Thread

1. famouswaffles · 1790178993 · focus · HN ↗

   What did they do to Astra so cracked at vision (and computer use). That ARC 3 score turned out to be no joke/fluke. That huge gap between Astra and Fable (in this case) is basically every hard vison/spatial benchmark i've seen including non-benchmarks like playing games (Portal, Factorio, RimWorld).

   SpatialBench - <a href="https:&#x2F;&#x2F;x.com&#x2F;spicey_lemonade&#x2F;status&#x2F;2096365630190698516" rel="nofollow">https:&#x2F;&#x2F;x.com&#x2F;spicey_lemonade&#x2F;status&#x2F;2096365630190698516

   ZeroBench - <a href="https:&#x2F;&#x2F;zerobench.github.io&#x2F;" rel="nofollow">https:&#x2F;&#x2F;zerobench.github.io&#x2F;

   Robot Arms - <a href="https:&#x2F;&#x2F;openai.robocurve.org&#x2F;gpt-6-astra&#x2F;">https:&#x2F;&#x2F;openai.robocurve.org&#x2F;gpt-6-astra&#x2F;

   1. smusamashah · 1790180010 · focus · HN ↗

      I think Opus 5.5 is at same level now. I have seen too many videos made by Opus 5.5 today on twitter.

      <a href="https:&#x2F;&#x2F;x.com&#x2F;victormustar&#x2F;status&#x2F;2102707412704919910" rel="nofollow">https:&#x2F;&#x2F;x.com&#x2F;victormustar&#x2F;status&#x2F;2102707412704919910 horse galloping pixel art

      <a href="https:&#x2F;&#x2F;x.com&#x2F;LexnLin&#x2F;status&#x2F;2102133072585965759" rel="nofollow">https:&#x2F;&#x2F;x.com&#x2F;LexnLin&#x2F;status&#x2F;2102133072585965759 moving train pixel art animation

      <a href="https:&#x2F;&#x2F;x.com&#x2F;jkeatn&#x2F;status&#x2F;2102441348075057539" rel="nofollow">https:&#x2F;&#x2F;x.com&#x2F;jkeatn&#x2F;status&#x2F;2102441348075057539 painting with code

      <a href="https:&#x2F;&#x2F;x.com&#x2F;LCSlates&#x2F;status&#x2F;2102503027340988559" rel="nofollow">https:&#x2F;&#x2F;x.com&#x2F;LCSlates&#x2F;status&#x2F;2102503027340988559 video, very detailed prompt though

      <a href="https:&#x2F;&#x2F;x.com&#x2F;aj_dev_smith&#x2F;status&#x2F;2102504509637587339" rel="nofollow">https:&#x2F;&#x2F;x.com&#x2F;aj_dev_smith&#x2F;status&#x2F;2102504509637587339 generated song&#x2F;music with code

      <a href="https:&#x2F;&#x2F;x.com&#x2F;aj_dev_smith&#x2F;status&#x2F;2102575577563570450" rel="nofollow">https:&#x2F;&#x2F;x.com&#x2F;aj_dev_smith&#x2F;status&#x2F;2102575577563570450 another song

      1. famouswaffles · 1790183379 · focus · HN ↗

         I don&#x27;t think so. These are cool but all of this is code to x. I&#x27;m talking about actual computer control.

         Stuff like: - <a href="https:&#x2F;&#x2F;x.com&#x2F;iam_zachi&#x2F;status&#x2F;2095992132620136677" rel="nofollow">https:&#x2F;&#x2F;x.com&#x2F;iam_zachi&#x2F;status&#x2F;2095992132620136677

         Puzzles, games, painting software, robotic control and now driving. I haven&#x27;t seen any other model fire on all cylinders like that.

Unofficial Hacker News client; not affiliated with Y Combinator.