I gave a Gemma 4 AI agent a sandbox and it taught itself physics | The Agent Factory Podcast
Gemma 4 Google deepmind Open source AI model Ai agent development Function calling Model deployment Self Hosting On Device AI Google cloud platform Mcp server Llm reasoning Apache 2.0 license
This episode explores Gemma 4, Google's latest open source model, through live demos including a food tour agent using MCP servers and a striking demo where the model independently reasons through physics to code a bouncing ball animation in Python. Targeted at developers and cloud architects, it covers practical deployment strategies in Cloud Run, the Apache 2.0 licensing benefits, and how to leverage features like native function calling and structured JSON output for building efficient, private AI agents.
Related videos
Develop and integrate AI agents with Google Workspace
Your AI agent is forgetful. Here’s how to give it a brain.
Gemma 4 production stack: Model Armor, ADK Agents, Tracing
Self host Gemma 4: Deploy LLMs on Cloud Run GPUs
Google just casually disrupted the open-source AI narrative…
AlloyDB Lakehouse Federation: Unified access to BigQuery and Google Cloud Lakehouse
Build an AI agent app with Go ADK, Cloud Run, and Flutter
Build full-stack apps with Google AI Studio, Cloud Run, and Cloud SQL
Agent-first workflows from prompt to production
From the I/O main stage to the terminal
The future of Cloud AI: Mastering MCP servers, Gemini, and agentic workflows
