Absolutely love your stack and you should tell more about how you did and write a decent blog post.

And I like and appreciate the idea of improving HN: we all love HN, are addicted and know that there is room for improvement.

But did you make before you started to code any kind of design mockup? In Fireworks or Photoshop? And looked at it a few days? If not you should. It's tempting and often faster to hack and test layouts in Jade and Stylus (or HAML and SASS) but I made the experience that if doing a mockup in Fireworks and just watch and get used to it a few days you get a much better feeling of your product. After a few days you realize bad design choices, start to change and your design and UI slowly matures into something perfect. Usually this takes up to 7 to 14 days.

Current flaws:

- Headers much too small in relation to thumb size and overall layout

- I question the thumbs: do they really offer any benefit? HN articles are often more about text than great visuals and even if i's about images: you'd have just to take one dominating pic of the site and not the whole site

- Too low content density compared to HN

- When showing rankings a single column list always works better than a grid layout because the reader instantly gets who is on #1, #2, #3, etc.