"AI Hype Keeps Me Up at Night (and I'm the AI Guy)" —Barry Dauber

Your Edge Isn't the Model Anymore, with Dr. Anthony Annunziata

Dr. Anthony Annunziata started at IBM building better memory and transistors out of magnetic materials. The work was good, but he kept feeling the same itch: get the thing out of the lab and into the world where it actually matters. That instinct shaped everything after, through quantum computing, AI for science, and now IBM's open-source AI strategy. The physicist's habit is first-principles thinking, stripping something down to what truly makes it work. But research and business part ways at the definition of done. In research, you discover, publish, and move on. In business, a demo is nothing. It doesn't work until someone else uses it, trusts it, and pays for it. He sees the same gap with agents today: demoing one is trivial, shipping one to production is the hard part. The throughline is where your edge actually comes from. Frontier models are extraordinary, but progress on them has slowed while the cost of pushing further climbs, so the model itself stops being the differentiator. What's left is yours to build: your data, an eval strategy grounded in your real use cases, and the choice to run smaller models on your own infrastructure rather than handing your proprietary edge to a tech company.

After the SaaSpocalypse, with Wade Foster

In 2011, three co-founders from Missouri walked into Y Combinator with a simple observation: every SaaS help forum had the same unanswered thread. "When will you integrate with X?" Nobody was building the infrastructure to close that loop. They did. Paul Graham told Wade and every YC batch the same thing: be a cockroach. Don't raise money. Survive. Adapt to whatever comes next. Almost nobody listened. Zapier did. Fourteen years, $1.3M raised, bootstrapped to a $5B valuation. Every time a funding opportunity came up, they asked one question: what is actually holding us back right now? The answer was never the balance sheet. Then the SaaSpocalypse arrived. SaaS stocks cratered. Every software company started asking: what part of what we built actually survives this? Wade's answer is specific. Not the 8,500 integrations. The governance layer, the auth infrastructure, the background automation reliability stack. And then Wade, in this episode of Revenue Engine Masters podcast, says something incredible. They built their entire business on drag and drop. It made them. And they're already moving past it. Natural language builders outperform it on activation. They measured it. They called it "yap to zap" and didn't look back. For Wade, this adaptability applies to humans too, not just companies. The best operators today aren't the ones prompting AI and moving on. They're the ones staying in the loop, iterating on context, pushing until the output is actually good. That's the human role now. In the outer loop around it. Paul Graham's advice was about never optimizing for the conditions of today at the expense of your ability to adapt to whatever comes after. Zapier understood that in 2011. They still do.

"AI Hype Keeps Me Up at Night (and I'm the AI Guy)" —Barry Dauber

Barry Dauber's job running GenAI GTM at Databricks is to convince enterprises to move faster on AI. So when he sat down at a VC dinner last year and everyone went around the table sharing what kept them up at night, the room got a surprise when his answer was "AI hype". Barry is not a skeptic. He believes AI is transforming everything, and he has the customer stories to back it up. But he has also watched enough enterprise pilots stall, enough Monday morning messages come in saying the data was not ready, to know that the gap between what AI promises and what organizations can actually deliver is real and wide. In this episode, Barry and Elio work through what that gap looks like from the inside: why most companies are not as ready as they think, what the early GTM lessons from MosaicML taught him about building before the market understood what you were selling, and what it actually takes to move from experiment to production at scale.

Why Selling Will Always Be Human

Remy Piazza, Principal & CRO at Powerhouse Consulting, started his career at Xerox, the godfather of solution selling. In this episode, we explore why despite AI and evolving tech stacks, selling remains fundamentally human. From mastering discovery to building trust, Remy shares why curiosity and genuine connection will always outperform automation.

The Art and Science of Revenue Operations - SK Ramakuru

Elio interviews SK Ramakuru about what RevOps really is: the “engine” behind selling that turns chaotic people/systems/data into repeatable motion. SK shares his path from Hyderabad, India—where early exposure to big tech sparked his ambitions—into consulting, then a Master’s program in the U.S. that deepened his stats and data skills. An experience of temporarily stepping into a selling role, and struggling badly, cemented a core belief: selling is art, not science. Data can point you toward opportunity, but performance depends on preparation, context, relationships, and understanding customer pain. SK explains his analytical approach: in RevOps, the goal isn’t perfect data—it’s decision-guiding data that drives behavior and feels fair, especially in territory planning. His mental model is: start by trusting nothing, validate assumptions, slice data from multiple angles, and build a clear story that leaders and the field can understand. As MongoDB scaled from ~600 to ~2,500 sellers, he emphasizes designing territories that match the right accounts to the right reps, remain equitable, require minimal additional headcount, and are explainable and auditable. SK and Elio also talk about how AI is changing both how RevOps works and what signals matter. Traditional indicators like website visits and content downloads are less reliable as buyers use answer engines (ChatGPT/Gemini) instead. That forces new time- and revenue-based indicators and stronger foundational data orchestration. SK argues the market is flooded with “AI wrappers”; the winners will orchestrate messy data across systems, not just generate answers.