News for hackers. Stuff that matters.
Re:Claude Opus 5.5 by adverbly on Tuesday September 22, 2026 @02:15PM (#49805746)
> The differences between the pelicans aren't huge, but the xhigh one has a better beak.

If you look carefully, everything except the last pelican has the two legs both in front of the crossbar as if the legs are all on one side of the bike.

The last pelican gets this correct.

Re:Claude Opus 5.5 by DenisM on Tuesday September 22, 2026 @05:11PM (#49808176)
I’ve been paying attention at this exact detail.

Misplaced legs clearly indicate lack is spatial reasoning - the llm can reason about verbal idea of a bicycle but not about the actual object. The fact that this model got it correct gives me a pause. Did they figure out spatial reasoning? Or did this complain trickle down to the training set?

Re:Claude Opus 5.5 by mjhagen on Tuesday September 22, 2026 @06:28PM (#49809148)
6 Astra Max is the only other model I’ve seen get this right.
Re:Claude Opus 5.5 by amativos on Wednesday September 23, 2026 @09:05AM (#49815476)
Interestingly, Astra Medium got it right as well.