The Data Exchange with Ben Lorica

The episode delivers solid insights on open models prevailing, the code generation attenuation funnel, and neo-consulting's outcome-based model, but relies heavily on conceptual frameworks rather than new data or surprising findings. The attenuation funnel (2-3x code written, 30% shipped, minimal end-user impact) is valuable, but much of the discussion confirms existing intuitions rather than challenging them.

“2x3x more code, but then the next step is how much uh new software is getting shipped that's not at the order of 2x or 3x, maybe it's in the 30%”

“the attenuation funnel works as follows. So there's no doubt the developers are writing more code, maybe 2x3x more. But then the next step is how much uh new software is getting shipped”

The framing of open models as a cost/sovereignty play and the attenuation funnel concept offer useful lenses, but the arguments are incremental improvements on existing tech industry narratives. The neo-consulting section rehashes the known shift from t&m to outcomes-based pricing. Limited contrarian positions; mostly validating current industry trends.

“cost of using the proprietary models and the need for a corporation to keep its IP which for the purposes of AI is really in the data that the corporation has”

“the attenuation funnel”

Evangelos Moody is a practitioner-focused guest with credible enterprise AI advisory experience and historical perspective (IBM AI division in 90s), but lacks current operational depth. He speaks about trends observed through client work rather than building/scaling AI products himself. Ben Lorica is a research-oriented host with some practitioner credibility but primarily a commentator.

“based on uh, I've been doing on legacy enterprises and the use of AI by them, as well as the work that our firm has been doing with, uh, such corporations”

“when I was running IBM's AI division this is back in the 90s”

The episode references some concrete examples (Datadog foundation model, Gemma and Nemotron open models, Distill/Palantir, Microsoft's $2.5B venture) and cites survey data (DevOps Research survey on AI as amplifier), but lacks specificity on client outcomes, financial impact, timelines, and measurable results. Most claims remain categorical or illustrative rather than evidence-backed.

“Datadog is famously built a foundation model for observability data uh ranging from parameter count in the several hundred million to a few billion”

“Microsoft allocating two and a half”

Ben presses Evangelos on contradictions (enterprises don't have their act together, so why use open models?), challenges the neo-consulting framing (is it really consulting or vendor services?), and pushes back on the J-curve speculation. Good follow-ups on business model implications and India's impact. However, some tangents meander and Ben occasionally accepts answers without deeper probing; conversational energy is professional but not sharply adversarial.

“Doesn't this contradict what you keep saying on this podcast, Evangelos, that uh, enterprises actually don't have their act together”

“Is there an easy way to think of NIO consulting as a clearly different category from Accenture McKinsey?”