frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Regression: An essay on statistics, eugenics, and my news pipeline

https://jakubdudek.com/writing/regression
1•jakubdudek•42s ago•0 comments

Pax Historia on WorldOS

https://worldos.cc/pax-historia
1•DomaLamma•1m ago•0 comments

The EnshAIttification of Day One

https://schwarztech.net/snippets/the-enshaittification-of-day-one
1•latexr•4m ago•0 comments

2025: The Year I Didn't Write Any Code

https://redfloatplane.lol/blog/11-2025-the-year-i-didnt-write-any-code/
1•chistev•8m ago•0 comments

Spotify Labs Launches Studio

https://labs.spotify.com/studio
1•chaidhat•9m ago•0 comments

Reflections on Do Things, Tell People (2012) (2025)

https://redfloatplane.lol/blog/10-thoughts-on-do-things-tell-people/#original
1•chistev•11m ago•0 comments

How could Britain look under proportional representation?

https://www.theguardian.com/politics/ng-interactive/2026/jul/22/how-britain-would-look-proportion...
2•ravenical•12m ago•0 comments

Samsung in talks to invest in Mistral at 20B euro valuation

https://www.reuters.com/business/finance/samsung-talks-invest-mistral-20-billion-euro-valuation-f...
1•iamflimflam1•13m ago•0 comments

Intelligence is not the main bottleneck

https://www.writingruxandrabio.com/p/intelligence-is-not-the-main-bottleneck
1•Luc•15m ago•0 comments

Codeberg: ToU extension to prohibit LLM-extrusions

https://codeberg.org/Codeberg/org/pulls/1253
1•robin_reala•17m ago•0 comments

OpenAI says AI models went rogue during testing

https://www.reuters.com/technology/openai-says-ai-models-went-rogue-during-testing-triggering-unp...
2•LordDefender•19m ago•0 comments

--end-of-options (git flag)

https://nesbitt.io/2026/07/21/end-of-options.html
2•latexr•20m ago•0 comments

The Vikings didn't wear horned helmets. So who did?

https://danjones.substack.com/p/history-etc-the-vikings-didnt-wear
2•ville•22m ago•0 comments

EU border agency unlawfully transferred data on migrants/activists to Europol

https://english.elpais.com/international/2025-07-13/for-years-the-eus-border-agency-illegally-tra...
2•robtherobber•22m ago•0 comments

AI Video and Audio Generator

https://silknova-ai.com/
1•amonte•26m ago•0 comments

Codeberg bans vibe coded projects

https://codeberg.org/Codeberg/org/pulls/1253#issuecomment-19820434
5•ambigious7777•27m ago•3 comments

Mapping Alien Continents: Achieving Optical VLBI for Exoplanet Imaging

https://www.nasa.gov/directorates/stmd/mapping-alien-continents-achieving-optical-vlbi-for-exopla...
1•fireflymetavrse•30m ago•0 comments

Updating IP Regulations for AI Distillation

https://www.marble.onl/posts/copyright_vs_ai.html
1•amarble•32m ago•0 comments

Hyra: A simple yet effective scaffold for general discovery

https://hy.tencent.com/research/hyra
1•ijk•32m ago•0 comments

Why I'd choose Tauri again for a cross-platform database client

https://tabularis.dev/blog/why-tabularis-runs-on-tauri
1•theanonymousone•33m ago•0 comments

Sharecoin, a Bitcoin Core fork that turns GPU mining into public randomness

https://github.com/TVHeroes/sharecoin
1•TVHeroes•39m ago•0 comments

Thoughts on Integers (2023)

https://blog.xoria.org/integers/
2•birdculture•39m ago•0 comments

Distributed Development with Reticulum

https://reticulum.network/manual/distributed.html
2•bergie•41m ago•0 comments

Show HN: Shirei v0.6 adds iOS/Android support for cross-platform GUI development

https://judi.systems/shirei/blog/release-v0.6-2026-07-22/
1•hsn915•46m ago•0 comments

Melaya – the agent builder that gives AI a governed Android phone

https://www.melaya.org
1•melaya•47m ago•1 comments

Have the Chips Fully Cooled?

https://robinhood.com/us/en/learn/weekly-rundown/3wGS8sjCYEufVoC3EfYph9/
1•dabinat•47m ago•0 comments

Aks HN: Why does GitHub grapql API returns missing data?

1•ThierryRkt•47m ago•0 comments

A hand-cranked device for running local LLMs

https://squeezlabs.github.io/handcrank/
2•antondd•51m ago•0 comments

Session Revocations at Scale

https://www.canva.dev/blog/engineering/session-revocations-at-scale/
1•SerCe•54m ago•0 comments

Show HN: Open benchmark for memory layers vs. pasting full chat history

https://github.com/leapmemory/memory-bench
1•canu21•55m ago•0 comments
Open in hackernews

AI is right about em-dashes

https://joeldueck.com/ai-is-right-about-em-dashes.html
12•freediver•1y ago

Comments

mattl•1y ago
Yeah I never used them before I worked at the FSF. The FSF used the Chicago Manual of Style.
neom•1y ago
Joel is a great guy who apparently loves the em dash. Personally, I find them particularly annoying, pedantic, and always have. Just use a comma, when I'm reading something with an em dash it's always unclear to me exactly how long I should pause, especially when reading aloud..., how dramatic...?!
damhsa•1y ago
a short dash - with spaces around it - is clearly a punction mark, not a word-joiner. what nonsense! anyway, you can not expect everyone to use 4 or more different key combinations depending what style they want. the only solution is to make firefox, word, etc choose the right dash for word-joining, punctuation, minus, number ranges, etc. somewhat similar to how they can choose the width of spaces in justified text.
dragonwriter•1y ago
> a short dash - with spaces around it - is clearly a punction mark, not a word-joiner. what nonsense!

That's not a "short dash", it is a hyphen with spaces around it (more precisely, an ASCII hyphen-minus, but that's really a hyphen that is overloaded in contexts where a proper typographic minus sign isn't available.) There are shorter dashes than the em-dash, one of which (the en-dash) is the normal print alternative for some uses of the em-dash, set open (with spaces) where the em-dash would be set closed (without spaces).

Also, both hyphens and en-dashes are used as word joiners in different contexts; this is not distinct from punctuation but a form of punctuation.

> anyway, you can not expect everyone to use 4 or more different key combinations depending what style they want.

For very informal writing, sure, people are often, according to their own personal comfort, going to be sloppy and approximate punctuation (heck, they are going to do that with spelling quite often.)

> the only solution is to make firefox, word, etc choose the right dash for word-joining, punctuation, minus, number ranges, etc.

Some apps have tools that automatically handle simple cases, but...it's not simple for software to figure out what your intent is and subsitute the correct punctuation mark (the are all punctuation). The actual solution is, as for other less common punctuation marks to accept that there's going to be lots of variation and substitution of easier-to-access characters in informal writing, while recognizing what is correct and maintaining it in formal contexts. But also recognizing that people that know what they are doing and know their tools are going to often use the correct ones in informal contexts, as well.

damhsa•1y ago
it is a dash. you could use an em dash - or a 2em dash - but this dash is probably shorter than 1em - unless like me you have a monospace font. if it can be a hyphen-minus, it can be a hyphen/minus/dash. and that is what it is, because that is how it is used, ASCII be damned. the english language and its typography is beholden to neither ASCII™ nor UNICODE™.
dragonwriter•1y ago
The hypen/minus/<various dashes> distinction long predates ASCII (much less Unicode), conflation between them is a product of ASCII with some contribution from the limitations of mechanical typewriters, the distinction isn't an imposition from ASCII onto the English language.
jasonthorsness•1y ago
I use a plugin[1] on my blog that automatically translates ‘--‘ or ‘---‘ into em or en dashes. I did not see it coming that now my articles will look AI-written :p

Edit: just discovered iOS keyboard input also does this when you type multiple hyphens

1 - https://daringfireball.net/projects/smartypants/