AI Applied: Covering AI News, Interviews and Tools - ChatGPT, Midjourney, Gemini, OpenAI, Anthropic Google's AI Week: Gemini 3.1 Pro, Lyria & Pomelli
7 snips
Feb 24, 2026 They unpack Gemini 3.1 Pro’s performance leap and what it means for competing models. They demo Lyria, Google’s new music generator, and compare it to other creative tools. They explore Pomelli, a product photography tool that turns simple shots into ad-ready images. Practical adoption tips and how these tools affect business workflows round out the conversation.
AI Snips
Chapters
Transcript
Episode notes
Gemini 3.1 Pro Leads The Benchmark Race
- Gemini 3.1 Pro delivered a notable performance jump on hard benchmarks, moving Humanity's Last Exam from 37.5% to 44.4%.
- Jaeden compares it to GPT 5.2 (34%) and Opus (40%), marking Gemini 3.1 Pro as the current top performer among large models.
Small Model Updates Can Change User Experience
- Incremental model updates matter because tools and feature additions can make the same core model feel substantially smarter to users.
- Jaeden highlights embedded tools (example: calculator inside GPT) as the kind of additions that improve usefulness beyond raw model weights.
Integration Beats Raw Usage Numbers
- Model usage counts (monthly active users) can be misleading; integration into productivity tools matters more for long-term impact.
- Conor emphasizes Gemini's advantage because it's embedded in Gmail, Workspace, and other Google services, not just raw app installs.