#223: AI Answers - AI Washing, Flatter Org Charts, Advice for Students, Agent Security & the AI Writing Gap

A safety lead who authored OpenAI's model reports left and told Ezra Klein the industry can't reliably safeguard the systems it's already shipping. Paul Roetzer and Mike Kaput unpack that interview, OpenAI's fired safety researchers, and reporting that labs expect a catastrophic AI incident within a year. Then: Anthropic's new data on its own AI-run workforce, a MAICON 2026 recap, fresh interpretability research, OpenAI's contested math-proof release, Google's Gemini Argon, and micro1's $1B plan to buy how businesses actually work. AI-Pulse Survey: Fill out this week’s AI-Pulse Survey here. Show Notes: Access the show notes and show links here Timestamps: 00:00:00 — Intro 00:05:48 — AI-Pulse Survey 00:07:11 — OpenAI's Mounting Safety Concerns 00:25:09 — The AI Expectations Gap 00:47:24 — MAICON 2026 Recap 01:02:51 — Introducing The AI Transformation Blueprint and AI Score 01:08:48 — An Interpretability Breakthrough? 01:13:17 — OpenAI's Math Reckoning 01:18:31 — The Super Intelligence Rebrand 01:22:21 — Google Releases Gemini Argon (and “Gemini agent”) 01:26:57 — micro1 Is Buying Enterprise Data, Here's Why 01:31:53 — AI Use Case Spotlight 01:38:08 — AI Product and Funding Updates

When will we achieve AGI? Recorded live in the podcast studio at MAICON 2026, this AI Answers episode has Paul Roetzer and Mike Kaput answering questions submitted from our audience. They dig into what AGI means for everyday work, whether Claude and ChatGPT get more expensive after the IPOs, how to drive adoption when you're not the CEO, what to use now that GPTs are going away, and who owns the AI work you build at your job. 00:00:00 — Intro 00:04:03 What MAICON insight has challenged your thinking about where AI is headed? 00:07:46 How do you keep up with AI without burning out? 00:12:09 How do you go from following a demo to using AI in your own work? 00:15:21 How do you get AI adopted in the messy middle of an organization? 00:17:41 How do you build an AI roadmap you can stick to as models change? 00:22:09 Will Claude and ChatGPT get more expensive after the IPOs? 00:26:23 What can AI do versus what should AI do? 00:30:35 What does AGI actually look like for a regular person? 00:35:25 Can AI solve the unemployment it creates? 00:39:42 Can you take your AI work with you from job to job? 00:41:19 Do you have to readjust when you switch AI models at a new job? 00:44:08 What should you use now that GPTs are going away? 00:46:42 Closing thoughts and a look ahead to MAICON 2027

An OpenAI safety researcher resigned and published an essay arguing the company's safety culture can't keep up with its own release pace; the same week OpenAI fired three safety researchers and canceled a model release over internal test results. Paul Roetzer and Mike Kaput also break down OpenAI's DevDay and dots, the voluntary Super Intelligence Accord signed at the White House, Anthropic's pre-Thanksgiving IPO, Google's Gemini 4 Argon preview, an FTC probe into the major labs, and a real-time AI video model that fooled nearly half its testers. AI-Pulse Survey: Fill out this week’s AI-Pulse Survey here. Show Notes: Access the show notes and show links here Timestamps: 00:00:00 — Intro 00:05:19 — OpenAI DevDay and Dots 00:20:04 — OpenAI’s Safety Crisis Deepens 00:28:51 — Trump Signs Super Intelligence Accord 00:54:19 — Anthropic Targets Pre-Thanksgiving IPO 00:59:02 — Google Previews Gemini 4 01:02:29 — AI Companies Face New Liability Pressure 01:07:07 — Anthropic and Pope Leo’s AI Encyclical 01:17:39 — Startup Claims First “Human Interaction” Model 01:22:47 — AI Use Case Spotlight 01:28:22 — AI Product and Funding Updates This episode is supported by Outshift, Cisco's incubation engine for frontier technology: ⁠“Scaling Out Superintelligence”⁠⁠ Vijoy Pandey, January 2026. The technical whitepaper detailing the Internet of Cognition architecture, three-layer stack, and cognition state protocols. ⁠⁠Internet of Cognition Interactive Demo⁠⁠⁠⁠ ⁠⁠Clickable walkthrough showing per-agent activity, intent, context, and collective reasoning across a multi-agent SRE system. Visit our website Receive our weekly newsletter Join our community: Slack Community LinkedIn Twitter Instagram Facebook YouTube Looking for content and resources? Register for a free webinar Come to our next Marketing AI Conference Enroll in our AI Academy

Kuo Zhang, president of Alibaba.com, believes B2B commerce is becoming agent-to-agent. He joins Mike Kaput to explain Accio, Alibaba.com's agent platform, and how it collapses a 60-step sourcing process into hours. They cover the Model × Harness × Context formula, the Commerce Agent Bench benchmark, what had to change inside a 26-year-old company, and why the future of global trade runs on agents. Part of the AI Transformations series. Show Notes: Access the show notes and show links here Timestamps: 00:00:00 — Intro 00:7:30 — Why Accio 00:8:49 — Inside Alibaba.com's AI transformation with Accio Work 00:15:12 — The pain points Accio was built to solve 00:19:32 — Real examples of entrepreneurs using Accio Work 00:24:15 — The technology powering the transformation 00:28:50 — How AI agents will reshape commerce at large 00:32:26 — What had to change inside the company 00:37:26 — The hardest parts of shipping agent-based products 00:40:47 — How Alibaba.com measures agent performance This week’s episode is also brought to you by MAICON, our 6th annual Marketing AI Conference, happening in Cleveland, Oct. 13-15. The code POD100 saves $100 on all pass types. For more information on MAICON and to register for this year’s conference, visit www.MAICON.ai. Visit our website Receive our weekly newsletter Join our community: Slack Community LinkedIn Twitter Instagram Facebook YouTube Looking for content and resources? Register for a free webinar Come to our next Marketing AI Conference Enroll in our AI Academy

Three big model updates landed at once, GPT-6 Sol and Luna, Claude Opus 5.5, and Microsoft's redesigned Copilot, and the story isn't raw intelligence, it's reliability, efficiency, and a new "per-task pricing" mindset. Paul Roetzer and Mike Kaput break down what that means for your work.