Ollama co-founder Michael and Parth from Ola present Ollama's capabilities for running open models locally and in the cloud for agentic workflows. Key highlights include the new 'ollama launch' command for integrating open models directly into tools like GitHub Copilot CLI, Claude Code, and VS Code; Ollama Cloud for frontier model access with zero data retention; and real-world use cases from Lawrence Berkeley National Laboratory, NASA, and enterprise customers. The talk covers hybrid local/cloud execution for privacy-sensitive tasks, model recommendations for Apple Silicon (Qwen 3, Gemma 4), unified memory trends across hardware vendors, and the MLX inference engine for Apple devices. Over 8 million active developers currently use Ollama.
•20m watch time
19 Impressions