Daniel sent us this one — he's asking whether some people are genuinely better at reading others than the rest of us, whether introverts have a particular edge in spotting deception, and if there's an actual framework civilians can learn for reading people and avoiding being lied to. And he frames it around this quiet superpower he seems to have, noticing when something's off in high-pressure interactions like property deals, where his gut just flags things his wife misses. Meanwhile she's an architect who sees building details he'd never catch. So there's a division-of-labor question here too.
It's a really good question, because it sits right at the intersection of psychology, evolutionary biology, and — this is the part I find fascinating — something called thin-slice judgment. Which is the technical term for what he's describing when he says his gut flags something in seconds.
Thin-slice judgment. Sounds like a deli counter technique.
It kind of is, metaphorically. The term comes from this landmark study by Nalini Ambady and Robert Rosenthal back in nineteen ninety-two. They showed participants silent ten-second video clips of professors teaching, with no audio, and asked them to rate the professors on things like confidence, warmth, competence. Those ten-second ratings predicted student evaluations at the end of the semester with remarkable accuracy. And the judgments held.
You're telling me students spend an entire semester forming opinions they've basically already formed before the professor opens their mouth.
That's exactly what the data suggests. And Ambady later shortened the clips to two seconds. Then she tried still photographs. The brain is making assessments at a speed that conscious reasoning can't keep up with.
Which would explain the vibe-check thing. The feeling that something's off before you can articulate why.
And here's the crucial distinction — thin-slicing isn't the same as snap judgment or prejudice. Snap judgments are when you make a decision with insufficient information and refuse to update it. Thin-slicing is when you extract the most diagnostic information from a very narrow window of experience and it turns out to be valid. The difference is whether the thin slice actually contains the relevant signal.
The question isn't whether your gut is fast. It's whether your gut is tuned to the right frequency.
That's a beautiful way to put it. And this connects directly to what the prompt is describing. Some people's guts are tuned to interpersonal frequencies. Other people's guts are tuned to spatial frequencies — like an architect who can walk into a room and immediately know the ceiling height is wrong or the proportions are off. Neither is a superpower in the comic-book sense. They're both examples of expertise-based pattern recognition operating below conscious awareness.
Hannah's architecture-trained pattern recognition and Daniel's people-reading pattern recognition are the same underlying mechanism applied to different domains.
And this is where the research gets really interesting. Paul Ekman, who's basically the grandfather of deception detection research, spent decades studying microexpressions — these fleeting facial expressions that last between a fifteenth and a twentieth of a second. They're involuntary. You can't suppress them. And they reveal the emotion someone is actually feeling before their conscious brain slaps on the socially appropriate expression.
A fifteenth of a second. That's faster than I move.
Most people don't move that fast. But here's the thing — Ekman found that most people are terrible at reading microexpressions. The average person performs at roughly chance level when trying to detect deception. About fifty-four percent accuracy. A coin flip plus a rounding error.
Fifty-four percent. So the entire human race is walking around thinking they're human lie detectors and they're statistically indistinguishable from guessing.
It's worse than that. There was a meta-analysis by Charles Bond and Bella DePaulo in two thousand six that looked at more than two hundred studies involving over twenty-four thousand participants. The average accuracy for detecting lies was fifty-four percent. For detecting truths, it was slightly better — about sixty-one percent. But here's the kicker: professional lie-catchers — judges, police officers, psychiatrists, customs officials — they don't perform any better than college students. In some studies, they actually perform worse, because overconfidence makes them less careful.
The people whose jobs depend on this skill are, statistically, no better than someone off the street.
With one massive exception. And this is the part that I think answers the first part of the prompt directly. The Secret Service.
The Secret Service.
The Secret Service. Ekman and his colleague Maureen O'Sullivan conducted a study where they tested over thirteen thousand people from various professions on their ability to detect deception. They identified a tiny group — less than one percent of participants — who scored exceptionally high. They called them wizards. Deception detection wizards.
Of course they did.
The single occupational group most overrepresented among these wizards was Secret Service agents. About fifty percent of the Secret Service agents tested scored in the wizard range. Not all of them. No other professional group came close.
What are they doing differently?
This is where it gets instructive. The wizards weren't using any single tell. They weren't looking for the classic signs — fidgeting, eye contact avoidance, all the stuff pop culture teaches you. Those aren't reliable indicators. What the wizards did was integrate multiple channels simultaneously — facial expression, body language, vocal tone, semantic content, and crucially, baseline behavior. They'd establish what's normal for a person in a low-stakes interaction before evaluating what changes under pressure.
So they're not reading the lie. They're reading the deviation.
That's it. That's the whole thing. A specific behavior — someone touching their nose, looking away, crossing their arms — means nothing in isolation. People have different baselines. Some people naturally avert their gaze. Some people are fidgety by nature. The signal isn't the behavior. The signal is the change in behavior when the stakes shift.
Which means you can't read someone you've just met with any reliability. You need a reference point.
And this is why the prompt's description is so interesting. He mentions he's been in enough of these transactions to know his gut is rarely wrong. That's the expertise component. He's built a mental library of baseline behaviors for people in property-deal contexts. He knows what normal looks like for an agent showing a property. So when something deviates from that pattern, his pattern-matching system flags it.
It's not a superpower. It's a database.
It's a database running on hardware he didn't know he was training. And this is where introversion becomes relevant. There's actually research on this.
I was wondering when you'd get to the introvert angle.
It's not that introverts are inherently better at reading people. That would be an oversimplification. But introverts do tend to process social information differently. There's a body of work by psychologists like Jonathan Cheek and Jennifer Grimes that suggests introverts spend more time in reflective observation mode rather than active social engagement mode. They're watching the interaction more than they're participating in it.
Which means they're collecting more data points.
They're processing them at a different depth. Extraverts tend to process socially in real time — they're thinking while talking, adjusting on the fly, drawing energy from the interaction itself. Introverts tend to do more post-hoc processing, but they're also doing something during the interaction that's subtle — they're monitoring. They're tracking patterns. They're not as cognitively occupied with maintaining the social flow, so they have bandwidth for pattern detection.
The introvert edge isn't some mystical sensitivity. It's just freed-up processing power.
Freed-up processing power and a natural inclination toward systematic observation. And here's a really interesting finding — there's research by Adam Grant and others showing that introverts are better at what's called integrative complexity. They're more likely to hold multiple perspectives in mind simultaneously, to notice contradictions, to see when things don't add up.
Which is exactly what the prompt describes. Noticing specific things that didn't add up.
And this connects to a concept from the deception literature called leakage. When someone is attempting to deceive, the cognitive load is enormous. They're managing the story, managing their emotional response, managing their self-presentation, managing the listener's reactions. Under that load, the truth leaks out through channels they're not monitoring as carefully. Inconsistencies in the narrative.
The deceiver is doing a juggling act, and the introvert observer is just sitting there noticing which ball is about to drop.
That's a very Corn way to put it. And the research supports this. There's a study by Aldert Vrij, who's one of the leading deception researchers — he's at the University of Portsmouth — that found that inducing cognitive load in the person being questioned significantly improves detection accuracy. When you ask someone to tell their story in reverse chronological order, for example, liars struggle much more than truth-tellers. Because constructing a false narrative in reverse is cognitively devastating.
There's a whole methodology around making it harder for liars to keep their story straight.
And this gets to the third part of the prompt — is there a framework civilians can learn? The answer is yes, and it's not what most people think. It's not about learning to spot specific tells. It's about learning to structure interactions in ways that amplify the signal.
Walk me through it. If I'm a regular person, no training, what's the framework?
I'm going to give you what researchers call the strategic questioning approach, and I'll organize it into something usable. First: establish baseline. Before you ask anything high-stakes, spend a few minutes on neutral topics. Observe how the person speaks, gestures, makes eye contact, breathes, holds their posture. This is your control condition. Everything that follows is measured against this.
Small talk isn't just social lubricant. It's calibration.
It's calibration. Second principle: ask open-ended questions and let them talk. The biggest mistake people make in trying to detect deception is they talk too much themselves. They fill the silence. They telegraph what they're looking for. Instead, ask a question like "Can you walk me through the timeline of how this property came on the market?" and then shut up. Let them fill the space.
Because the more they talk, the more data you get.
The more cognitive load they're under if they're fabricating. Third: listen for what's not being said. Truthful accounts tend to include contextual details — sensory information, irrelevant specifics, things that don't directly serve the narrative. Fabricated accounts tend to be cleaner. They're stripped of the messiness of real memory. If someone's story is suspiciously linear and every detail serves the point they're trying to make, that's a flag.
Real life is messy. Lies are tidy.
Fourth: introduce unexpected questions. If you suspect you're being deceived, ask something that doesn't fit the script they've prepared. "What was the landlord's reaction when you told them about the water damage?" If they've rehearsed a story about the property's condition, they might not have rehearsed the landlord's emotional response. The hesitation is informative.
You're basically creating a pop quiz for a story they haven't studied for.
That's the cognitive load approach. Fifth: pay attention to clusters, not isolated behaviors. One thing — a nose touch, a glance away, a hesitation — means nothing. But if you see a cluster — voice pitch rises, posture shifts, and a microexpression of fear flashes across their face all within the same two-second window — that's a signal worth noting.
The framework is: establish baseline, shut up and let them talk, listen for what's missing, throw curveball questions, and look for clusters. That's five principles. That's learnable.
It's entirely learnable. And here's what I find so encouraging — there's research showing that even brief training in these principles significantly improves accuracy. A study by Porter and ten Brinke in two thousand ten found that just a few hours of training in systematic observation techniques moved detection accuracy from the mid-fifties into the mid-seventies percentage-wise. It's not wizard-level, but it's a meaningful improvement.
Seventy-five percent accuracy versus fifty-four. That's the difference between guessing and actually having a useful skill.
There's an important caveat here that I think gets missed in a lot of popular discussions. None of this is about becoming a human lie detector in casual social situations. That's a terrible idea. The research is very clear that trying to constantly analyze your friends, family, and colleagues for deception destroys trust and relationships. The frameworks are for high-stakes situations where deception has real consequences — negotiations, major purchases, safety-relevant interactions.
That's an important boundary. This isn't a party trick.
It's absolutely not. And the people who treat it like one — who walk around thinking they can read everyone — are usually the worst at it. The Dunning-Kruger effect applies brutally to deception detection. The most confident people are often the least accurate.
The fact that the prompt describes this as something that just happens to him, almost passively, rather than something he actively does, is actually a point in his favor.
The wizards in Ekman's research didn't describe themselves as actively trying to catch liars. They described it as something that just registered for them. The pattern recognition was automatic, not effortful. That's a hallmark of genuine expertise in any domain — the expert doesn't have to consciously work through the steps. The recognition just happens.
Which brings us back to the division of labor with Hannah. She sees building details he doesn't. He sees people details she doesn't. Is there any research on how couples or teams divide perceptual labor?
There's fascinating work on this from organizational psychology. The concept is called transactive memory — the idea that groups develop distributed systems for encoding, storing, and retrieving knowledge. In a long-term relationship, partners implicitly divide cognitive responsibilities. One partner handles social calendar management, the other handles financial monitoring. One remembers names, the other remembers directions.
You don't both need to be good at the same things. You just need to know who's good at what.
And you need to trust that division. The prompt says his wife is often surprised by his assessments. That suggests she doesn't share his perceptual filter for interpersonal data. But the fact that his track record is impressive means the division works — her skepticism probably keeps him honest, and his accuracy means the information is valuable.
It's like having an in-house quality control process. She's the second set of eyes that forces him to articulate what his gut is flagging.
Articulation is key. Because one of the risks of relying on gut instinct is that you can't explain your reasoning to others. The framework we just discussed — baseline, open questions, listening for gaps, unexpected probes, clusters — that gives you language for what your gut is doing. It lets you say "three things didn't add up in that conversation" instead of "I don't know, something felt off.
Which is more persuasive and also more checkable. You can go back and verify whether those three things were actually inconsistent or whether you were just in a suspicious mood.
And this is where I want to add a note of caution about the research. There's been some replication controversy around certain deception detection findings. Some of Ekman's microexpression work has been harder to replicate than initially reported. The -analyses are more reliable than any single study. The broad consensus is that deception detection is difficult, most people are bad at it, training helps somewhat but doesn't make you infallible, and the best approach is to use structured information-gathering rather than trying to intuit truth from facial expressions alone.
The microexpression stuff is real but overhyped, and the real value is in the strategic questioning approach.
That's the current state of the science, yes. And I think it's worth mentioning that there's a whole parallel literature on interrogation and interviewing that's relevant here. The Reid technique, which was dominant in American law enforcement for decades, was essentially a confrontational approach designed to break down resistance and extract confessions. It produced false confessions at alarming rates.
Because it was designed to confirm what the interrogator already believed.
It was confirmation-bias-driven. The alternative that's gained traction in the last fifteen years is called the PEACE model — it originated in the UK. Planning and preparation, engage and explain, account, closure, evaluation. It's non-confrontational. It's information-gathering rather than confession-seeking. And studies comparing the two approaches consistently show that PEACE-style interviewing produces more accurate information and fewer false confessions.
The same principle applies whether you're a cop or a civilian looking at a rental property. Information-gathering beats confrontation every time.
And I think that's the most practical takeaway from all of this. If you walk into a high-stakes interaction thinking "I'm going to catch this person in a lie," you've already compromised your ability to actually detect deception. You're primed to see confirmation, not information. If you walk in thinking "I'm going to gather as much information as possible and pay attention to what doesn't fit," you're in a much better position.
The mindset shift is from detective to scientist. You're not building a case. You're running an observation.
And the prompt's description of how he operates — noticing small things, feeling that the general vibe is off, identifying specific inconsistencies — that maps perfectly onto the scientist mindset. He's not interrogating. He's observing. And his brain is doing the pattern-matching automatically because he's logged enough hours in these situations.
If someone wants to develop this skill deliberately, the prescription would be: one, get exposure to the domain you want to read well. Two, learn the five-point framework. Three, practice articulating what you're seeing. Four, stay humble about your accuracy. And five, don't try to use it on your spouse.
That last one might be the most important. There's a reason the research shows that couples who try to read each other's deception are miserable. The framework is for asymmetric information situations — business transactions, negotiations, situations where the other party has incentives to misrepresent. It's not for relationships built on trust.
Unless you're trying to figure out who ate the last piece of pizza.
In which case the framework collapses instantly because both parties are equally motivated to deceive and the stakes are existential.
For sloths, pizza is always existential.
You and the pizza origin story again.
I'm not letting that go. The historical record is incomplete.
The historical record, as far as I can tell, consists entirely of you saying it happened.
Oral tradition is tradition.
Moving on from sloth mythology — I want to circle back to something about the introvert piece, because there's a nuance here I think is worth unpacking. It's not just that introverts have more processing bandwidth during social interactions. There's also evidence that introverts are more sensitive to what's called prediction error.
That's when reality doesn't match your brain's forecast.
The brain is constantly generating predictions about what's going to happen next — what someone will say, how they'll react, what the outcome will be. When reality deviates from prediction, that generates a neural signal — literally a spike in dopamine neurons in some cases — that says "pay attention, something unexpected just happened." Some people's prediction-error signals are more sensitive than others.
Introverts tend to have more sensitive prediction-error detection?
There's some evidence for this, though it's not conclusive. But it would explain the phenomenon the prompt describes — the feeling that something is "off" without being able to immediately identify what. Your brain's prediction system registered a deviation before your conscious mind could categorize it.
The "vibe check" is literally your prediction-error system firing. The vibe isn't a feeling. It's a calculation.
A calculation you don't have conscious access to, but a calculation nonetheless. Your brain has modeled what a normal property showing looks like based on dozens of prior experiences. When the current showing deviates from that model in a way that crosses some statistical threshold, the prediction-error signal fires. You experience that as a gut feeling or a vague sense that something's off. Only later, if you're good at this, can you trace the feeling back to specific inconsistencies.
Which means the skill isn't in having the feeling. The skill is in being able to reverse-engineer it.
That's the whole game. Everyone gets gut feelings. The difference between someone who's good at reading people and someone who isn't isn't in the initial feeling — it's in what they do with it. Do they dismiss it? Do they act on it blindly? Or do they use it as a starting point for systematic observation?
That's the framework again. The gut flags something. Then you go to the framework — establish baseline, ask open questions, listen for gaps, introduce unexpected probes, look for clusters. The framework is how you test whether your gut was right.
The gut is the hypothesis generator. The framework is the hypothesis tester. Neither is sufficient alone.
There's something almost scientific-method about this. Your unconscious brain generates a hypothesis — "this person is being deceptive." Your conscious brain then designs observations to test that hypothesis. And you update your belief based on the evidence you gather.
The best practitioners — the wizards, the people the prompt describes — they do this so fluidly that they don't even experience it as a two-step process. The hypothesis generation and testing happen simultaneously, in real time, below the level of conscious deliberation. That's what expertise looks like in any domain. The chess grandmaster doesn't consciously evaluate moves. The pattern just registers.
The short answer to the prompt's first question — "are some people better at reading people than others?" — is yes, but it's not a fixed trait. It's a learnable expertise that some people develop implicitly through experience and that others can develop deliberately through training.
That's the synthesis. And the second question — "is there a framework civilians can learn?" — also yes, and we've outlined it. Baseline, open questions, listen for gaps, unexpected probes, look for clusters. Plus the -principles: information-gathering over confrontation, humility over confidence, and don't use it on people you love.
The introvert edge is real but it's not magic. It's just that introverts tend to naturally do more of what the framework prescribes — observe more, talk less, process at depth.
Which means extraverts can learn this too. They just might have to work a little harder at the "shut up and observe" part.
I've been telling you that for years.
You have, and I've been ignoring you, which I believe is consistent with the research on extravert behavior.
The research also says extraverts are worse at detecting deception because they're too busy being charming.
I don't think the research uses the word "charming.
It's implied.
Let me add one more thing that I think is relevant here. There's a concept from the deception literature called the truth-default theory, developed by Timothy Levine. The idea is that humans have a truth-default — we assume people are telling the truth unless we have strong evidence otherwise. And this is actually adaptive. In most interactions, most of the time, people are being truthful. If we walked around constantly suspicious of everyone, social life would be impossible.
The baseline assumption of honesty is a feature, not a bug.
It's a feature that sometimes gets exploited. But here's the interesting thing — people who are good at deception detection seem to be people who can temporarily suspend the truth-default in specific contexts without becoming globally suspicious. They're not paranoid. They're contextually vigilant.
That's a good phrase. In a property negotiation, vigilance on. At dinner with friends, vigilance off.
The ability to toggle that switch appropriately seems to be one of the characteristics of the wizards. They're not walking around thinking everyone is lying. They're just able to shift into a higher-observation mode when the situation calls for it.
Which the prompt describes perfectly. He's not suspicious of everyone. He's good at reading people in specific high-pressure interactions where deception has incentives.
He knows his limits. He doesn't claim to be able to read buildings. That's Hannah's domain. Knowing what you can't do is as important as knowing what you can.
There's a humility in that division of labor that I think is underrated. A lot of people who fancy themselves good at reading others try to apply it universally. They read motives into everything. They become insufferable.
The armchair psychoanalyst at the dinner party who's figured out everyone's childhood trauma based on how they hold their wine glass.
And they're almost certainly wrong, because they're operating without baseline, without domain expertise, and with massive overconfidence.
The fifty-four percenters who think they're ninety percenters.
The real superpower isn't reading people. It's knowing when you can and when you can't, and staying humble either way.
That's the conclusion the research keeps pointing toward. The best deception detectors are characterized not by swagger but by caution. They're more likely to say "I think something might be off here, let me check further" than "this person is lying.
Which is a much less dramatic skill, but a much more useful one.
A much more learnable one. Anyone can learn to say "let me gather more information before I conclude." Anyone can learn to ask open questions and listen for gaps. Anyone can learn to look for clusters rather than single tells. This isn't a mystical gift. It's a set of habits.
Habits that introverts might find slightly easier to adopt, but that are available to anyone willing to practice them.
Worth practicing, given how much of life involves high-stakes interactions with information asymmetry. Buying a house. Negotiating a salary. Choosing a contractor. These are all situations where the other party has incentives to manage the information you receive, and where a little bit of structured observation can save you a lot of grief.
The framework doubles as a life skill and a financial protection mechanism.
And I want to emphasize one last point from the research. The single most effective deception-detection technique that shows up across studies is surprisingly simple: just ask more questions. Not accusatory questions. Not leading questions. Just open, curious, information-seeking questions that let the other person talk. The more they talk, the more data you have. The more data you have, the harder it is for a deceptive narrative to hold together.
The entire framework could be reduced to "ask questions, shut up, pay attention.
That's the ninety-second version. The full version adds the structure around baseline and clusters and cognitive load, but the ninety-second version will get you surprisingly far.
I appreciate frameworks that compress well.
They're the best kind. And they're also the kind people actually use. Nobody's going to remember a twenty-point checklist in the middle of a tense negotiation. But "ask questions, shut up, pay attention" — that you can hold in your head.
The prompt's author seems to be doing this naturally already. The framework just gives him language for what he's already doing and maybe some techniques for getting even better at it.
That's exactly what good frameworks do. They don't replace intuition. They give intuition a structure to work within. They make the implicit explicit. They let you teach other people what you've learned.
He could actually teach Hannah some of this, and she could teach him how to spot bad architecture.
Cross-training in perceptual superpowers. I love it. Though I suspect the architecture-reading might be harder to compress into a five-point framework.
"Look at the ceiling, check the proportions, notice the light." There, three points.
You've just summarized four years of architecture school. Hannah's going to be thrilled.
I'm a synthesizer.
You're a sloth who thinks he invented pizza. But yes, you're also a synthesizer.
And now: Hilbert's daily fun fact.
Hilbert: In the nineteen eighties, linguists working on Inuktitut documentation in the eastern Canadian Arctic discovered that the language's polysynthetic word-building capacity had an unintended consequence — a single Inuktitut verb could accidentally encode an entire English sentence's worth of semantic nuance, which meant that machine translation efforts of the era, already struggling with French, simply crashed when fed Inuktitut input. The systems would attempt to parse what looked like one word and produce output buffers so long they overflowed. Some early NLP researchers in Eritrea, working on a parallel project for Tigrinya, referenced these crashes in their grant proposals as a cautionary tale about the limits of Indo-European assumptions in computational linguistics.
Eritrea enters the chat.
This has been My Weird Prompts. Thanks to our producer, Hilbert Flumingtop. If you enjoyed this episode, leave us a review wherever you listen — it helps other people find the show. We'll be back next week.
