Track This Theme

Open Weights and Closed APIs Are Diverging at the Worst Possible Layer

As frontier AI capabilities advance, the gap between what closed-API labs will release and what open-weight models make freely available is becoming the central proliferation risk. This theme tracks how the safety architectures of Western frontier labs are being structurally undermined by open-weight releases operating outside their policy frameworks.

2 signals tracked