I find it really hard to see how productive a collective pause and 'think' about something so inherently unpredictable will be.

> "implement a set of shared safety protocols for advanced AI design and development that are rigorously audited and overseen by independent outside experts"

Who deserves to be called an expert on this? Feels like Economics or something, where you have camps of thought advocating for themselves but little way of knowing who's right. Best to break things and develop antibodies whilst the stakes are still low.