Hacker News

Favorites Setup
☰
Comment by demibabs | original | Hy4 preview
[−]demibabs · 2026-08-30 Sun 00:35 UTC · link
No one’s talking about how good the final product is.

Edit: someone else commented that as I was typing this, lol.

[−]blackhaz · 2026-08-30 Sun 07:29 UTC · link
I wonder, do we need a new benchmark? There's quite a bit of feedback data floating around about pelicans on bicycles already.
[−]ziofill · 2026-08-30 Sun 09:05 UTC · link
That’s a fair question, but it seems that it’s not yet necessary. See here

https://dylancastillo.co/posts/pelicanmaxxing.html

https://simonwillison.net/2026/Jul/22/

[−]stymaar · 2026-08-30 Sun 15:50 UTC · link
I don't think this argument is a good one though, as it would be quite natural for a lab rhat want to macimize the performance of their model on the pelican bench to train it for “text-to-svg simple image generation” rather than just “pelicans on bicycle”.