Curated by
More in Harness Management
See all 38 →More from Mike Boscia
See all stacks →Your Local AI Harness Can Still Be a Landlord
Local AI harnesses can create a new form of lock-in despite appearing to offer sovereignty, as the harness defaults and configurations determine what an agent can access and do rather than the model weights themselves. The article advocates for Meta AI's Rule of Two principle, which limits agents to holding at most two of three capabilities—untrusted input, private data access, and ability to change things or communicate externally—within a single session.
Built for AI agentsACO · 3222 tokens
Summary
Local AI harnesses can create a new form of lock-in despite appearing to offer sovereignty, as the harness defaults and configurations determine what an agent can access and do rather than the model weights themselves. The article advocates for Meta AI's Rule of Two principle, which limits agents to holding at most two of three capabilities—untrusted input, private data access, and ability to change things or communicate externally—within a single session.
Tags
local-ai · agent-security · model-autonomy · landlord-problem · ai-governance · harness-design
Key entities
Alex Veremeyenko (person, 0.95) · Simon Willison (person, 0.9) · Anthropic (organization, 0.95) · Meta AI (organization, 0.95) · lethal trifecta (concept, 0.85) · Rule of Two (concept, 0.9) · harness (concept, 0.88)
Classification
analysis · language en · status final
Provenance
claude-haiku-4-5 via @stacklist/be@0.1.0, confidence 0.85, 10 Sep 2026