
AI Pod by Wes Roth and Dylan Curious | Artificial Intelligence News and Interviews With Experts Google's INSANE new AI Agent
7 snips
May 8, 2026 They unpack Google’s push for a Gemini-powered personal agent that acts continuously and takes actions for users. The conversation jumps to agentic coding tools, how agents tie into Search, Android and Chrome, and implications for work and software. They debate privacy and safety trade-offs, competition among big AI players, and why agents could reshape everyday tech use.
AI Snips
Chapters
Transcript
Episode notes
Robots Rely On A Data Flywheel
- Household robots will scale via a data flywheel where early, imperfect units provide crucial real-home data for rapid improvement.
- Wes Roth explains companies deploy robots at loss to collect user-task data, then improve firmware and models to leapfrog competitors.
Unfinished Desk Shows Real Demand For Home Robots
- Wes Roth shares his unfinished standing desk story to illustrate mundane tasks people procrastinate on and would gladly outsource to robots.
- He describes the desk left 90% assembled for weeks and imagines a robot finishing it instantly.
Robotic Slop Comes Before Dexterous Robots
- Many early household robot tasks will be "robotic slop" — low-dexterity, repetitive chores that still deliver high user value.
- Dylan Curious and Wes Roth stress simple tasks like gathering clutter or basic tidying are immediately useful while fine dexterity lags.
