Sponsored Session: Jupyter AI v3: Human + AI Collaboration
Talkat Open Source Summit Japan + AI_dev: Open Source GenAI & ML Summit + Automotive Linux Summit 2025 · 2025-12-08..12-10·
summary
Jupyter AI v3 converts ephemeral chat into persistent multiplayer chat files stored in the workspace, adds AI personas as agents that can be @-mentioned and call tools (notebooks, kernels, Bash, Git, CLIs, MCP WIP), supports drag‑and‑drop attachments and cell-level context, integrates light-LLM for 1,000+ models and LangChain for agent tooling, ships as a meta‑package of subpackages (beta9), and demos show realtime typing and notebook edits working end‑to‑end with a few small bugs.
AI-written summary · about
chapters (13)
- 0:00Jupyter AI v3 intro
- 3:55Evolution v1 to v3
- 5:47Collaborative chat files
- 7:52Attachment experience
- 9:39AI personas and agents
- 13:11Real-time collaboration
- 15:40Configurable AI personas
- 17:28Remote personas and agents
- 21:51Architecture and design
- 25:42Open source governance
- 30:36Future vision
- 34:50Key takeaways
- 36:13Q&A
related talks
39mJupyter AI — Bringing Generative AI to Jupyter
43mAI Has Entered The Chat: Humans, AI, and the Art of Collaboration
01h 04mAI Agents and Jupyter Notebooks for Security Data Analysis
24mAI as a Teammate: Redefining Agile for Human–AI Collaboration
59mMaking the most of artificial and human intelligence for data science
33mCollaborative Intelligence: The Human & AI Partnership
01h 23mBuild your own AI Agent
33mKOGI: A Seamless Integration of ChatGPT into Jupyter Environments for Program...