Billion-dollar platforms rush to block AI agents after one rogue LLM broke customer data

IgorCarron · x · 2026-11-19

oanaolt spoke this week with executives at platforms each clearing $1B+ in transactions, and the same request kept coming up: help us block AI agents. The trigger was an incident where a customer handed an LLM their own login to update records; it modified data it should never have touched, broke downstream reporting for other customers, and cost days of engineering time to trace. Terms of service already banned it — nobody reads those.

Every platform reached the same conclusion: close the gates — no agents unless authorized, then build one sanctioned path and force everyone through it. IgorCarron's repost ("And they fall...") signals that indiscriminate OpenAPI-style openness is closing, and agent governance/authorization infrastructure is becoming a platform necessity.

Original post →

More from coding & agent

coding & agent channel →