The Operator's Library
Frameworks and field notes from the podcast, optimized for operators who'd rather read than listen.
Each conversation on Between Two COO's runs 45 to 90 minutes. The articles here distill the most useful frameworks our guests have shared into 5 to 15 minute reads. If you don't have time to scrub a podcast feed for the one operating lesson you actually need this week, start here.
Lab in the loop: teaching an AI what toxic looks like
Anybody can build a lab in the loop. Deep Genomics COO Tom Masterson says the advantage is in what you feed it: data designed for machine learning, including test molecules that are intentionally toxic so the model learns what toxic looks like.
More recent
AI token cost keeps falling. The debt behind it does not.
AI token cost is falling 60 to 80 percent a year. That is good news for anyone buying tokens and a serious problem for anyone who borrowed to make them. Paul Kedrosky of SK Ventures explains the math, and what operators should take from it.
What a forward-deployed team finds inside a mortgage company
Moburst sent a team on-site to a mortgage company to find where AI could help. What they found was forms checked by hand, blurry phone photos, and weekend calls with no context. The first fixes were not AI, and the rule that came out of it applies to any operator.
Seven agencies and a calendar: what actually decides whether a data center in orbit gets built
An orbital data center is not a launch, not a communications satellite, not an imaging satellite, so seven agencies each own a slice of it and the FDA is one of them. Cowboy Space COO Joe Yaffe on the unglamorous constraints that actually decide whether this gets built.
The physics is the easy part: Joe Yaffe of Cowboy Space on what it actually takes to put a data center in orbit
Joe Yaffe spent 31 years as a Silicon Valley lawyer before becoming COO of Cowboy Space. He argues an orbital data center needs no new science, only execution, and that the decision which closes the economics is letting the rocket double as the radiator.
AC vs DC power: the 1890s design decision that could become the next bottleneck in AI
AC vs DC power, explained: why every conversion between the grid and an AI chip burns energy as heat, and why that path is now worth rebuilding.
Jevons paradox and the AI infrastructure bet: the 160 year old idea behind hundreds of billions in spending
An 1865 observation about coal is the argument underneath hundreds of billions in AI infrastructure spending. Jevons paradox, and what would break it.
Generative engine optimization for physical products: Riikka Söderlund of Katana on why AI search cannot see what you cannot count
Katana COO Riikka Söderlund on generative engine optimization, why AI search hides out-of-stock products, and the brand selling at a loss unaware.
AI agent security after a model walked out of its own sandbox: what operators should actually do
A model escaped its test environment and attacked another company. The AI agent security failures were operational, and seven of them are yours to fix.
The full library
Every episode, most recent first.
The COO's Execution Playbook
Frameworks, templates, and hard-won lessons from operators who've been in the chair. Every Tuesday.
No spam. Unsubscribe anytime.