There is a sentence I said to myself for a while before I was willing to hear it: "I am using the agent a lot more lately — so why do I feel more tired?"
The first reflex is to blame myself: I must not be skilled enough, I have not learned enough. The second reflex is to blame the agent: it must have gotten worse lately. Both are comforting in their own way — one gives me more work to do, the other gives me someone to fault. Both are wrong.
The harder truth: my approach had gone stale. It once fit the kind of work I did at the start. The work changed; the approach did not.
01The inertia trap: reluctant to drop it because you already invested
Your current approach has an unfair advantage: you have already paid for it. You put in the work to build those templates, you got used to the rhythm, you memorized the little tricks. All that effort creates a pull: dropping it would be a waste.
But "waste" is a backward-looking way of thinking. The effort you already spent is money already gone — it does not come back whether you keep the approach or drop it. The right question is not "how much have I invested in this approach," but "from today forward, does this approach still pay off." Those two questions produce very different decisions, and the trap is that we almost always answer the first while believing we are answering the second.
Inertia is especially smooth with an agent, because everything still runs. There is no crash forcing you to stop and look. The old approach still produces output, still gets the job done — it just costs you a little more each time to keep it done. And because there is no breaking point, you have to learn to read small signals in place of the loud noise that never comes.
02Three signals an approach has gone stale
One lone signal might just be a bad week. Two at once are worth noticing. All three, and it is no longer about you failing or the agent being bad — the approach has expired, and the move is not to try harder, but to change.
The third signal deserves a longer pause, because it is the only one you feel before you can measure it. The first two need you to watch the numbers. The third knocks with a feeling of fatigue — and that feeling, this time, is real data, not an illusion. More-tired-despite-using-it-more is a contradiction, and a contradiction is always where something needs a second look.
03A review once a quarter
Signals are only useful if you have a moment to stop and read them. Weekly is too dense — no approach should change every week. Yearly is too sparse — enough to grind away for a whole year. A quarter is the right cadence, and fifteen minutes is enough.
If it is different but the templates, the rules file, the way I delegate are unchanged — that is the gap. The approach should follow the work, not the work be bent to fit an old approach.
Avoiding a kind of work usually means the current approach makes it painful enough that you would rather do it by hand. That is exactly where a new approach would pay off the most.
This question cuts straight through sunk cost. If the answer is "no, I would do it differently now" — then that different thing is the work to do, regardless of how much you already invested in the old way.
The review does not force you to change anything — it only forces you to look. Most quarters, the answer will be "the old way is still fine," and you move on, relieved. Its value is in the one quarter where the answer is no — the quarter where, without this review, you would grind away three more months before admitting it.
04Changing the approach is not admitting failure — it is refusing to grind on
The real reason people cling to an old approach too long is rarely laziness. It is something more delicate: changing feels like admitting the old way was wrong, and the old way is the one you chose.
But an expired approach does not mean it was ever wrong. It was right for the work of its time. The work changed, you changed, the agent you use changed — an approach needing to change along with all that is the most ordinary thing in the world, not a confession.
What is actually worth dreading is not changing the approach. It is not changing: it is month nine still running the exact mold of month three, still a little more tired each week, still telling yourself it must be you. The next time you catch yourself thinking "using it more, but somehow more tired" — do not swallow it down. That is not you getting weaker. That is your approach asking to retire.