What's running you?

Your threat system has found enough ambient uncertainty to stay warm, but not enough specific data to tell you what to do — so the body mobilises and then stands there, ready for something unnamed. This is why reassurance keeps failing: it addresses the story your mind built afterwards, not the activation that came first.

The system was built for short bursts followed by real recovery. What you're running is a line instead: continuous input, no resolution, never fully standing down. The wired-and-tired combination is the signature — too activated to rest, too depleted to work well, and the baseline has crept up far enough that you've stopped noticing it.

If weekends used to restore you and now don't, you're not short of sleep — you're running an unfavourable exchange that a pause can't repay. The giveaway is dread arriving before the demand: Sunday evening doing to you what Monday used to. That's not fatigue. Fatigue is a shortage of sleep; this is a shortage of everything.

Something got assessed as wrong and the energy to correct it was released before you were consulted. The detection is usually accurate — something did cross a line. What's unreliable is the scale, the target, and whether it's even about today, because current anger carries every similar unresolved one behind it.

Physical heaviness, disrupted sleep, appetite off, energy low — this is what processing a large load looks like from inside, and it's a common output after a loss of any kind, including ones that weren't a death. It can also be depletion, or something physical. The point of naming it is to stop reading it as a character problem.

Thinking is supposed to end in a small click where you know enough and the question gets released. That click is produced by certainty, not effort — so when a situation genuinely doesn't support certainty, it never arrives, and your mind reads its absence as instruction to keep going. The loop isn't malfunctioning. It's waiting for a signal the situation can't generate.

The replay is looking for an account in which you came out better, and that account isn't available — the conversation is finished and the data is fixed. So it can't close on its own terms. It'll keep offering you the scene, each time with a slightly better script, and none of them can be submitted.

Your body reacts to imagined events with real chemistry, which is why rehearsing a conversation forty times costs you something close to what having it forty times would. The rehearsal feels like preparation. Mostly it's building a case against an opponent you've invented, and you'll arrive braced for that version regardless of who turns up.

You're evaluating a choice made with the information available then, using information that only arrived later. That isn't a fair test, and whatever verdict you reach answers a question the earlier version of you was never asked. Meanwhile the parallel life you keep visiting was built to order — it contains what you want now and none of the costs the real path would have carried.

More processing isn't producing more clarity — it's producing more doubt, because each pass invents new scenarios rather than new facts. Somewhere in here the honest answer is usually already present and is being declined because it costs something. The deliberation is doing work, just not the work it appears to be doing.

There's a genuine assessment in there before the answer comes out — I don't want to, I disagree, I'm exhausted — and a faster process is overwriting it with a different priority: keep them comfortable. The tell is what your body does immediately after you agree. A true yes settles. A pleasing yes braces.

Addressing it looked too expensive at the time, so it went below instead of being expressed — which converted an acute signal into a background process with no closing condition. That's why time hasn't helped. It's also distorting: everything they do now gets read through the accumulated pile, so neutral behaviour keeps arriving as evidence.

If you've said it three times and nothing changed, what you've been making is a request — and requests can be refused, forgotten, or agreed to and ignored. A boundary is the other half of that sentence, the half almost nobody says out loud: what you will do if it happens again. That part needs nobody's cooperation.

There's the thing they did, and separately there's the collapse of a model you were navigating by — and the second is usually larger. It's why the past starts rewriting itself and why the question that follows isn't about them at all: if I was wrong about this, what else am I wrong about?

Check the evidence — things that happened, that another person could verify, not tone or a short reply. If the list is thin, this is a model built somewhere the thing genuinely wasn't safely held, and it's being applied to someone it wasn't built from. The damage usually comes from reorganising your behaviour around the prediction, which tends to produce it.

A working apology has three parts — what you did, that it was yours, what changes — and the middle one is where nearly all of them fail, because your self-image defends itself first. That produces the standard distortions: it wasn't that bad, they're oversensitive, look what they did. None of that makes you a bad person. It's a story protecting itself in real time.

Something in the task is being declined, and it's usually one of three: you don't actually know the first physical step, you're afraid of the verdict that finishing invites, or the work belongs to a version of yourself you're performing but don't believe in. They have different fixes, which is why generic advice about this fails so reliably.

If this were about standards, finishing would feel good. It doesn't, which is the clue. The bar is set where the work can't be assessed — and unassessed work can still be excellent. It works, in its way. It just charges you the entire body of work to do it, which is why the projects die at ninety percent rather than at the start.

Each individual deferral had a plausible reason — that's how this works, and it's why you can't catch it by examining one decision. The pattern is the evidence. And the physical response is the giveaway: a genuine priority question produces a thought, a reroute produces something in the chest or the stomach.

Your reward system funds effort on observed returns, and before you begin there's nothing to observe — only a projection, which it's entitled to decline. Start, and it gets real data: this is moving, that exists, the room is half done. The funding tends to follow. You don't have a motivation shortage, you have an ordering error.

Sustained focus is an override of a scanning default, it costs energy, and most people get two to four genuinely deep hours a day. The error almost nobody notices is spending that budget on low-stakes input first and then meeting the work that needed it with whatever's left — then reading the result as a discipline problem.

You're running an old evaluation while continuing to accumulate evidence you don't log. Feelings will keep insisting you're getting away with something; they aren't consulting the data. This is a records problem rather than an emotional one, which is good news — records can be updated and feelings mostly can't be argued with.

Guilt says you did something bad; this says you are something bad — and because it aims at all of you, there's nothing to repair, which is why apologising and making amends leaves it unchanged. Often there's no specific act underneath at all, and that absence is itself the finding: a verdict with nothing under it isn't a response to anything you did.

There are two kinds and they feel identical: one reports that you broke your own code, the other is somebody else's code installed early and still firing. Worth sorting, because only one of them has a repair attached. And if it's been processed and repaired already, what's left isn't information — it's a loop.

You have the complete file on yourself — the cost, the doubt, the parts that nearly didn't work — and a curated exterior of theirs. That comparison would come out against you even if your life were objectively the better one. And the reference pool has no edges any more, so what used to be occasional useful data is now a permanent deficit reading reporting nothing true.

The reading sets what registers as normal, so treatment that matches it doesn't trip any alarm — being treated as worth less feels unremarkable rather than outrageous. That's why nothing from inside flags it, and why the gap only becomes visible when you borrow a correctly calibrated instrument.

Flatness about one task is ordinary. Flatness across everything, including things you used to like, sustained for weeks, is a different reading — and it's the one case here where the honest answer is that a page can't help you much. Burnout tends to be domain-shaped and improves when the demand is removed. This doesn't, and it comes on holiday with you.

It's checking whether connection occurred, not whether contact did — past logistics, seen rather than performed-for, exchange rather than two broadcasts, more resourced rather than less. Miss those and the reading stays low regardless of attendance. Which is why a busy week can leave you lonelier than a quiet one.

This usually turns up when the immediate problems are handled — which is why it so often follows getting what you wanted rather than failing to. Nothing is wrong, and that's precisely the condition under which the larger question becomes audible. It isn't a malfunction and it doesn't resolve by finding a better distraction.

Knowing and wanting are produced by different systems and they don't have equal access. Your assessment can be entirely accurate about the cost while the reward circuit continues reporting that this is the most important available thing — and that circuit is older and wired deeper. The clarity is real, and it loses. That isn't weak character.

The control drive scales with perceived threat, and under sustained uncertainty the reach keeps growing until you're managing other people's behaviour and other people's opinions. It never succeeds — it just spends effort on variables that don't respond. The hard part isn't letting go; it's that letting go registers as exposure, because the machinery equates control with safety.