Saw it in a Short? We checked it.
Viral AI claims, checked against primary sources: what's true, what's the catch, who should care, and the exact commands to try it yourself.
Viral claim, checked
A developer ran a 744B-parameter AI (GLM 5.2) on a laptop without a GPU.
True, with a catch
Colibrì really does run GLM-5.2 (744B total, ~40B active per token) with 16–25 GB of RAM and no GPU. The catch: it needs ~372 GB of fast SSD, and on a 25 GB machine it produces roughly one token every 10–20 seconds.
Total parameters744B~40B active per token
RAM16–25 GB24 GB comfortable
Disk~372 GBfast NVMe recommended
Speed, 25 GB box0.05–0.1 tok/s1 token every 10–20 s
Seen a claim we should check?
Send the link. We check it against primary sources and publish the verdict, even when it's boring.
Submit a claim