Hacker News

Top stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. Does Georgism work? Five years later (astralcodexten.com)
    100comments
  2. PipePipe: NewPipe hard fork implementing SponsorBlock (github.com/infinityloop1308)
    190comments
  3. DeepSeek Elastic Compute (DSec) (arxiv.org)
    55comments
  4. Go Concurrency Distilled (antonz.org)
    13comments
  5. Show HN: Reladraw – A diagram language where you decide where to place things (github.com/reladraw)
    57comments
  6. Turning GLM-5.3-Flash into a Jev-like decision model (privatemode.ai)
    22comments
  7. Evolving programming languages in the AI era (dashbit.co)
    33comments
  8. A searchable library of forgotten public-domain film clips from 1915 onward (movingimagearchive.com)
    25comments
  9. Reverse-engineering the Intel 8087's tangent algorithm: more than CORDIC (righto.com)
    6comments
  10. Drawgent: Coding agent on a live Excalidraw canvas (tangled.org/yanndegat.tngl.sh)
    32comments
  11. Fifteen years later, the Apple Cards origin story (lexontech.org)
    92comments
  12. Biology might not be quantum, but its math is quantumlike (quantamagazine.org)
    8comments
  13. Promising discoveries about the potential for life on one of Saturn’s icy moons (fu-berlin.de)
    17comments
  14. Welcome to the Medical Clinic at the Interplanetary Relay Station (lightspeedmagazine.com)
    9comments
  15. ASML says it sold 'absolutely nothing' in Europe in 2026 (tomshardware.com)
    466comments
  16. The Evolution of Vending Machines (saturdayeveningpost.com)
    4comments
  17. Snap Wants to be a State Actor??–Kansas v. Snap (ericgoldman.org)
    1comments
  18. Generate fonts where every LLM token is the same width (mesh.host)
    7comments
  19. Modern Object Pascal Introduction for Programmers (castle-engine.io)
    60comments
  20. How I changed teaching after AI managed to do all my homework assignments (thelastsoftwareengineer.substack.com)
    146comments
  21. The Lost Atomic Update on Loongson CPU (jia.je)
    6comments
  22. How one Twitch chat message became code execution on a streamer’s PC (scrt.ch)
    11comments
  23. How to keep enjoying programming in a world of LLMs (haskell.org)
    218comments
  24. LA Metro has some of the slowest escalators on Earth (basin.la)
    68comments
  25. Sousveillance (wikipedia.org)
    4comments
  26. HomeBody: A humanoid that explores, remembers, and acts on its own (stanford.edu)
    2comments
  27. Dutch designer made DE9: Closer to the Edit into a playable web-based instrument (creativeboom.com)
    2comments
  28. Reading’s Bayeux Tapestry (diamondgeezer.blogspot.com)
    2comments
  29. Breaking Up with Google Play: Why Conversations Is Now Free (gultsch.de)
    256comments
  30. The Rise of Audio AR (dbreunig.com)
    13comments

Generate fonts where every LLM token is the same width

32 pointsby 1d agoampdot.mesh.host
7 comments
14h agoHN ↗

The kerning is absolutely awful in Safari. Looks fine in Chrome though...

13h agoHN ↗

This page reliably hangs Firefox 155 at 100% CPU for me.

10h agoHN ↗

I did most of my testing in Firefox 156 and it's fine for me

1h agoHN ↗

smooth as butter on 152 in windows 10, as long as we're all reporting

12h agoHN ↗

Haha this did help me generate empathy for the assistant. Neat idea.

7h agoHN ↗

I wonder how this would look in Chinese Mandarin.

2h agoHN ↗

Completely the same as regular text, except punctuations are centered instead of staying at the bottom of the line. CJK Han likely encodes tokens to character one on one. Which brings an interesting question, are Chinese characters more efficient for NLP? In the sense that semantic meaning of a word is not chopped up into partial "tokens".