Comment by bryan0

Dario has been pretty clear and consistent that:

1. AI model safety is the most important thing.

2. open weights decreases model safety

So it seems unlikely that this suggestion would be well-received