The operating intelligence platform for operators & podcast hosts
The smartest operators keep saying the same things. We found them all — and how they work.
Operators is a structured corpus of operating principles, frameworks, and cross-episode patterns extracted from the best business podcasts — organised around the decisions founders actually face.
Founders and operators browse the corpus for free. Podcast hosts get a full intelligence layer over their show — and a public dashboard their audience can use.
For founders & operators
Podcast wisdom that actually compounds
You've listened to hundreds of hours. The insight from Tuesday's episode is gone by Thursday. Operators turns the best operating knowledge into a structured, searchable corpus — organised around the decisions you're actually facing, not just the episode that was fresh.
Free. Or browse the corpus →
For podcast hosts
Give your show a permanent intelligence layer
Every episode you publish is richer than your newsletter shows. Operators extracts all of it — then gives you a private brief for Thursday and a permanent public dashboard at operators.work/show/your-podcast that your listeners can actually use.
5 free briefs · No account needed · Limited time
The Operator Signal Corpus
Eight shapes of knowledge. One extraction pipeline.
Operator knowledge doesn't fail because the insight isn't there. It fails because a durable principle and a one-week tactic sound identical coming out of a speaker. We extract and classify both — in eight distinct types — so every insight you find tells you exactly what it is and how to use it.
A reusable truth
Claims backed by evidence across multiple independent conversations. Not one person's opinion — the same insight surfacing across shows, guests, and stages.
Example
“Speed compounds when learning loops are tight”
A repeatable method
Named structures with steps, conditions for when to apply them, and honest failure modes. What to do, when it works, and exactly where it breaks.
Example
“Barrels vs. Ammunition — diagnosing why output isn’t scaling with headcount”
What actually happened
Specific incidents, named and real, with a transferable takeaway. History that teaches rather than stories that entertain — bound to the episode where they surface.
Example
“Shipping a UI feature ahead of the underlying ML — the costs that weren’t visible until month two”
What’s shifting early
Forward-looking observations about markets, tools, and behaviour — with evidence of momentum before something becomes consensus. The difference between early and obvious.
Example
“Non-developers running production workflows through AI agents — adoption steeper than reported”
A gap worth targeting
Specific market gaps with a named buyer, a revenue path, and a reason the gap is open right now. Not trends — exploitable whitespace with a shape to it.
Example
“Wrapping enterprise data in MCP for non-engineer AI reasoning — no one owns this category yet”
Where operators disagree
Two valid positions held by credible operators in direct conflict — with the condition that determines which one wins. The most decision-useful content in the corpus.
Example
“Move fast to learn vs. move right to protect trust — both true, not always compatible”
What to do this week
Verb-first executable actions tied directly to evidence of them working. Not theory — the specific move, the context it applies in, and the outcome attached.
Example
“Send the referral ask within 24 hours of a customer saying ‘this is working’ — 3–4× higher conversion”
How operators decided — and what happened
Narrated choices made at a specific stage, with a stated outcome and a transferable lesson. When the same decision appears across multiple episodes, an outcome distribution emerges. That’s not advice — it’s data.
Example
“Cut sales headcount at Series A to extend runway — chose survival over growth — reached profitability 14 months later”
Aggregates into Decision Patterns across episodes
Every object in the corpus carries its evidence source, a confidence level, a stage tag, and — for principles and frameworks — an honest failure condition. The corpus doesn't just tell you what smart operators believe. It tells you when they're wrong.
Browse the corpus →Principles
Every principle with its evidence, stage, and failure condition
A principle without a failure condition is just an opinion. Every insight in the corpus carries the context where it applies and the condition where it breaks.
Barrels define throughput — not headcount
Organisational output is constrained by people who can own outcomes end-to-end without supervision. Adding ammunition without adding barrels produces diminishing returns.
Speed compounds when learning loops are tight
Fast iteration is not cosmetic. It changes product quality, resource efficiency, and strategic position over time — the gap between fast and slow teams widens, it doesn’t close.
Hire for trajectory above current altitude
The only defensible hiring signal is velocity of growth over current skill level. A person who has improved three standard deviations in two years will outperform a static expert within months.
Cross-episode pattern recognition
When the same principle appears across independent sources, it earns a higher confidence
Every cross-episode pattern shows its evidence — attributed, labelled by type, and traceable back to the original source. A principle seen across 9 evidence points in 4 podcasts is categorically different from one person's anecdote.
Direct contact before delegation
The strongest operators maintain direct personal contact with a function before handing it off. Not to micromanage — but to build the judgment to know what good looks like and when something is wrong.
Founders / Interview · Tony Xu — DoorDash
“I delivered food myself for months. By the time I hired my first ops lead, I knew every edge case they’d face.”
Founders / Book breakdown · James Dyson
“Dyson spent years on the factory floor. The delegation model worked because he’d built the judgment to know what good looked like.”
Tim Ferriss Show · Jim Collins
“Great companies are built by people who understand the work at every level before they systematise it.”
My First Million · Sam Parr & Shaan Puri
“Founders who skip the early grind don’t miss operational knowledge — they lose the credibility to know when something is wrong.”
Live tension in the corpus
Keith Rabois argues this has a natural limit — at scale, a founder cannot be hands-on in every function. Resolution: direct contact is required before the first delegation, not perpetually.
Frameworks
Named decision tools — with when they break
A framework without a “don't use when” is just a slogan. Every framework in the corpus is a reusable decision tool with constraints built in.
Barrels vs. Ammunition
Keith Rabois · Founders Podcast
Definition
Barrels are people who carry an idea from concept to shipped outcome without supervision. Ammunition executes tasks given by barrels. Organisations are constrained by barrel count, not headcount.
Use when
Diagnosing why output isn’t scaling with hiring. Building a hiring scorecard for senior roles. Deciding who should own a new function.
Don't use when
The first 10 people, where everyone must be a barrel. Deep craft roles where expertise outranks ownership instinct.
The Earned Secret Test
Peter Thiel / DoorDash case
Before entering a market, test whether your insight was earned through direct experience or assembled from public information. Earned secrets produce defensible strategy; unearned beliefs produce consensus trades.
The Delegation Readiness Filter
Synthesised across DoorDash, Dyson, Jim Collins
Only delegate a function when you can do it well yourself — not to be the best at it, but to audit quality and recognise failure. Delegation of functions you cannot evaluate compounds silently.
Tensions
Where smart operators fundamentally disagree
The most useful operating knowledge often lives in the contradiction. The corpus surfaces disagreements between credible sources, attributes both sides, and resolves them where possible.
Should founders talk to customers?
▾Keith Rabois · Founders Podcast
No — intuition is contaminated by articulation
“When you ask users what they want, they describe a shallower version of what they actually need. Building to stated preference produces mediocre products.”
Context
Consumer products at zero-to-one, where latent behaviour is the signal.
Brian Chesky · Lenny’s Podcast
Yes — direct contact is the discipline
“I called customers every night for a year. Not to gather requirements — to understand what they were actually experiencing. That’s categorically different from a survey.”
Context
Any stage where the founder needs to understand the emotional texture of the problem.
Resolution
The tension dissolves on method. Survey-style feedback contaminates product intuition. Deep observational contact builds it. Chesky and Rabois are using the same word to describe different activities.