Replay: Open Source Friday Cua with Francesco Bonacci
Cua Computer Use agents AI agents Open source Lum Policy Driven agents GitHub Open Source Friday Agent benchmarks Francesco Bonacci Coding agents AI agent security
Francesco Bonacci, CEO of Cua, discusses the openly-sourced Cua ecosystem for building computer-use agents, covering Lum, Cua Driver, and their benchmarking approach. The conversation explores the shift from screenshot-based computer use 1.0 to more capable policy-driven agents, with insights on security, evaluation, and how the team builds with coding agents. Developers and AI researchers interested in agent infrastructure and open-source AI projects will find this valuable.
Related videos
Stop letting AI write code until you fix this one thing (with Antigravity) | The Agent Factory
The Download: GitHub Copilot app, Anthropic news, and Minecraft
Vibecoding your Own Multi-Agent Workstation
Draw in your terminal! tmux + Terminal Maxxing w/ Ben Vinegar
Microsoft Dataverse plugin: unleashing coding agents on the enterprise | OD849
Post-Training and Deploying Open Source Reasoning Models in Foundry | DEM321
Observe and control agents across any framework with open source tools | BRK250
Open Source Friday: Governance for AI agents with Imran Siddique
Extending AI Agents with WebAssembly
Open Source Friday: Mozilla AI | Project cq
Open Models Coding Essentials – Running LLMs Locally and in the Cloud Course
