The next 6 to 12 months is going to get very weird. AI is getting very close to AGI or at least close enough… The times they are a changing…

Anthropic’s Claude Mythos is putting in that overtime…

This is about to create a huge shift in AI if regulations start coming through. I am not sure how exactly this can be handled with 100% accuracy given the pure non-deterministic state of Agentic AI

I have literally spent the last six months designing a secure AI proxy and gateway architecture at enterprise scale. This includes AI security controls and AI SOC guidance for monitoring, alerting, and threat hunting. It includes AI use-case governance. It covers six

Opus 4.7 is literally broken. It contradicts itself with every message. It hallucinates, then admits it when questioned, only to hallucinate again. Completely unusable right now.

LMFAO