Latest / Artificial Developer Intelligence
Claude Watermarks, Zuck's Superintelligence Essay, Zed's ZDB & Multi-Agent Turf Wars
Anthropic put AI agents on one computer with conflicting goals. They wrote self-replicating malware and killed each other's processes.
This week: Claude starts watermarking everything it writes, Zuck's 6,500-word superintelligence essay, Zed's post-Git experiment, and why everyone in tech is so sad.
Co-hosts: Shimin Zhang and Dan Lasky. Rahul is away on vacation number two — one apparently wasn't enough.
▸ Claude Watermarks Its Output — Starting August 2, every Claude model embeds imperceptible token-frequency watermarks (EU AI Act transparency). It defeats the lazy slop grenade, but removal…
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.