Two people at a corner table, coffee going cold, arguing about AI. One says it will take every job in the building. The other says it won't touch a thing. Neither is asking a question. Both are reading out a conclusion they arrived at somewhere else, before they sat down, and neither seems bothered that the other one disagrees.
That's the tell. Not the disagreement. The comfort.
Nobody at that table asks which jobs, on what timeline, or what "take" is even supposed to mean. And nobody asks because the questions would slow things down, and slowing down is the one thing this kind of certainty exists to prevent. You get a stalemate, a last sip, and both of them leave more sure of themselves than when they arrived.
You've had this conversation. Maybe not about AI. About the restructure, the new supplier, whichever acquisition is currently being talked about like weather. The subject changes. The shape doesn't.
Move the same two people into a building with a lanyard policy and watch it happen again, dressed differently. Someone leans back in a meeting and says, "well, the truth is," and the room quietly stops arguing. Nobody notices the shift because it doesn't feel like a shift. It feels like relief.
Here's what's actually going on. Certainty and truth do different jobs, and organisations reach for each at exactly the wrong moment.
Certainty turns up early, before anyone has agreed what the words mean, because arriving early is what lets you skip the argument about definitions. Truth turns up late, after something has already been agreed, because by then the room needs it to sound finished rather than provisional. One jumps the queue. The other cleans up afterwards. Neither one is doing the job you'd want it to do, which is telling you what's actually happening while there's still time to act on it.
Most of what you treat as solid at work isn't discovered, it's agreed. Job titles. What counts as strategic. Which projects get called ambitious and which get called risky. None of it fell from the sky. It holds together because enough people keep behaving as though it's fixed, and it keeps working right up until the moment somebody tests it.
But testing it is exactly what a firm "truth" is built to discourage. Once a decision hardens into "the truth", reopening it stops sounding curious and starts sounding disruptive. You can hear the difference. Explanations get smoother than the original decision ever was. Doubt goes missing without anybody announcing it left.
And that's because every organisation is running two contests at the same table, not one. The first is the work: what actually happened, what will actually happen next. The second is the career: who gets to survive what just happened. They usually want different things, and truth, when it finally shows up, tends to serve whichever one is paying closer attention.
Which is why the old line still holds up: nobody ever got fired for buying IBM. If the safe option works, truth calls it professional judgement. If it fails, truth calls it bad timing and moves on. Either way, the person who chose safe never has to explain themselves twice.
Try the same test on the option that wasn't safe. Succeed, and the story tidies itself within the week: the risk drops out, and it becomes the plan all along. Fail, and the patience that was there in the pitch meeting is nowhere to be found in the post-mortem. Judgement gets discussed. Invitations thin out. Nobody announces that either.
Twice. That's how often that person gets to be wrong before the invitations stop, in most rooms.
None of this is a conspiracy. Nobody is sitting in a room deciding to punish good ideas. It's habit, learned one outcome at a time. You watch which stories calm a room down and which ones extend the meeting, and you adjust, the same way everyone else at the table already has. Bold gets proposed less often not because anyone banned it, but because everyone quietly worked out which version of true keeps them employed.
So next time someone tells you "the truth is" three sentences after a decision has already been made, don't hear a fact arriving late. Hear a decision being made comfortable. And next time two people are equally certain about opposite futures before either of them has asked what the other one means by the words they're using, don't referee it. Ask the boring question instead: which jobs, by when, and what would actually count as "take".
Treat truth the way you'd treat stock, not furniture. Furniture gets carved once and sat on for years. Stock gets checked, used, and replaced the moment it's past its date, and nobody thinks that's disrespectful, they think it's how a business stays solvent. Truth that never gets checked again isn't more true for having lasted. It's just older, and it's had longer to get comfortable.
The two at the corner table are still there, still certain, still not asking each other the question that would actually settle it. Their coffee's cold. Somebody should ask what "take" means.