10 comments

[ 3.1 ms ] story [ 23.6 ms ] thread
Somewhat different from my idea of how to get to the front page of Hacker News (from http://ineptech.com/index.html?blog&post=20):

* Use a plainly wrong headline which you intend to weasel out of in the body of the essay (e.g. "Why Javascript is lower-level than Assembly", "Mobile is dead")

* Any "<app> in <techology>" project that hasn't been done yet (e.g. "Building a reverse-Polish calculator in Rust", "Show HN: Umud, the first MUD client for Urbit")

* Devise a variant on an obscure technology or algorithm and give it a quirky name (e.g. "Soda-can steganography", "Homomorphic hashing the Monty Python way")

* Vague counter-intuitive advice for startup founders (e.g. "Raising your valuation through the power of song", "How beating my children saved my B round")

so basically: use a clickbaity title
To be honest most of those articles look like something I want to read. Especially soda-can steganography.
"These queries reveal that Github, the New York Times, and Medium are the sources the Hacker News community trusts the most."

I agree it says something about how HN regards the NYT; but isn't saying Github or Medium are "sources the community trusts" like saying paper is a source someone trusts because they get their knowledge from books?

In the case of Medium, it says essentially that right on the tin. Because, you know, the medium is not the message.
I always interpreted the name as being "it (the content) is not high quality, or low quality, it's medium quality"
Maybe it has multiple meanings, but I thought medium referred to the length of content. As opposed to short form (Twitter) or long form (articles, papers, etc)
This article has been posted 4 hours too early.
Normally, the trick to getting on the front page of Hacker News is to submit an article about Hacker News, but we're all a bit distracted by the GDPR at the moment...
The author draws conclusions based on a sample dataset of 4000 submissions, whereas currently there are 2,962,457 submissions (stories, askHN and showHN combined). Seems a bit risky at best, unless the sample was designed in a very clever way.