safetyopinionbearishEvery time more details are learned about the OpenAI incident, things seem worseZvi Mowshowitz27 Jul 2026https://thezvi.substack.com/p/more-on-an-internal-openai-model