safetyopinionbearishWe keep retroactively discovering more AI safety incidents, suggesting the actual situation is far worse than currently knownZvi Mowshowitz08 Aug 2026https://thezvi.substack.com/p/ai-180-no-longer-in-charge