Great Data Minds

The episode delivers several genuinely operational insights - the three-barrier framework for data sharing (practical, ethical, legal), the internal-consultancy CDO model, the incremental 'spend 50 to save 70' funding loop, and the Snowflake organisations approach to virtual centralisation without central ownership. However, roughly a quarter of runtime is housekeeping, affirmations, and standard change-management platitudes that dilute the density.

“start, start with 50 grand use 50 to save yourself 70, and you funded your next project right”

“Can I do it practically like, what is the development project to do this data share? Can I do it ethically? Right? Is it the right thing to do? And can I do it legally right?”

The Snowflake organisations-as-virtual-centralisation model and the three-barrier data-sharing framework are fresh and practically useful, and the sewer-permit story is a genuinely striking illustration of configuration debt. But the broader arc - listen first, build champions, start small, raise literacy slowly - is well-trodden consulting doctrine dressed in government clothes.

“centralizing in the in the old model of like, we're gonna buy a data lake and put data from each agency into this lake was an absolute nonstarter”

“it's the frog in the in the boiling water analogy, right where they almost don't notice that they're becoming data Literate, because we're we're feeding it to them so slowly”

Adam Carpenter is an actual sitting state CDO who speaks entirely from direct operational experience, citing real project costs, timelines, and regulatory constraints; he is emphatically not a circuit thought-leader. His cross-industry consulting background in AI/ML before joining government adds genuine credibility, though the Montana scale (just over a million people, limited agency headcount) keeps the complexity ceiling lower than a large-state or federal equivalent.

“I went to consulting a few years before joining the State, and went there to lead their machine learning and AI practice for the Western United States”

“for I think the total project cost was 208,000 we digitize 9 million records”

The episode is unusually concrete for a government-sector webinar: hard dollar figures, named vendors, statutory timelines, rejection rates, and square-mile population densities all appear. The child-safety digitisation project ($208K cost, 9M records, $400K/year labour saving, $30K/year FedEx saving) and the sewer-permit breakdown (12 years, 82% rejection, 90-day statutory window, one-month fix) are textbook examples of evidence-backed storytelling.

“for I think the total project cost was 208,000 we digitize 9 million records. And now there is a portal you can go to...it saved 400,000 per year in labor and 30,000 per year in Fedex costs alone”

“An agency can get in for $6,000 and have enough capacity for the full year”

Mike Lampa functions more as a warm prompt-giver than an interviewer - questions are broad scene-setters ('How did you get this started?', 'Where are you guys going next?') and every answer is met with affirmation rather than follow-up or challenge. No claim is pushed on, no failure asked about, no tension introduced; Kalia's one pointed question about public-versus-private balance was the sharpest moment in the session.

“How did you get this started. How did this come about?”

“Where where are you guys going next? What's gonna you know kind of things are we going to be doing in the future?”