Enterprise AI Innovators

The episode contains several genuinely useful operational insights - the 50% AI-code threshold as a role-redefining inflection point, the CTO's practice of cloning repos before meetings, and the CPA labor-gap framing - but roughly half the runtime is career biography, company overview, and generic change-management commentary that a B2B operator would already know.

“when we hit 50 plus percent of our code being written by AI...no longer was a human the primary controller and contributor to the code base”

“I can, if I've got a meeting with a team coming up on Friday and we're talking about a product like can literally go clone the repo and spend an hour like with Claude or with Codex”

The copilot-to-colleague framing and the specific 50% threshold as a meaningful cognitive shift in engineer identity are genuinely useful framings; however, much of the rest - trust barriers, T-shaped talent models, 'curiosity' as a skill, change-management challenges - is standard enterprise-AI discourse that circulates widely.

“your job as an engineer shifted towards being like the contributor and owner of the code base, to more being like the controller and governor of the code base”

“the value that you create as an engineer is not the lines of code that you write. And in fact, it's never been the lines of code you write”

Joel is an actual operating CTO with direct responsibility for 5,000 engineers, 100-plus products, and a company that ships AI into regulated professional workflows; he came from an acquired AI startup, giving him both startup and enterprise practitioner depth rather than a pure thought-leader profile.

“there's 5,000 engineers in my organization. Like it was impossible for me to ever like look at a line of code of any of the people”

“we now have agents that can do all parts of that process. And so the tax professional's job now turns much more into advisory”

Named products (Westlaw Deep Research, Co-Counsel, Claude, Codex), a concrete user count, a specific percentage threshold, and a detailed end-to-end tax workflow description add real texture; however, there are no productivity multipliers, error-rate data, cost figures, or timeline evidence to anchor the claimed outcomes.

“A million users of co counsel today”

“we've actually built products that really end to end, automate the development of tax returns...to take like a shoebox of W2s, 1099s and put it through all the tax regulation and tax law and compute a tax return”

The hosts frame the episode well in the intro and ask reasonably organized questions, but they never challenge a single claim, let the most interesting threads (accuracy of automated tax returns, hallucination risks in legal research, actual productivity data) pass without follow-up, and the lightning round is purely surface-level.

“What about kind of like on the internal AI use run kind of AI adoption? Are there some ways you and the team are using AI that actually, let's start with the team”

“Are there tools they're using, are there workflows have kind of dramatically changed?”