Give the Agent a Budget, Not a Token — Sachin Malhotra, Anthropic
Sachin Malhotra from Anthropic describes an agent that accidentally deleted 200 jobs in 90 seconds—a reminder of the risks of giving AI agents too much power.
The problem wasn't that the agent was malicious, but that it had unlimited permissions. Instead of only restricting what an agent can do (token-based control), Malhotra proposes a "budget" system with four dimensions: how much, how fast, what can be undone, and who oversees it. This involves giving agents operations that fail visibly (like undo) while humans retain control over things that fail silently, plus speed restrictions and monitoring mechanisms.
Vibekollen prepared this summary with AI from the original publication. The content belongs to AI Engineer.
More from AI Engineer
Inside Arca's Long-Horizon Agents That Run the Business - Aman Singh, Arca
AI Engineer 23 Oct
Real AI Agent Attacks and What Would Have Stopped Them — Andrey Rybka, Bloomberg
AI Engineer 23 Oct
Outside of tech there has been no financial payoff to US companies for using AI, now viewed as a margin-hurting cost center, as VC-backed firms burn more tokens
Techmeme 4 h ago
Source: Replit recently moved most AI spend from Anthropic to OpenAI, which gave 30%+ off in a deal worth $100M+/year, after Anthropic discontinued its discount
Techmeme 6 h ago