The AI model that writes text in parallel blocks instead of one token at a time

Article automatically generated from technical news.

Almost every language model you have used works the same way underneath. It predicts one token, appends it, then predicts the next, one at a time, left to right. That sequential nature is why generation feels like watching a typewriter and why speed has mostly been a hardware problem. DiffusionGemma, an open model Google DeepMind released this summer, quietly breaks that assumption, and it is worth understanding even though it did not get the attention it deserved. Instead of writing