Podcasts

What happens when every enterprise gets AI agents, but nobody rethinks the work? Most leaders aren't asking that question. They should be. Because the window to get this right is shorter than they think, and the upside for those who do is enormous. The substrate is that every AI conversation should be an operating model conversation. Today, Anthropic and Infosys announced a collaboration to build the next enterprise agentic layer. Between Claude Cowork , their legal tool , a

OpenAI just hired Peter Steinberger (the creator of Clawdbot, briefly Moltbot, now OpenClaw) to build its next generation of personal agents. What they build next feeds into ChatGPT's 800-million-strong user base. A chess move ahead of what could be the largest tech IPO in history. ChatGPT's lock-in is surprisingly thin. Custom GPTs never became the App Store, memory is shallow, and most users could switch to Claude or Gemini tomorrow and lose nothing. That's a problem when y

There’s a tectonic shift underway. Enterprise AI spending tripled to $37 billion in a single year. That’s a 3X increase in 12 months. Where does this money come from? It is carved out of existing software budgets, especially traditional SaaS. And therein lies the tension. For twenty years, SaaS was the greatest business model in technology. Recurring revenue. Per-seat pricing that scaled beautifully with headcount. It turned startups into empires and minted more millionaires

There are two groups of leaders forming right now. The gap between them doubles roughly every few months. And it’s invisible to the group falling behind. I call it the Compounding AI Fluency Gap. One group uses AI for real work. Daily, messily, imperfectly. They’ve felt it fail. They know where the edges are. The other group knows AI matters, has a strategy for it, and has never once been surprised by what it can or can’t do, because they’ve never pushed it themselves. That s

When I was a kid, the “serious” adults told my friends that thousands of hours in StarCraft and Age of Empires were a waste of time. They were wrong. Those kids were training for the real job of the 2030s: commanding swarms of AI agents under pressure. The fortunes will be made by the people who decide where the trains go. The substrate is AI orchestration: coordinating dozens of AI agents in parallel, allocating compute, routing decisions between humans and machines in real

OpenAI ,  Anthropic , and  SpaceX  are all preparing for potential IPOs this year. If you don’t follow tech markets, that might not sound like news. But all three going public in the same window, at combined valuations in the trillions, represents something bigger than just stock listings. For context, Saudi  aramco ’s roughly 29 billion IPO in 2019 was the largest in history. For the first time since the original internet boom, a new wave of genuinely transformative technolo

PwC is training its 75,000 U.S. employees in AI skills through its new AI Learning Collective . It is built around a recognition that the half-life of usable knowledge is now shorter than most corporate training cycles. Models evolve faster than curricula. Even well-funded upskilling programs are structurally late. PwC’s insight is subtle and decisive. They are not teaching employees how to use tools. They are rebuilding the firm’s operating system. The Substrate: We have cr

AI just found 800 cosmic anomalies in 35 years of Hubble images. The discoveries were sitting there the entire time. 60 hours of compute analyzed 100 million image cutouts and revealed gravitational lenses, galaxy mergers, and objects scientists can't classify yet. The implications extend beyond astronomy. Every industry sits on decades of legacy data. Most of it was analyzed once with the tools available at the time, then stored. Reanalyzing old data was always possible. It

Last week I wrote about the Hubble archive. Thirty-five years of data hiding 800 cosmic anomalies. This week, McKinsey found its own. Here’s what happened. McKinsey receives 1 million applications a year and hires less than 1% of them. They wanted to know: What distinguishes candidates who become successful partners from those who don’t? They analyzed 20 years of hiring data with AI. The pattern jumped out immediately. Candidates who’d experienced setbacks and recovered were