This is an automated rejection. No LLM generated, heavily assisted/co-written, or otherwise reliant work.
Read full explanation
I want to propose a failure mode that I suspect affects a subset of high-agency people.
The short version:
After an interruption, some people systematically rebuild their execution systems instead of simply resuming them.
I don’t mean quitting. I mean reinitializing. New rules, new structure, new “fresh start.”
My tentative claim is that this isn’t primarily about laziness or classic akrasia. It’s about identity-level instability under variance.
This seems relevant to LessWrong because it touches on self-model stability, robustness to noise, and how agents behave when reality contradicts their internal narrative.
The Pattern
Roughly:
You construct a structured plan.
Execution goes well for several days.
A small interruption occurs (missed day, stress spike, fatigue).
Instead of resuming, you feel a strong urge to redesign the system.
You restart.
Empirically (at least anecdotally), the goals don’t change much. The constraints barely change. But the restart feels psychologically necessary.
That’s the part I’m trying to explain.
A Possible Mechanism
Here’s the current model. I’m not highly confident in it.
1. Planning encodes an identity claim. When you build a system, you’re implicitly asserting: “I operate at this standard now.”
2. A miss creates identity dissonance. Even a small failure becomes evidence that the new identity isn’t stable.
3. Restarting restores narrative coherence. Instead of tolerating “I operate at this standard imperfectly,” you reset to “Now I operate at this standard.”
Restarting is cheaper (psychologically) than integrating the contradiction.
There’s also an activation-cost component. After a miss, the system feels heavier — more backlog, more self-judgment, more friction. Restarting reduces perceived complexity even if objective work stays constant.
So restart may be locally rational (reduce immediate discomfort) but globally suboptimal (destroy compounding).
Distinction From Nearby Concepts
This overlaps with:
Perfectionism
Akrasia
Status-quo bias
Narrative self-protection
But I think it’s slightly distinct.
Classic akrasia predicts procrastination. This predicts strong initial execution followed by system reinitialization after variance.
Perfectionism predicts dissatisfaction. This predicts redesign frequency specifically triggered by interruption.
If this is just a special case of those, I’d like to understand how.
A Falsifiable Prediction
If this model is roughly correct, then:
Return latency (time between a miss and re-entry) should predict long-term output better than streak length.
Prohibiting catch-up behavior should increase total output over longer horizons.
Forcing a small, time-boxed re-entry should reduce restart frequency.
I don’t have data yet. This is a pattern-level observation, not a study.
Why This Might Matter
If true, then some “low agency” outcomes are actually high standards interacting poorly with variance.
In that case, improving agency isn’t about increasing motivation.
It’s about designing systems that survive contradiction.
In other words: robustness > intensity.
I’m unsure how novel this is. It may just be a reframing of existing ideas in The Sequences about self-deception and local vs global optimization.
If anyone has seen related work, I’d appreciate pointers.
I want to propose a failure mode that I suspect affects a subset of high-agency people.
The short version:
I don’t mean quitting. I mean reinitializing. New rules, new structure, new “fresh start.”
My tentative claim is that this isn’t primarily about laziness or classic akrasia. It’s about identity-level instability under variance.
This seems relevant to LessWrong because it touches on self-model stability, robustness to noise, and how agents behave when reality contradicts their internal narrative.
The Pattern
Roughly:
Empirically (at least anecdotally), the goals don’t change much. The constraints barely change. But the restart feels psychologically necessary.
That’s the part I’m trying to explain.
A Possible Mechanism
Here’s the current model. I’m not highly confident in it.
1. Planning encodes an identity claim.
When you build a system, you’re implicitly asserting: “I operate at this standard now.”
2. A miss creates identity dissonance.
Even a small failure becomes evidence that the new identity isn’t stable.
3. Restarting restores narrative coherence.
Instead of tolerating “I operate at this standard imperfectly,” you reset to “Now I operate at this standard.”
Restarting is cheaper (psychologically) than integrating the contradiction.
There’s also an activation-cost component. After a miss, the system feels heavier — more backlog, more self-judgment, more friction. Restarting reduces perceived complexity even if objective work stays constant.
So restart may be locally rational (reduce immediate discomfort) but globally suboptimal (destroy compounding).
Distinction From Nearby Concepts
This overlaps with:
But I think it’s slightly distinct.
Classic akrasia predicts procrastination.
This predicts strong initial execution followed by system reinitialization after variance.
Perfectionism predicts dissatisfaction.
This predicts redesign frequency specifically triggered by interruption.
If this is just a special case of those, I’d like to understand how.
A Falsifiable Prediction
If this model is roughly correct, then:
I don’t have data yet. This is a pattern-level observation, not a study.
Why This Might Matter
If true, then some “low agency” outcomes are actually high standards interacting poorly with variance.
In that case, improving agency isn’t about increasing motivation.
It’s about designing systems that survive contradiction.
In other words: robustness > intensity.
I’m unsure how novel this is. It may just be a reframing of existing ideas in The Sequences about self-deception and local vs global optimization.
If anyone has seen related work, I’d appreciate pointers.