Concept

Social Radar

conceptfounder-evaluationstartupsy-combinatorinvesting

Social Radar

The Social Radar is the skill of judging founders by reading their social and emotional signals under pressure, rather than by their credentials, their idea, or their technology. The term began as Paul Graham‘s nickname for Jessica Livingston, who ran founder evaluation for Y Combinator‘s first batches: her three technical co-founders questioned the product while she watched the people, then surfaced the human red flags they had missed.

Why it works at the earliest stage

At the seed stage there is almost no data — often only an idea, and the idea will usually move. PayPal began as money transfer between PalmPilots; Airbnb’s idea was one YC actively disliked. What survives the pivots is the founders. So the evaluation that matters is of people, and the only instrument available in a 10-minute interview is a read of how they behave. The radar is the attempt to make that read deliberate.

What it reads

Livingston’s signals, in rough order of weight:

  • Defensiveness — the master negative signal. A founder who bristles under questioning will struggle to educate sceptics, to pivot, and to learn from others. The best founders treat hard questions as a tennis match, not an interrogation.
  • Earnestness / authenticity — genuine care for the problem and the user, as against doing a startup because it is cool or looks like easy money. Reads as humility under questioning (‘I don’t know — here is where I got to’).
  • Commitment — willingness to quit the day job and ‘burn the boat’. Founders who keep a salary and health insurance tend not to push through the hard parts.
  • Co-founder dynamics — do they get along, finish each other’s sentences, share a history? Co-founder conflict is a leading cause of startup death. Teams formed only to start a company, with no prior relationship, are a major red flag.
  • Resourcefulness — making something happen regardless of obstacles (Paul Graham‘s ‘relentlessly resourceful’). Airbnb’s hand-glued cereal boxes are the canonical example.
  • Domain rootedness — fixing a real, often unsexy problem the founder is personally connected to (Parker Conrad on broken HR plumbing), as against an idea invented to chase a market.

Charisma is explicitly distrusted as a primary signal — many charismatic people are ‘full of baloney’. It helps only with substance behind it.

Is it trainable?

Partly. Livingston frames the gift itself — a compulsive interest in people, present since childhood — as untrainable, and offers external evidence of trait-level perceptual skill: a perfect 36/36 on Simon Baron-Cohen’s Reading the Mind in the Eyes test, read by asking ‘what is this person trying to tell me?’ rather than matching emotion labels. But the method is teachable to those without the gift: acknowledge you are bad at it and therefore be deliberate; carry a short mental checklist (do they get along, are they defensive, do they understand the product); ask plain questions during and after the interview. There are no trick questions — the discipline is attention, not interrogation.

She calibrates the radar by following up on strong gut instincts years later to check whether she was right.

Where mainstream views differ

The Social Radar is a defence of structured intuition, and it sits against two currents.

  • Data-driven and structured-interview evaluation. A large body of hiring and selection research finds unstructured, gut-feel interviews to be weak predictors relative to structured, criterion-based ones, precisely because intuition is noisy and over-confident. Livingston’s checklist is a partial concession — it structures the read — but the core claim is still that a trained perceiver beats a rubric at the earliest stage, where there is too little data for a rubric to bite.
  • Bias. Reading people on ‘authenticity’, ‘earnestness’ and ‘fit’ is exactly the channel through which pattern-matching encodes demographic and cultural bias — backing founders who resemble past winners. Livingston’s own ‘could I have dinner with them every week?’ filter is candid about this risk. Defenders argue the alternative (credentials, warm intros) encodes bias at least as strongly; critics argue gut-feel launders it as taste.
  • Falsifiability. Her reported record on strong negative vetoes is clean — no founder she put her foot down against became a major success — but this is self-reported and selection-prone, and she volunteers a catastrophic miss (a funded founder later convicted of laundering stolen Bitcoin). The predictive validity of a 10-minute read remains genuinely uncertain; the honest position is that the perceptual gift is real and its hit-rate unmeasured.

See also