Images vs. Editable Layers Problem

Beth opened episode 601 with Andy joining early and Karl arriving later. The show kicked off with browser based agents, Google’s Nano Banana expansion into Workspace, and a live demo of Slides using AI to beautify content. From there, the conversation shifted toward the limitations of Gemini generated infographics, the need for human oversight, the rise of agent powered browsers, and early signals about OpenAI’s new hardware team. The hosts explored cultural pushback against wearable AI, the gap between real world adoption and tech hype, and the long term impact of AI on management skills, jobs, and public trust.

Key Points Discussed

Perplexity’s Comet agent comes to mobile with full web action support

Google rolls out Nano Banana AI in Docs, Slides, and Notebook LM

Gemini 3 image models still make factual mistakes in diagrams and labels

Google confirms layered image editing is on the roadmap

Manas launches a browser operator extension that turns Chrome into an AI agent

OpenAI builds a hardware division and hires dozens of Apple engineers

Public resistance grows against AI wearables like the Friend pendant

Western media messaging reinforces AI as a threat, slowing adoption

Singapore’s AI rollout reveals a management and leadership gap

Human interpersonal skills emerge as a key competitive advantage

Robotics accelerates as Google DeepMind hires Boston Dynamics’ former CTO

Visionary hardware concepts likely push toward AI native devices with voice first design

Sora, agent tools, and multimodal models still struggle to break into mainstream awareness

Timestamps and Topics

00:00:00 👋 Opening, Thanksgiving week, Andy joins

01:01:00 🤖 Perplexity Comet mobile agent overview

02:21:00 📝 Nano Banana comes to Google Workspace

03:12:00 🎨 Slides demo with AI generated infographics

05:04:00 🚗 Andy reviews Nano Banana Pro car diagrams and labeling errors

08:43:00 🧩 Discussion on image limitations and lack of editable text layers

11:49:00 💬 Community notes, Google confirms layered images are coming

14:07:00 🧭 Karl joins, new browser operator from Manas

16:00:00 🛠️ OpenAI’s hardware division poaches Apple engineers

17:40:00 📱 What an AI native device might look like

21:08:00 🚇 Anti AI backlash, Friend pendant ads defaced in Chicago

22:52:00 🌍 Western fear framing versus Asian AI optimism

24:01:00 📉 Media narratives shape public adoption and trust

27:03:00 🇸🇬 Singapore as a case study in AI driven workforce disruption

29:15:00 👔 Management skills become a rare and valuable human advantage

33:23:00 🤝 Interpersonal skills and face to face client work outcompete automation

34:59:00 🔄 AI agents cannot replace real rapport and live collaboration

38:59:00 🤖 DeepMind hires Boston Dynamics CTO to build robot capabilities

41:12:00 🗣️ Future devices shaped around voice first AI

45:15:00 ❓ Growing public “why would you build this” skepticism

48:34:00 🧩 Designing use cases that actually solve problems

52:28:00 📰 Upcoming stories this week: OpenAI internal memo, Meta updates