AI Snips
Chapters
Transcript
Episode notes
Gemini 3 Pushes Multimodal, Factful AI
- Google launched Gemini 3 Pro as a natively multimodal, more factual, and reasoning-focused flagship model.
- Brian McCullough highlights Google pushing richer visual and interactive responses across Search and the Gemini app.
Agentic Models Aim To Do Real Tasks
- Gemini 3 Pro can process text, images, and audio simultaneously and create visual magazine-style interfaces.
- Google says the model shows improved reasoning, reduced sycophancy, and agentic capabilities for multi-step tasks.
Try Antigravity For Agent-First Coding
- Try Antigravity for agent-first coding workflows during Google's public preview on Windows, macOS, and Linux.
- The tool reports artifacts like plans and recordings to help you verify agent work as it runs.