Thread

1. alasano · 1790557209 · focus · HN ↗

   If enjoy AI Pixel Slop, I had made this little Yamanote Line experience to test GPT 5.6 when it came out

   <a href="https:&#x2F;&#x2F;yamanote-window-loop.aljosa-asanovic.chatgpt.site&#x2F;v3" rel="nofollow">https:&#x2F;&#x2F;yamanote-window-loop.aljosa-asanovic.chatgpt.site&#x2F;v3

   Sound on for nostalgia if you&#x27;ve been to Tokyo.

   Edit: forgot I had added a little behind the scenes page

   <a href="https:&#x2F;&#x2F;yamanote-window-loop.aljosa-asanovic.chatgpt.site&#x2F;making-of" rel="nofollow">https:&#x2F;&#x2F;yamanote-window-loop.aljosa-asanovic.chatgpt.site&#x2F;ma...

   1. augment_me · 1790558458 · focus · HN ↗

      Nice! Were the assets generated or imported? Also, if generated, natively or MCP? And which?

      1. alasano · 1790558810 · focus · HN ↗

         All generated, I don&#x27;t normally use the ChatGPT&#x2F;Codex desktop app but in this instance I had it do everything directly.

         So GPT 5.6 Sol using GPT images 2.0 (fake transparency by generating pink which it then removed) and it deployed to ChatGPT sites which I also wanted to just see in action for fun. It&#x27;s basically just CloudFlare with D1 underneath or something.

         So tldr yeah just generated with the built in image gen, it does have an advantage there over Claude. With images 2.5 I imagine the result would be more consistent.

Unofficial Hacker News client; not affiliated with Y Combinator.