Next in Tech

The episode delivers a handful of genuinely useful concepts - three root causes of agent hallucination, the semantic/ontology foundation argument, and policy-bounded autonomy - but these are diluted by significant host-and-guest agreement loops, restating of premises, and motivational-tone filler that slows the idea rate meaningfully for a 24-minute runtime.

“Automation without governance isn't a feature, it's a liability you'll be spending years to unwind later”

“The three things would be the ground truth was not retrievable, so therefore it made something up. Uh, or the data was there, but it was ambiguous.”

The reframe of 'autonomy' as the wrong question, replaced by 'how much responsibility do we want to safely hand over,' is a genuinely useful inversion, and the ontology-before-MCP argument is underappreciated in the market; however, the build-vs-buy section and trust-in-capital-markets framing are well-worn territory with no contrarian edge.

“the noise about autonomy is probably the wrong question to a certain degree. And um, the reason I say that is the real question is how much responsibility do we want to safely hand over to an agent”

“a connection without a shared meaning is just plumbing”

Krishna Ventimuri is a genuine practitioner - CTO at S&P Global Enterprise Solutions - with evident hands-on exposure to production agentic systems in a regulated data context, which gives his warnings about governance real credibility; however, this is an internal corporate podcast, limiting the candor and external proof points a truly independent operator voice would bring.

“early on, all of us, like everybody else in the industry, we got very excited about Agent I there, mateen there, and we built a demo. It was honestly wickedly cool. It lit up the room, everybody was excited. And I was sitting thinking, this is Great. But demo was phenomenal. But if I put this into production, how do I know that the actions the agent is taking is right?”

“I went to the team and I said, okay, can you add these more rules and say Krishna, can you have a baseline? They said, no, never.”

Named technologies (OPA, MCP, FIBO, Common Domain Model/FENOS, data mesh) and a concrete example of semantic ambiguity across capital markets entities give the episode more grounding than pure abstraction; but there are zero customer names, no metrics, no dollar figures, and the one production anecdote ('demo lit up the room') stays deliberately vague.

“we use OPA as our policy agent. So every action now is governed means the same thing, it is auditable.”

“the common Damian model, which is now under phenos, or the FIBO model, which has been defined”

The host shows genuine domain literacy - correctly distinguishing MCP context from ontological grounding and connecting themes across episodes - but questions are frequently long, leading, and self-answering, and there is no instance of productive pushback or challenge to any claim the guest makes.

“You mentioned MCP earlier and I think oftentimes we get into discussions about it's the model context protocol. And through the interface you now have the ability to define context around that data. And that's certainly better than raw APIs without ability to handle it. But you pointed out that thing that, however boring, is still so critical”

“How do we move across that and what are the approaches that we can take to ensure that. In fact, we're able to fully realize the many potential benefits that we could pull out of these capabilities.”