Lance Martin

Lance Martin works on agent systems at Anthropic and speaks publicly about harnesses for reliable, secure long-horizon work. His talks connect model capability to the execution environment around it.

Original profile or publication ↗

Role comes from the linked Anthropic talk description; portrait comes from the linked public GitHub profile. Checked September 19, 2026.

Referenced in 101

1. Managed agents 101

   Claude for Long-Horizon Tasks ↗

   Long-horizon reliability comes from the harness and environment around Claude: decomposition, state, feedback, security and recovery.

About this page

This is a Mudpie reference profile, not an account, endorsement or claim that Lance Martin contributed to Mudpie. The profile identifies an author, editor, maintainer, speaker or source publication connected to the linked work; it is not necessarily the article byline. Summaries are Mudpie's; the linked work remains the source of truth.