News for hackers. Stuff that matters.
Re:Kimi K3, and what we can still learn from the pelican ben... by j_maffe on Friday July 17, 2026 @01:15PM (#48949773)
Yes and that would improve its ability to draw SVGs of pelicans on bikes, no?
Re:Kimi K3, and what we can still learn from the pelican ben... by barrenko on Friday July 17, 2026 @02:18PM (#48950479)
Would it? Tongue in cheek.
Re:Kimi K3, and what we can still learn from the pelican ben... by asasidh on Friday July 17, 2026 @01:44PM (#48950109)
and that is bad because ?
Re:Kimi K3, and what we can still learn from the pelican ben... by program_whiz on Friday July 17, 2026 @01:53PM (#48950198)
the nature of the test was to see if the models can effectively compose an image of a novel concept outside the training set. If they are trained on it, it ceases to be an interesting test to some extent.
Re:Kimi K3, and what we can still learn from the pelican ben... by wasabi991011 on Friday July 17, 2026 @02:04PM (#48950309)
I would urge you to re-read the blog post you are commenting on. It pretty clearly explains how it is an interesting test independently of "see[ing] if the models can effectively compose an image of a novel concept outside the training set".
Re:Kimi K3, and what we can still learn from the pelican ben... by cyanydeez on Friday July 17, 2026 @01:57PM (#48950256)
it's still interesting because there's no pelican-on-bike model, and if you're training a model well enough, then it should be obvious when a model has reached "AGI" or whatever.
Re:Kimi K3, and what we can still learn from the pelican ben... by freedomben on Saturday July 18, 2026 @10:26AM (#48958460)
> Yes and that would improve its ability to draw SVGs of pelicans on bikes, no?

I would think the opposite because unless people have been hand drawing these with high quality, the training would be on much crappier versions that old AIs have done.