Set hard time-based triggers for sleep agents — fatigued brains systematically misjudge their own tiredness
For sleep agents, set a hard time-based trigger (like 9:30 PM) rather than a subjective state trigger (like 'when I feel tired'), because fatigued brains systematically misjudge their own tiredness.
Why This Is a Rule
Sleep is the domain where subjective state triggers fail most spectacularly, because the very state that should trigger the agent (fatigue) impairs the cognitive system that detects the state. Research on sleep deprivation consistently shows that subjective sleepiness plateaus while objective impairment continues to worsen — people stop feeling increasingly tired long before their performance stops declining. "I'll go to bed when I feel tired" is a trigger that degrades as a function of the very thing it's supposed to detect.
A time-based trigger — "At 9:30 PM, begin the shutdown routine" — fires regardless of your subjective assessment of tiredness. It doesn't ask whether you feel ready for bed. It fires because the clock hit 9:30, which is a fact that fatigue cannot distort. This is Agent triggers must be observable or measurable — vague triggers like "when I feel ready" never fire reliably (observable triggers over subjective ones) applied to the domain where the gap between subjective assessment and objective reality is widest.
The "but I'm not tired yet" objection is itself evidence of the problem. At 9:30 PM, after a full day of cognitive work, you are almost certainly tired enough that sleep quality would benefit from starting the wind-down process. The fact that you don't feel tired is the fatigue talking.
When This Fires
- When designing or redesigning your sleep routine agent
- When you consistently stay up later than intended because you "didn't feel tired yet"
- When screen time reports show evening phone use past your target bedtime
- When applying Agent triggers must be observable or measurable — vague triggers like "when I feel ready" never fire reliably (observable triggers) to health-related behavioral agents
Common Failure Mode
Negotiating with the time-based trigger: "It's 9:30 but I'll just finish this episode/chapter/email." The trigger fired successfully — the agent's job is done. The failure is in the action execution. Design the action to minimize negotiation: the 9:30 alarm means "phone goes in the other room, not 'consider whether to start winding down.'"
The Protocol
(1) Set a hard time-based trigger for your sleep shutdown routine. Choose a time that allows 30-60 minutes of wind-down before target sleep time. (2) The trigger must be impossible to miss: phone alarm, smart home automation, or partner agreement. (3) The action should be physical and non-negotiable: phone in another room, lights dimmed, screens off. Not "think about going to bed" but "do these physical actions." (4) Do not override the trigger based on subjective tiredness assessment. The whole point is that your assessment at this hour is unreliable. (5) Track compliance and sleep quality. Most people find that consistent time-based triggers produce better sleep within one week, even though they "didn't feel tired" at trigger time.