Years ago, in a self-defense class, the instructor set up an impossible position. You are on your knees, ankles crossed, hands laced behind your head. Someone stands two metres back with a pistol to your skull. What is the correct move?

There is no move. You comply. The lesson he wanted wasn’t about the position. It was about everything upstream of it: the street you shouldn’t have walked down, the situation you should have read earlier, the chain of small and individually reasonable decisions that delivered you, kneeling, to a place with nothing left in it. Look up sooner. See it coming. By the time the gun is at your head, all the interesting choices are already behind you.

I’ve carried that parable around for the better part of two decades, and I’ve started to think it describes the wrong moment. The lesson is fine as far as it goes. But almost none of the hard situations I’ve actually been in had a gun in them, or any clean terminal point at all. I wasn’t kneeling with the decision already made for me. I was still moving, still able to act, and I just couldn’t see where any of it led.

There’s a familiar piece of wisdom for exactly this trap, and I believe most of it: look up. Raise your eyes from the task in front of you and ask whether the thing you’re building is still worth building. Crunch is dangerous precisely because it takes that question away — you stop being someone who looks up, and the thing you should have asked goes invisible from inside the work. I’ve made that case myself. But it smuggles in an assumption: that looking up would have helped, that the answer was sitting out there for anyone who raised their eyes. Sometimes it is. Sometimes you raise your eyes and there is nothing out there but fog.

In the fog, the only thing that stays manageable is the next small thing. You answer the email. You ship the small fix, keep the standup, make the dinner, do the dishes. There’s a story we tell about this, and it’s a good one: one foot in front of the other, progress measured in inches but progress all the same. I’ve told it to myself and meant it. But the cadence wasn’t carrying me toward anything. It was keeping me from having to look at the question I couldn’t answer — the dishes done carefully while the house burns.

Which is strange, because that same cadence is the disease. The routine that crowds out the question, the wrapping work that leaves no room to look up: keeping the rhythm going is how the question gets buried in the first place. And yet the rhythm is also the only thing holding me upright. Same motion, opposite verdict. So either I’m contradicting myself, or there are two cadences and the whole problem is telling them apart.

I can’t tell them apart. From the inside, the rhythm that’s saving you and the rhythm that’s burying you feel identical. Whether the motion is wisdom or avoidance depends on the ground you happen to be crossing, and the ground is the thing you can’t see. I met a bloodless version of this years ago, doing standards work. We contributed to a slew of standards bodies, each with its own meeting calendar and submission deadlines, and most of the time the calendars stayed politely out of each other’s way. Every so often they collided. When they did, every team needed the same simulation capacity in the same two weeks, and there was nowhere near enough to go around — this was before you could rent a hyperscaler’s worth of compute by the hour. We never solved it. The closest we came was to look further ahead, guess who else would need the machines and when, pace our own work to leave ourselves room, and otherwise accept that some quarters the collision would simply happen and we would eat it.

Even there, in the tame version, the right move wasn’t always to keep grinding. Sometimes it was to stop — to not start a long run you couldn’t finish before the deadline took the machines away, to sit on your hands and hold the capacity for when it mattered. There’s a reason mountaineers caught in a whiteout are told to stop and dig in rather than walk. In zero visibility, motion is how you stroll off the edge you can’t see. The comforting little asymmetry I’d been running on — that standing still is a guaranteed nothing while moving at least buys a chance — turns out to hold only in some terrain. In other terrain it reverses, and the thing that feels like surrender is the only thing that keeps you alive.

The comfort of motion is exactly what I should trust least. I once spent two years on a project where nobody froze. We kept perfect cadence the whole time, shipping and planning and hitting our dates, building toward a market that had quietly closed while we worked. It felt like motion because the legs were moving. Stasis wears motion as a disguise, and it’s a good one, because from the inside it is indistinguishable from progress.

If I can’t tell the saving cadence from the killing one, and I can’t read the terrain, then the choice to keep going or to stop is not being made on the merits. It can’t be. The merits are precisely the thing I don’t have. It is being decided somewhere else, and the basis isn’t flattering. I keep going because going is the failure I can forgive myself for.

I keep going because going is the failure I can forgive myself for.

Stopping and continuing can land me in the identical bad place. The outcome doesn’t sort them. What sorts them is the account I get to give afterward. “I kept trying” is a story I can live inside. “I stopped” is one I would have to forgive, and I’m not sure I know how. So stopping isn’t the worst option. It’s the one I can least forgive. Those are not the same thing, and I’ve spent a fair amount of my life mistaking the second for the first.

I can’t decide whether that’s the most human thing about me or the most dangerous. Maybe the refusal to be the kind of person who stops is what keeps me a person worth being, the part that won’t fold when folding would be easier. Or maybe it’s a rigged accounting that walks me off cliffs to protect a self-image, the same mechanism that keeps a room full of capable people holding their cadence toward a chart that already says they were wrong. I don’t know which. And I can’t tell from the inside any better than I can read the terrain. The conscience is as fogged as the road.

Most of what’s gone wrong, I can at least argue about — the standards collisions, the closed market, the cliffs you walk off in the fog. They have outcomes. The deadline arrives, the verdict lands, the terrain you couldn’t read resolves into something you can finally name, and you find out. A few decisions never give you that.

The gun was never the point of that parable. The corridor was. You don’t take one wrong turn. You take a hundred, each reasonable standing in it, each forgivable on its own, none of them the kind you get to walk back out of, until the choices fuse into something you can’t take apart again. They stop being decisions. They become the ground you stand on. After that, whether you chose right is permanently unanswerable, and for a reason worse than missing data: there’s no counterfactual to check against, and the version of you who could have judged it cleanly is gone, because making the choice is part of what made you. You only ever get to live one side. You don’t get the verdict. Not later, not ever.

So I keep moving. Not because I’ve found a reason that survives being looked at, but because stopping is the one failure I’ve never learned how to forgive. Some days that feels like the bravest part of me. Some days it looks exactly like how a man ends up on his knees, having made every small decision sensibly, in a place with nothing left in it. I only ever ask which one it is while I’m still moving, never standing still, never long enough for the question to cost me anything.