cwillu comments on an effective ai safety initiative

[cwillu]

1. Literally does not apply to any existing AI

3. Does so by attacking open source models

1 contradicts 3.