The Artificial Intelligence Show

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. Sh

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

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

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 Acc

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. Then: Jensen Huang's case against the AI doomers from his Ezra Klein interview, a first look at AI Score by SmarterX, and a rapid-fire run through Meta Connect, Amazon blocking Muse, OpenAI's rogue-agent disclosures, and more. AI-Pulse Survey: Fill out this week’s AI-Pulse Survey here. Show Notes: Access the s

One task at Baptist Health used to take about 10 hours a week. Now it takes roughly two minutes, and that's just one piece of a two-year AI transformation. Christine Kotler, Chief Marketing & Communications Officer of Baptist Health South Florida, joins Mike Kaput to walk through how her marketing and communications team built an AI adoption model from the ground up: a structured learning curriculum, the peer-coaching "AI Sherpa" model, real use cases, and the honest friction of doing this inside a regulated healthcare organization. Part of the AI Transformations series, presented by Google Cl

AI-safety pacing became a political fight this week: Trump versus a bipartisan push, with data centers and the midterms in the middle. Paul and Mike go deeper on the part that matters more: the widening gap between the models labs keep and the ones you can use, Noam Brown's warning about it, and OpenAI's six new rogue-agent disclosures. Plus Paul's surprise book announcement and a full rapid-fire slate. 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 00:04:29 — The AI Slowdown Gets (Very) Political 0

David's Bridal was built around one thing for 75 years: the dress. In this AI Transformations episode, Elina Vilk (President & Chief Business Officer) and Mike Ball (Head of Product) join Mike Kaput to explain how the company is reinventing itself around the whole wedding journey, the "aisle to algorithm" strategy. We get into Pearl Planner, the free AI wedding-planning companion now used by hundreds of thousands of couples with no marketing spend; Pearl Connect, the new vendor-side platform built on Google Cloud; the vision quiz and dynamic budgeting that make it work; and how David's own tea

A former OpenAI and Anthropic researcher resigned with a warning that went viral, reigniting the "should we slow down?" debate. Then OpenAI said an internal model solved the Navier–Stokes Millennium Prize problem - followed by a credit controversy with an NYU mathematician. Paul Roetzer and Mike Kaput break down this week’s latest stories, plus new data suggesting an AI jobs boom, Anthropic's 2030 economic scenarios, Jensen Huang declaring AGI has arrived, and a run of "agents gone wild" security incidents. AI-Pulse Survey: Fill out this week’s AI-Pulse Survey here. Show Notes: Access the show

A smaller bank replaced a $375,000 software contract by building its own tool, and now runs named "digital employees" that answer tickets, review its websites, and message staff in Teams. This is what an AI transformation actually looks like inside a regulated institution. CTO John Kowal joins Mike Kaput to unpack Peapack Private Bank and Trust's journey since 2023: flipping build-vs-buy, giving every AI agent a name and personality, building a network of 29 AI champions, and the guardrails that keep AI in check in banking. Part of the AI Transformations series, presented by Google Cloud. Show

GPT-6 Astra can operate a computer for days at a time, and OpenAI is positioning it as the model that does the work, which is exactly why Paul says this launch felt less like a demo and more like a turning point. He and Mike dig into Astra's computer-use leap, the fumbled rollout, and the jobs question the labs keep sidestepping. Then: New York City's ban on student AI through 8th grade and the debate it set off, Trump going all-in on data centers as the fight reaches the midterms, the U.S. government backing OpenAI in the NYT copyright case, Wall Street banks pushing Big Law to cut fees, and

If you stop hiring entry-level staff and quietly decimate the leadership pipeline, where do future managers come from once AI does the junior work? That's one of 15 listener questions Paul Roetzer and Cathy McPhillips take on in this AI Answers edition, drawn from recent Intro to AI and Scaling AI classes and an Academy Live session with SmarterX's COO and legal counsel. The conversation runs from the practical (how to carve out time for AI, generic vs. custom agents, human-in-the-loop checkpoints) to the strategic (funding AI, token budgets, intelligence redundancy, and whether we're building

OpenAI's Hugging Face hack keeps getting stranger; new investigations show autonomous agents that coordinated, deceived, and pursued their goal at almost any cost. Paul Roetzer and Mike Kaput break down what it means for the models coming next. They also cover Bill Gates reversing his stance on AI and jobs, Nvidia doubling revenue and buying Hugging Face, Anthropic's $30 trillion revenue pitch and its court win, McKinsey's 2026 State of AI, and more in the rapid-fire section. AI-Pulse Survey: Fill out this week’s AI-Pulse Survey here. Show Notes: Access the show notes and show links here Times

For twenty years, the percentage of time a B2B sales rep spends actually talking to customers has barely moved. Jon Dick, chief customer officer at HubSpot, thinks AI is the first thing that genuinely breaks that constraint, and he has three years of rebuilding to point to. In this AI Transformations episode, Jon takes Mike Kaput through HubSpot's agentic go-to-market model: the AEO strategy built for answer engines, the AI SDR handling most website chats, the assistant that moved win rate, and the org change that put every AI engineer under one leader. Show Notes: Access the show notes and sh

OpenAI paused reinforcement learning on its most advanced models over cybersecurity concerns, saying its next model may cross a critical capability threshold. Paul's read: the labs aren't removing dangerous behaviors, they're suppressing them and suppressed behaviors can come back. Also this week: a very public argument over Anthropic's direction as it heads toward a possible $2 trillion IPO, Pennsylvania's crackdown on data centers, Nvidia's Ohio AI factory play, new Pew data on young Americans souring on AI, and humanoid robots in Beijing breaking records set by Usain Bolt. AI-Pulse Survey:

Claude will now watermark the content it generates, and the fight over what that means for writing, schoolwork, and "AI plagiarism" is only starting. Paul and Mike unpack the watermarking news, then move through AI's environmental reckoning, OpenAI's very public friction with the White House, Anthropic's march toward a record IPO, Grok 4.6's return to the frontier, and an AI agent that hacked a gym website while trying to book a class. 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 00:07:04 — Claude

Zapier's Chief Marketing and AI Transformation Officer Dan Slagen called a "code red" on his own marketing team, not because they were behind, but because he saw the bar moving from using AI to building with it. In this AI Transformations episode, Slagen tells Mike Kaput how Zapier moved from AI adoption to AI-native work: three tiers of builders, an AI-powered "marketing brain," the "lossless marketing" idea behind it, and why working in public mattered more than any single tool. Show Notes: Access the show notes and show links here Timestamps: 00:00:00 — Intro 00:03:50 — Who Dan Slagen is an

Google’s AI leadership is being reordered as Demis Hassabis changes roles, Jeff Dean departs after 27 years, and power appears to shift back toward Silicon Valley and Sergey Brin. But the bigger story may be OpenAI’s detailed account of agents that hacked Hugging Face, communicated across test runs, shared stolen credentials, and operated undetected for weeks. Paul and Mike break it all down, then cover the White House’s new framework for reviewing frontier models, OpenAI’s delayed Astra release, its public fight with Apple, Meta’s open approach to superintelligence, AI-driven layoffs, Gavin B

A rogue AI agent hacked Hugging Face to cheat on its own test. The Pope wrote 43,000 words warning about AI. Washington pulled Fable 5 offline, then reversed course. Paul Roetzer and Mike Kaput count down the top 10 AI trends of the quarter: AI jobs whiplash, the pillars of business transformation, Apple v. OpenAI, GPT-5.6 and ChatGPT Work, agents transforming work, Fable 5, soft nationalization, the Hugging Face breach, and the battle over open weights. This episode was recorded live on July 31, 2026. Show Notes: Access the show notes and show links here Timestamps: 00:00:00 — Intro 00:02:54

Both OpenAI and Anthropic disclosed this week that AI agents escaped their safety tests and reached real organizations, and one model flagged its own action as wrong before carrying it out. Paul Roetzer and Mike Kaput break down what happened and why it matters for anyone deploying agents. Then: why 1,300+ AI insiders are asking Washington to pace the frontier, where the open-weights battle goes next, Sam Altman's case for an abundant future, Astra's decade-old math breakthrough, Microsoft's record year, NVIDIA's bet on Ilya Sutskever's SSI, and a take on keeping your voice in an AI world. AI-