Five pieces a week, each with my own commentary. Not a news ticker: the commentary is the value and the link is the proof. 20 editions so far, 99 pieces, all still readable.

Five pieces on liability for agents. A lawsuit over some 700 OpenAI agents, the first known one to lean on California’s new rule. A former computer-crime prosecutor who reads the federal hacking law as probably missing the case. An announced bill whose first part covers operators. An industrial security firm on two utilities where attackers used AI, and on why most plants could not tell afterwards what happened. And a German product liability law that covers software and AI from December, but not damage between companies. Curated and commented, not aggregated.
Related: Agentic AI — whether it pays, build vs buy, and where you stand →
The reading list by email once the mailing starts: five pieces, my commentary. Unsubscribe anytime.