The INTJ Who Optimised the Wrong People Out

You don't dislike most people. That would require feeling something. What you do is faster and quieter: you assess them and assign them a tier. Within a few interactions you've decided whether someone is worth your limited bandwidth — sharp enough to be interesting, reliable enough to plan around, or none of the above and therefore filed under 'low priority.' It doesn't feel like writing anyone off. It feels like resource allocation, the same triage you run on every other demand on your time. You have finite attention and a long list of things to build, so you cut the people who don't clear the bar.

But notice what that decision actually is. You've taken a person you've observed for a handful of moments and permanently set their future allocation to zero. No more attention, no more updating, no more chances to surprise you — because you've already routed your energy elsewhere and stopped watching. You'd never freeze a forecast that hard, that early, on that little data in any domain you respect. People are the one system you run this lazily on, and the cost doesn't show up as a dramatic failure. It shows up as a life that quietly gets narrower while you congratulate yourself on being efficient.

You're confusing a snapshot with the trajectory

Here's the modelling flaw a strategist should hate. When you write someone off, you're treating your first read as if it captured the whole person across time. But you of all people know that systems aren't static — you live in trajectories, second-order effects, what something becomes over years, not what it looks like on day one. You'd laugh at an analyst who valued a company on a single quarter and never looked again. Then you do exactly that with human beings, sampling a person at one low-resolution moment and locking the verdict in for good.

People are slow-revealing systems. The interesting parts — depth, loyalty, a strange competence in a domain you didn't probe — rarely surface in the first few interactions, especially with anyone who, like you, doesn't perform themselves for strangers. The quiet person you filed as 'boring' was reserved, not empty. The one you tagged 'difficult' was testing whether you were worth opening up to. By writing them off on the snapshot, you guarantee you never see the trajectory, and then you mistake your own missing data for proof they had nothing to offer.

The cost is invisible because it's an opportunity cost

The reason this error survives is that it never sends you a bill. When you write someone off and they drift out of your orbit, nothing visibly breaks. There's no failed project, no betrayal, no moment that flags 'you got that wrong.' The connection that would have formed simply never forms, and you can't grieve a relationship you don't know you forfeited. So your judgement looks costless from the inside, which is precisely why you keep doing it.

But opportunity cost is still cost — you understand that everywhere else in your thinking. Every person you optimised out was a potential ally, collaborator, or one of the rare few who could have genuinely met you, and you closed each of those doors on a first impression to save attention you had plenty of. Run the asymmetry honestly. Keeping a file open costs you a little ongoing observation. Writing the wrong person off costs you a connection you'll never even learn existed. A strategist who systematically incurs large invisible losses to avoid small visible ones isn't being efficient. He's miscounting the ledger because half of it is hidden.

Your filter is calibrated to the wrong signal

Look at what your tiering actually measures. You sort people fast on the cues you can read quickly — visible competence, sharpness in conversation, whether they track your reasoning. Those are real signals, but they're the surface ones, the traits that happen to show up early. The qualities that actually make someone worth knowing over a decade — steadiness, decency, the kind of loyalty that acts when you're not in the room — are exactly the ones that never register in a quick scan. Your filter is fast because it's tuned to the shallow variables.

So the people who pass your bar aren't the highest-value people; they're the ones whose value is legible on first contact. That's a selection bias, and you'd flag it instantly in anyone else's process. The person who can't keep up in a debate may be the one who shows up when your plan collapses. The unimpressive first impression may belong to someone whose best traits are simply slow to surface. Optimising for early legibility doesn't find you the best people. It finds you the people who are good at being read early, which is a different and much less useful thing.

Run a longer evaluation window

The fix is a protocol, not a personality change, which is the only kind of advice you'll actually adopt. The instant you catch yourself assigning the tier — 'not worth the effort,' 'nothing here,' 'low priority' — treat it as a flag, not a finding. The speed of the verdict is the tell that you sampled too little, not that the answer was obvious. Then, instead of cutting allocation to zero, drop it to a small non-zero amount: stay observant a few interactions longer, ask one real question, leave the file open.

You're not lowering your standards or pretending everyone is fascinating. You're correcting a sampling error by extending the evaluation window before you commit to a conclusion you can't easily reverse. Most people you watch a little longer will confirm your first read — and now it's confirmed instead of assumed, which is strictly stronger. But a meaningful fraction will reveal something the snapshot couldn't show, and those are the connections your quick verdicts were quietly shutting out. Holding the file open costs you almost nothing. Closing it early costs you exactly the people most worth keeping.