hasn't been solving the problem, it has only been making it worse year over year.

← PreviousPage 8 of 34Next →