19 August 2026

Researcher tests small AI models as code sandboxes

First reported

Simon Willison ran this on 19 August.

Simon Willison used Claude Fable 5, a smaller version of Anthropic's Claude chatbot, to test running untrusted Python and JavaScript code safely.
The experiment explored whether small AI models could serve as sandboxes, isolated environments where potentially dangerous code runs without harming the main system.
The testing used Claude Code for web, a tool that lets Claude write and execute code within a browser.

How it was covered

Simon WillisonDaily notes and links

Simon Willison conducted research using Claude Fable 5 in Claude Code for web to explore whether smolmachines/smolvm could serve as a sandbox for running untrusted Python and JavaScript code safely.