← Browse

World's Fair 2026

Agents Need Feature Flags - Sachin Gupta

Sachin Gupta

Overview

This talk argues that AI agents, capable of performing high-impact actions like sending money or modifying databases, are being deployed without the safety infrastructure common in traditional web development. It introduces six specific types of feature flags—prompt variants, tool access, model routing, memory policy, autonomy level, and kill switches—that are necessary to manage the unique behavior surfaces of agents and prevent incidents. The core thesis is that applying established web development disciplines, particularly robust kill switches and phased rollouts, is crucial for controlling and safely iterating on AI agent functionality.

Who should watch

Key takeaways

Notable quotes

*We are shipping them the way web team used to ship in 2008.*
*The boring infrastructure that keep deploy safe is already a solved problem.*
*Ship the kill switch first. If you do nothing else, give your agent one agentwide kill switch and one per tool kill switch.*

Watch on YouTube →

Unofficial community note. Prefer the recording for nuance.