Re:The first general computer action model
by rio_popper on Monday February 23, 2026 @12:06PM
(#47125137)
Curious about the masked diffusion IDM choice. They mention CTC loss and cross-entropy both underperformed — I'd love to see ablations on that. The claim that typos were "extremely common" with non-causal cross-entropy is interesting but hand-wavy without numbers.
[ Parent
| on HN ]
Re:The first general computer action model
by nee1r on Monday February 23, 2026 @12:15PM
(#47125292)
the main chain of experiments was trying causal => non-causal => non-causal with ctc and CE. i think a good intuition here is that you need a generative approach fundamentally because there definitely are multiple correct IDM labels.