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.

1194Corpus objects
135Episodes processed
59Cross-episode patterns
12Podcasts tracked

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.

Browse 1194 corpus objects across 12 of the best business podcasts
Filter by the decision you're facing — 24 decision domains
See which ideas have real evidence vs. one person's strong opinion
Every principle carries a failure condition and a stage tag
Get the week's highest-signal drop — free, every Tuesday

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.

Private brief: angles, pull quotes, cross-corpus patterns — ~4h delivery from transcript submission
Public episode page: full principle stack, frameworks, tensions
Show dashboard: emerging canon, fault lines, episode clusters
Shareable URL in every show note — free for your audience
Compounds with every episode — your archive becomes navigable
Start free — 5 episodes →

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.

Principle

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”

Framework

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”

Lesson

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”

Signal

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”

Opportunity

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”

Tension

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”

PlayNew

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”

Decision MomentsNew

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.

01

Barrels define throughput — not headcount

New96%

Organisational output is constrained by people who can own outcomes end-to-end without supervision. Adding ammunition without adding barrels produces diminishing returns.

9 evidence pts3 podcastsSeries ASeries BHiringOrg Design
02

Speed compounds when learning loops are tight

93%

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.

11 evidence pts4 podcastsSeedSeries AProductStrategy
03

Hire for trajectory above current altitude

94%

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.

8 evidence pts3 podcastsSeedSeries AHiring

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.

High confidence12 evidence points5 podcasts

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.

direct

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.

paraphrase

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.

paraphrase

Tim Ferriss Show · Jim Collins

Great companies are built by people who understand the work at every level before they systematise it.

synthesis

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

Series A+

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.

DomainHiring · Org Design

The Earned Secret Test

Pre-seedSeed

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

SeedSeries ASeries B

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.

Move fast and iterate, or move right the first time?