AIE Europe Keynotes & Coding Agents ft. Pi, Google Deepmind, Anthropic, Cursor, Linear, & more

on YouTube ↗09h 10m

" summary

Gemma 4 ships as small on‑device models with an E2B per‑layer embeddings design and the family already hits millions of downloads; MCP tooling reports ~110M monthly downloads and a push toward programmatic tool calling for lower latency. Cursor replaces whole features—replaced 15,000 lines with a 200‑line Markdown skill—and Pi shows a minimal, extensible harness while warning of agent‑generated technical debt. Hardware and infra choices matter too, e.g. Codex Spark 1,200 TPS enabling instant validation and new fast‑model workflows.

AI-written summary · about

" chapters (24)

1. 0:00Pre-show and opening film
2. 10:40Tejas Kumar opens Day 2 of AI Engineer Europe
3. 15:44Omar Sanseviero (Google DeepMind): Gemma 4's on device capabilities and E2B architecture
4. 31:00David Soria Parra (Anthropic): The future of MCP and programmatic tool calling
5. 49:44Ido Salomon (MCP Apps): AgentCraft and the visual orchestration of multi-agent coding swarms
6. 1:01:05Mario Zechner (Pi): Building the Pi agent and the dangers of AI generated technical debt
7. 1:19:33Armin Ronacher & Cristina Poncela Cubeiro (Earendil): Designing agent legible codebases and embracing friction
8. 1:38:12Benjamin Dunphy: AI Engineer World's Fair announcement
9. 1:44:14Break: Morning coffee
10. 2:26:10David Gomes (Cursor): Replacing 15,000 lines of code in Cursor with Markdown skills and Git Worktrees
11. 2:46:17Matthias Luebken (TAVON): Embedding OpenClaw and Pi into multichannel production environments
12. 3:08:39Sarah Chieng (Cerebras): Adapting developer habits for ultra fast models like Codex Spark (1200 TPS)
13. 3:27:11Lawrence Jones (Incident io): Using AI to evaluate, debug, and manage complex AI systems
14. 3:45:47Luke Alvoeiro (Factory): Architecting long running, multi day agent missions with Factory
15. 4:04:47Break: Lunch
16. 5:41:46Ben Burtenshaw (Hugging Face): Using coding agents for AI Systems Engineering and writing CUDA kernels
17. 6:00:33Michael Richman (Cmd+Ctrl): Curing FOMAT (Fear Of Missing Agent Time) with mobile command and control
18. 6:17:29Liam Hampton (Microsoft): Orchestrating local, background, and cloud agents simultaneously in VS Code
19. 6:35:28Break: Afternoon
20. 7:41:28Tuomas Artman (Linear) with Gergely Orosz (The Pragmatic Engineer): Fireside chat on Linear's design philosophy and Zero Bug Policy
21. 8:10:48Jacob Lauritzen (Legora): Vertical AI and why complex agents need durable UI artifacts over chat
22. 8:25:11Peter Gostev (Arena ai): The "Bullshit Benchmark" and what top models still fail at on LMSYS Arena
23. 8:45:32swyx: Automating a $9M conference business using AI agents for non coding tasks
24. 8:59:02Closing remarks by Tejas Kumar

" related talks

• 05h 56mFrom the main stage to the terminalAja Hammerly, Addy Osmani, Jason Davenport, Omar Sanseviero, Logan Kilpatrick, Philip Kiely, Katie Nguyen, Stephanie Wong, Jay Rodge, Dave Elliott, Ben Gilbert, David Rosenthal, Shubham Saboo, Yasmeen Ahmad, Khulan Davaajav, Tomek Porozynski, Muhammad Farooq, Sam Witteveen · Google Cloud Next '26
• 01h 01mAI Agents In-Depth – Function Calling, MCP and Tool Use Under the HoodAlan Smith · NDC Copenhagen 2026
• 01h 45mAI Engineering with the Google Gemini 2.5 Model FamilyPhilipp Schmid · AI Engineer World's Fair 2025
• 29mTools and Roadmap for AI-Native DevelopersManikantan Krishnamurthy · SuperAI Singapore 2026
• 39mBuild the new generation of AI AgentsGerard Sans · Uphill Conf 2024
• 08mAI-Powered transformation: Driving innovation and reshaping organizationsJoana Carrasqueira · Women in AI Summit 2024
• 27mSoftware engineering with GenAIGergely Orosz · LDX3 London 2025
• 45mPanel Discussion on AI for Code GenerationReact Nexus 2025