Picture a job interview. The question sounds harmless, almost like a filler while the interviewer checks their notes. You answer, maybe after half a beat of thought. What you don't know is that your answer barely mattered. What the interviewer actually logged was how long you took, where your eyes moved while you thought, and whether you glanced back at them afterward to check how you'd landed. Chase Hughes built an entire interview model — the HBI — on this one design principle: ask a question whose surface content is disposable, and mine the response for behavior instead.1
Hughes developed the HBI to assess incoming employees at an intelligence agency; he reports it was adopted almost wholesale and is still used there as a standard hiring metric.1 Every question in the set is paired with a stated rationale — not "here's what a good answer looks like," but "here's the behavior I'm trying to elicit, and here's what different behavioral responses tell me." A question like "would you rather be successful or significant?" isn't measuring which word the candidate picks. It's measuring hesitation, eye-accessing direction, whether they check the interviewer's face for a reaction afterward — a nonverbal habit Hughes reads as a tell for people who are used to seeking approval.2
The questions escalate in a deliberate order. Early questions (effective/efficient, successful/significant) establish a comfort and honesty baseline — how this specific person's face and body move when they're not under real pressure. Later questions ("what hardships have you overcome," "is there someone who continuously lets you down") open windows into locus of control, emotional regulation, and pity-seeking patterns, while lower-stakes novelty questions ("can you text as quickly with one hand as with two?") get inserted deliberately between heavier ones — Hughes calls this a way to introduce novelty, slightly sharpen focus, and reset the emotional temperature before the next serious question lands.3
Hughes runs roughly two dozen questions total, each with its own diagnostic rationale. A representative cross-section:234
The model's design logic is the same one running through the rest of the book's elicitation material: bury the actual instrument (behavioral observation) inside something that looks like ordinary conversation (a job interview). Hughes is explicit that the questions themselves are close to interchangeable — what matters is what response they're built to elicit, and that the model is meant to be woven into "almost any conversation," not confined to hiring.1 He pairs it explicitly with two other tools already documented in this vault: elicitation technique for surfacing information indirectly, and the Six-Axis Model's Openness axis, which linguistic skill is meant to level up during the same conversation.1
You're running a screening conversation — a hire, a vendor vetting call, even a first real conversation with someone you're sizing up socially. Before you ask anything, decide what you actually need to know: honesty under low stakes, locus of control, deviance tolerance, capacity for creative problem-solving. Open with a low-stakes calibration pair (effective/efficient, successful/significant) and watch — not listen to — the response: latency, eye movement, whether they check your face afterward. Once you have a baseline, move to a question in the register you actually care about, and insert a novelty question (the one-handed texting question, or an equivalent) between two emotionally loaded ones to reset their state before the next probe. Never react visibly to an answer that surprises you; the value of the technique depends entirely on the subject believing this is small talk, not assessment.
Evidence: Hughes presents this as his own field-developed instrument, with a specific institutional adoption claim (an unnamed agency "adopted almost all of it") that cannot be independently verified from the text.1 The individual diagnostic rationales (hesitation patterns, eye-accessing, divergence-testing) draw on genuine, established behavioral-observation concepts, but the specific pairing of question-to-diagnosis is Hughes's own construction, untested outside his own practice. [PLAUSIBLE — needs corroboration] [SINGLE SOURCE]
Tensions: The book insists repeatedly that "there are no bad answers" to these questions — but several of the accompanying rationales (the karma question's "moral food chain" reading, the toilet-papering/cursing pair's deviance framing) clearly do treat some answers as more hireable than others, particularly for roles like sales versus nanny work named directly in the text. The claim of neutrality and the actual diagnostic use pull in different directions.
Psychology — Cognitive Dissonance (Festinger). Several HBI questions (the karma question, the "biggest challenge with a previous employer" question) are explicitly built to surface how a subject resolves uncomfortable self-narratives — whether they take partial blame or externalize it entirely onto a "toxic" employer. Cognitive dissonance theory explains why this diagnostic works at all: a person under dissonance pressure reliably reaches for whichever story restores their self-concept fastest, and the HBI is essentially a battery of situations engineered to force that reach in front of a trained observer, then reads which direction they reached.
Behavioral-Mechanics — The Six-Axis Model. Hughes names this connection himself: HBI skill is meant to be paired with linguistics that level up Openness on the Six-Axis Model during the same conversation.1 The insight the pairing produces: the HBI isn't just a passive assessment tool sitting beside the Six-Axis Model — it's a live-fire diagnostic for it, letting an operator locate a subject's actual Openness baseline in real time rather than guessing at it before deploying the rest of the book's influence machinery.
Sharpest implication: If the actual content of an answer is genuinely secondary to the behavior around it, then most interview training — coaching candidates on what to say — is optimizing the wrong variable entirely, at least from the interviewer's side. An interviewer running the HBI isn't grading your answers. They're using your answers as a delivery mechanism for something else.
Generative Questions: