Private Equity FunCast

There are genuine practitioner insights buried here—sell-side diligence timing, AI feature telemetry gaps, agentic development maturity signals—but they're heavily diluted by Kansas City BBQ tangents, personal app-building anecdotes, VW settlement stories, and sports jokes. Probably 20-25 minutes of substantive content in an 84-minute episode.

“The miracle of SaaS is potentially dead... you will have deeply unprofitable users, you will have deeply unprofitable user flows and workflows unless you embrace AI”

“Where they've put AI on a feature and they're losing money on the feature because they haven't put the telemetry in to know what the value of that feature is”

A few genuinely fresh framings—documentation as a degraded diligence signal in the AI era, continuous DD as a product rather than one-off event, shielding target tech teams from multiple simultaneous bidder workshops—but these are islands in a sea of recycled wisdom about cloud spend, right-tool-for-right-job, and architecture basics.

“Great documentation used to be a good signal, a positive one. Yeah. Nowadays you can just churn out documentation. Right. Like we see more and more cloth decks where you can definitely slap”

“Our goal also in the mid to long run is actually to do have a continuous due diligence where DDs don't just become this one off thing, but you continuously DD and improve a company”

Code & Co are genuine domain practitioners—200+ tech DDs per year, founded 2016, real operator backgrounds (ThoughtWorks, Rocket Internet, K1 add-on)—with credible, lived expertise. Not marquee names, but legitimately the people who actually do this work at volume, which earns real credit.

“Last year we did probably like 200 DD's this year we hope to double it, knock on wood”

“We work from pre loi to post exit and everything in between”

The $6M cloud savings figure, the step-by-step Draft Punk agentic workflow via MCP, and the pen-test-uploaded-night-before-signing story are concrete and vivid. Most diligence war stories are kept deliberately vague for confidentiality, which is understandable but limits the evidence base.

“We identified cost saving potential of $6 million per year”

“Someone reports a bug in our internal software. So I write a message in Slack. Then an agent that we call draftpunk because we're engineers and we like stupid puns. Draft punk picks up the ticket or a message and moves it over by MCP to Linear and creates a ticket”

The host has genuine domain fluency and lands some sharp setup questions, but routinely lets conversations drift into personal anecdotes (his own app-building, the Kobe-shirt-at-Celtics story), and the promised 'phone booth' structured answers never materialise. Guests are rarely pushed to defend claims or quantify assertions.

“Do you guys believe in this concept of if they've got data and expertise at the data layer, that's going to be a moat? Or do you think that Milk goes away with AI just because data's everywhere?”

“I ended up writing my own exercise tracking application. I posted in the App Store. We did a video about it. Not because it's a Goldilocks problem. I've been training a certain way for my entire life and no app ever”