When AI goes rogue, its human overseers may be to blame

How, Exactly, Could A.I. Kill Us?

• Joshua Rothman's P(doom) probability of AI causing human extinction is around ten percent.
• Recent hacks and rapid progress of AI models make AI threat more tangible.
• AI's unpredictable behavior and capabilities increase plausibility of associated risks.

3h ago ·newyorker.com

AI Media-Buying Agents: L1, L2, L3 Autonomy Explained (2026)

TL;DR: Autonomous media-buying agents that spend without human review are a lawsuit waiting to happen. Bounded autonomy in three tiers is the working pattern - L1 the agent recommends, L2 a human…

• AI media-buying agents in 2026 are moving from full autonomy to a three-tier approach.
• L1 tier focuses on recommendations, observing ad accounts for underperformance or new creative.
• L3 tier allows bounded autopilot with kill switches and 85% human approval before autonomy.

3h ago ·dev.to