Pydantic Monty & Logfire
summary
A minimal Python interpreter written in Rust called Monty runs AI-written code with zero builtin syscalls and requires explicit capability grants to call host functions, supports snapshotting of interpreter state (~1 kilobyte snapshots), and claims microsecond startup (~5 microseconds) so code-mode agents replace many sequential tool calls. Integration with Pydantic AI and Logfire shows demos halving LLM requests, cutting tokens dramatically, and a live bounty program offers $10,000 to find any escape.
AI-written summary · about
chapters (10)
- 0:00Pydantic & Monty intro
- 3:55Traditional tool calling issues
- 13:00Monty minimal interpreter
- 14:24Live Monty demo
- 16:27How Monty works
- 23:50Production with Monty & Logfire
- 25:35Pydantic AI code mode demo
- 29:50E-commerce agent example
- 33:42Logfire prompt optimization
- 34:02Q&A
related talks
01h 04mBuilding AI Applications the Pydantic Way
30mGenerative AI Monitoring with PydanticAI and Logfire
29mRust is the language of the AGI
34mGood or Bad Vibes? Coding with AI in Rust
50mHow to build your own fun and absurd pair programmer
24mDurable Agents: long running AI workflows in a flakey world
59mLightning Talks
32mProgramming language evolution in the AI era