Beyond the Lethal Trifecta: Agentic Commerce on the Open Internet — David Levine, Kiduna Club
- Source
- youtube.com
- Author
- AI Engineer
- Date
Context
An identity could help connect an automated action to the organization responsible for it. The proposal in this talk starts from what Simon Willison named the lethal trifecta: access to private data, exposure to untrusted content and the ability to communicate externally in a way that could steal that data. The speaker says that combination is why enterprises keep agents inside closed platforms and give up the they lose there.
The proposed answer is identity. According to the talk's description, agents would carry JWT-based identity credentials that resolve through a registry to an organization, much as DNS resolves a name, with an audit trail underneath. That would make an agent's activity traceable to a responsible organization. The description does not explain how this stops an agent from being tricked by untrusted content, and an identity record does not by itself establish that; this is an accountability proposal, not a proven defense against .
- AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
- context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
- prompt injection — An attack that hides instructions in content an AI will read — a webpage, email, or document — tricking it into following the attacker instead of the user.
videoMemory-Aware Agent Harnesses for Anti-Money Laundering — Nacho Martinez, Oracle
videoThe AI Bill Is Too Damn High: Tokenomics 101 — Eugene Cheah, Featherless AI
videoStop Chatting With Spreadsheets — Francesco Galletta & Felippe Felisola Caso, Monzo
videoWhy One AI Agent Isn’t Enough (And How to Run 1,000) - CJ Avilla & Harrison Stall, Anthropic
Checking sign-in…
Loading comments…


