16 comments

  • codingdave 48 minutes ago
    This is cool, and clearly a popular post, yet also... doesn't the 'Show HN' guidelines specifically exclude lists and reading material?

    With the massive drop in quality that AI has brought to Show HNs, this is a great example of something that truly is cool... but would be equally cool without 'Show HN' in front of the title. We're failing to moderate the site well unless we recommend OP does what the guidelines say and "make a regular submission instead."

    https://news.ycombinator.com/showhn.html

    • jonahx 20 minutes ago
      I'm not sure how this would "ruled on", but my take is that the website itself has enough functionality, enough ways to view and slice and dice the info, so to speak, that it can be tried out, and is not merely a "list".
  • gchamonlive 1 hour ago
    Looks a lot alike Web UIs SOTA models produce, and you know what I'm really glad.

    No more people doing crazy stuff with UX just because.

    No more frontend dark magic only the most dedicated zealots can make sense of.

    Democratised UI for all. And this might not even have had an AI assisted development, I'm just abusing this comment and going slightly off topic to say I'm happy for this tech, as someone that only dealt with backend and DevOps/SRE to be able to confidently create useful graphical applications for regular users and learning a ton about frontend architecture in the meantime.

    Btw, a hall of fame for comments would also be interesting!

    • oyster143 54 minutes ago
      This was 95% made with gpt 5.6 sol high

      I actually did the comments internally and will probably post later in the week.

      • gchamonlive 51 minutes ago
        Great work! I'm doing something similar, but using upvoted submissions I did so I can rank posts better, filter out posts about AI but most importantly flag potentially interesting submissions in New that couldn't have made frontpage.
      • georgemcbay 29 minutes ago
        You might have forgot to tell it 'no mistakes' in the prompt?

        The 4 boxes at the top are somehow double loading items on refresh (didn't look at the code, just the output) and generating an awful flicker.

    • basilikum 29 minutes ago
      I find you're perspective quite baffling. It is interesting in that you like the standardization and lack of certain web design crimes and I should not be overly negative to the OP showing their work^1. But man, I truly cannot understand how anyone can see that slop design and feel anything positive. It makes me want to instinctively close the tab immediately as soon as I see it.

      [1] It's mostly not their work and just generated though that's not as relevant. But still I think we should be polite when someone shows something that has the chance of having some significant amount of human effort in it.

    • thi2 21 minutes ago
      And yet the search on mobile opens half cut, the dark mode icon isn't centered like the search icon and in the resurfaced section are overlapping texts. And thats just from glancing through the page on my phone.

      Maybe I'm too picky but it gives a sloppy, unpersonal feeling.

      • gchamonlive 9 minutes ago
        I could take your comment and use it as a prompt and depending on the model quality it will have been an ex-bug
  • jaggederest 8 minutes ago
    Also notable, the internal list that this may or may not derive from: https://news.ycombinator.com/highlights
  • thi2 20 minutes ago
    There is a code freeze coming up at may job, I'll probably browse through some of the links in the spare time. So thanks for curating this.
  • ghosty141 48 minutes ago
    One thing that most LLMs do in their UIs are way too small fonts and too much information density.

    For example, the "max pts" text has no reason to be this small, there is easily enough space to the left.

    • CPLX 3 minutes ago
      This concept has annoyed me so much that I decided to give it a name and create a bunch of shortcuts in my AI tools to stop it.

      I call these things "sprinkles" Because I feel like AI can't stop sprinkling dumb shit all over its designs. Small text that annotates things, little callouts are examples but to me, the canonical thing that AI does when you ask it to build a UI that makes no sense is to create a bunch of stuff on the page that couldn't possibly come from any kind of deterministic database.

      For example, if you ask it to sort and filter a library UI for a book finder, it'll add some element on a SERP page that's like "22 first time novelists" Even though that piece of information doesn't exist anywhere and couldn't possibly wire into a dynamic website.

      Drives me crazy. Thought I'd share.

  • embedding-shape 1 hour ago
    I'm surprised how many of those I've never seen before, even after spending more than a decade on HN and arguably way too much every day. Such a good surprise :)

    I wish the page used the CSS :visited selector and provided a link on the index page, so we could easily see which one we've visited before or not.

    • jperras 1 hour ago
      18 year HN vet here. I've forgotten most of these, even though I was around for nearly every single one of them!
    • oyster143 1 hour ago
      some of the older ones are pure gold
  • artemonster 9 minutes ago
    bookmarked. will never open again
  • mattrighetti 1 hour ago
    Love these websites! Design is very cool as well
  • arjunvrofficial 34 minutes ago
    Good initiative
  • dvaun 1 hour ago
    I was initially skeptical, but after looking through it I see some articles that I distinctly remember (eg Gervais Principle) and blogs like SSC which led me down rabbitholes.

    Lots of good things come from the orange site. I’m sure myself and others will dig through these to find those hidden gems that we missed :)

  • dgeiser13 1 hour ago
    The ole legendary blog post. Get these submitted to the National Register of Historic Blog Posts.
  • fragmede 1 hour ago
    This is cool! Had to scroll way too far to find "wat" but it's a candidate.
  • drumih 1 hour ago
    Looks cool! Some of the listed posts are such a blast from the past
  • bckr 1 hour ago
    Can 3100 links be legendary? I feel like there can be at most 310 legendary links
    • nonethewiser 1 hour ago
      Please elaborate. What issue do you have with the stated methodology?
    • oyster143 1 hour ago
      fair comment, i've probably exaggerated somewhat, but i think there are quite a few links outside of top 300 that are very interesting
    • nonameiguess 1 hour ago
      This was also my first instinct. Wikipedia says the US Baseball Hall of Fame has 351 inductees currently, for a sport that has existed in professional form since 1876. That feels like a more appropriate rate for this kind of enshrinement. Over 3000 in the span of less than 20 years is far too fast. The entire English language literary canon doesn't have that many entries.
      • oyster143 1 hour ago
        on the other hand, there are much more links than baseball players. also there are only around 1k hall of fame links, the rest are candidates (i.e. have not met all requirements - 5+ submissions · 3+ distinct submitters · 3+ active years · 2 qualifying breakout events from different submitters at least 720 days apart)

        i do accept thought that i have exaggerated in the title)

  • toomuchtodo 2 hours ago
    Very cool. Do you have any plans to aggregate book recommendations and papers (via doi perhaps or other link hints) as well?