Don't be fooled by @GoogleCloud remixing existing services at Next. That's the whole point. Knowledge Catalog is Dataplex reframed. ADK is Vertex AI agent tooling consolidated. Agent Runtime is Cloud Run with agent-specific optimizations. The cross-cloud lakehouse is built on Iceberg and Cross-Cloud Interconnect that already existed. The industry looked at 250 announcements and said "nothing new." They're right. And they're missing what matters. When a vendor announces a brand-new product, you're looking at a roadmap. When a vendor remixes existing capabilities into a new control point, you're looking at something closer to production-ready. The pieces have been running already. Google didn't spend three days in Vegas talking about models or benchmarks. The keynote wasn't about the labs. It was about the gap between demo and production. That's the gap where every enterprise is stuck right now. And here's what the neocloud price war is missing. The $4/hour GB200 is a Layer 0 fight. Compute. Hardware. Infrastructure. Real value, real price advantage. But Google isn't competing at Layer 0. They're competing at Layer 1C and Layer 2C — the judgment layers. The catalog that models what your data means. The agent governance that determines what your agents are allowed to do. The identity framework that controls which agent accesses which context. We've seen this before. How much do you pay for CPU in your Lambda functions? You don't know. You don't care. AWS won by making the infrastructure invisible and the abstraction indispensable. Google is running the same play. The GPU is not the moat. The borrowed judgment above it is. I wrote the full analysis. Link below.

As my sons started their professional careers (logistics and CS), I warned them the average co-worker isn’t that productive. I warned them not to measure their output by the standards of their coworkers. They were still surprised. Most people view work as something they do to pay bills. That’s perfectly fine. What companies are looking for are people passionate about infrastructure. You want a list of those people? Take a look at who has attended multiple @TechFieldDay events as a delegate.

Another way to read Satya Nadella’s recent comments on AI: AI isn’t going to change jobs as fast as advertised. Not because AI isn’t powerful. But because organizations change more slowly than models improve. Most jobs aren’t a single task that can be automated. They’re bundles of workflows, decisions, approvals, and accountability. AI is already very good at parts of work: – drafting – summarizing – analyzing – synthesizing inputs But roles don’t disappear when parts get faster. They change when systems, incentives, and responsibility change. That takes time. What we’re seeing instead: – AI augmenting decisions before replacing roles – productivity gains showing up unevenly – expectations rising faster than headcount falling In the near term, AI won’t eliminate jobs. It will expose: – slow processes – fuzzy ownership – brittle workflows – weak decision discipline AI will change how decisions get made long before it changes how many people are employed. If you’re thinking about AI this way — decisions and workflows first, seats and org charts later — I unpack that more deeply here: 👉 ctoadvisor.substack.com/p/st…