Leveraging AI 249 | Fast-takeoff fears, $1 B Disney-OpenAI pact, GPT-5.2’s pro-grade leap, Gartner yells “block AI browsers,” and Apple bleeds AI talent—our mega AI recap for the week ending on December 13, 2025
97 snips
Dec 13, 2025 Exciting developments in AI are unfolding! The recent release of GPT-5.2 promises to outperform humans in various tasks. There's buzz over a $1 billion partnership between OpenAI and Disney, focusing on AI-generated content. Recursive Self-Improving AI raises governance concerns, with industry leaders urging caution. Meanwhile, users are pushing back against perceived ads in ChatGPT. The job market is shifting as the Fed acknowledges AI's impact, and Apple's AI talent exodus poses competitive risks.
AI Snips
Chapters
Transcript
Episode notes
GPT-5.2 Is Built For Real Work
- GPT-5.2 is explicitly tuned for real-world knowledge work, long context, agentic tool use, and vision tasks.
- OpenAI positions it as a professional-grade model to unlock economic value across spreadsheets, presentations, and multi-step projects.
Agents Move From Fun To High-Value Work
- Perplexity + Harvard found agentic browser use shifts from casual tasks to productivity and research as users gain trust.
- Over time agents are used mainly for productivity workflows, learning, and complex research rather than only entertainment.
Copilot Patterns Mirror Human Routines
- Microsoft found copilot usage patterns vary by device and time: mobile skews health and late-night philosophical queries, desktop skews business work.
- Usage peaks align with human routines: coding on weekdays, gaming and exploration on weekends.