There were years, and I’ll spare you the exact count, when I treated my bed like an opponent. I climbed in expecting a fight and I usually lost. I’d lie there running the day’s tape backward, negotiating with the ceiling, doing arithmetic on how many hours were left if I fell asleep right now. Then right now. Then right now. Morning arrived like a bill I couldn’t pay.
Now that I’ve spent a couple of years learning to footnote my hunches instead of just broadcasting them, I want to tell you what actually changed, and to do it the way I was trained to: with the receipts attached. Because the honest version of my story is not the one I would have told you back then. Back then I thought I had a sleep problem. I didn’t. I had a living problem that only sent its invoice at night.
The expensive way to miss the point
Here’s the trap, and it’s a very modern one. When something breaks at night, we shop for a night-time solution. I bought the mattress. I bought the pillow with opinions. I bought supplements with confident labels and tried the apps that promised to hack my sleep, as though sleep were a lock and I’d simply lost the combination.
Rory Sutherland has a line I’ve come to live by: the reframe is usually cheaper than the fix, and usually better. My mistake wasn’t that I chose the wrong gadget. It’s that I was solving the wrong sentence. I kept asking how do I fall asleep when the useful question was what am I bringing to bed with me.
The two changes that finally mattered cost almost nothing. One of them was free. It was the light switch, turned off.
What I changed in the room, and why it works
I did eventually fix the room, and the room turned out to matter more than any device, because it speaks the one language your body has never stopped understanding.
I stopped lying to my brain about what time it is. Deep in the hypothalamus sits a cluster of cells called the suprachiasmatic nucleus, your master clock, and it sets itself almost entirely by light landing on your eyes. Evening light, especially the blue-white kind pouring off screens and LEDs, tells that clock the sun is still up, and the clock responds by holding back melatonin, the hormone that ushers in sleep. A 2015 study from Charles Czeisler’s group at Harvard put this in plain terms: volunteers who read on a glowing e-reader before bed took longer to fall asleep, showed a delayed body clock, and were groggier the next morning than those who read a paper book. A systematic review the same field produced found that even a couple of hours of evening light suppresses melatonin, and that the suppression is worst at exactly the short, blue wavelengths our devices love.
So I blacked out the bedroom. Not “dimmed.” Blacked out. I hunted down every little status LED, the ones on chargers and smoke detectors and the television that glow like tiny interrogations, and I killed them or taped over them. Then I did the harder half, which is the other end of the clock: I started getting real light into my eyes in the morning, early, because morning light does the opposite of evening light. It nudges the clock earlier, so the whole rhythm arrives on time that night. If evening light is a lie your body believes, morning light is the correction.
I made the room cold. This one surprised me until I understood the mechanism. Sleep doesn’t begin until your core temperature falls, by roughly one to three degrees Fahrenheit, and the temperature of the air around you is the biggest lever on how cleanly that drop happens. A long-standing review in Sleep Medicine Reviews landed on a range most people find absurd at first: somewhere around 60 to 67°F. I set the thermostat down at night and stopped apologizing for it. (There’s a lovely counterintuitive twist here, too: a warm bath before bed helps cool you. The heat pulls blood to your skin, and you dump warmth into the room once you’re out. The alchemist in me appreciates a hot bath that works by making you colder.)
A note I’ve earned the right to make now that I’m the age I am: the clock changes as we age. Older bodies make less melatonin and less of the deep, slow-wave sleep that does the heavy lifting, and they often want the room a touch warmer than a twenty-five-year-old would. Recognizing your own rhythm, not a guru’s, is the whole game.
I gave the night a cue, not a cage. I built a wind-down: lights down, the same slow, wordless music, the same short sequence every night. It helped enormously. But I’ll say the quiet part out loud, because this is where routines go wrong: a ritual is a beautiful thing right up until you stop noticing you’re performing it. The music isn’t magic. It’s a signal to a body that runs on signals. The day I felt anxious because I’d skipped the playlist was the day I realized I’d turned a helpful cue into one more thing to get right. Hold your routine the way you’d hold a bird: firmly enough that it stays, loosely enough that it can breathe.
The night keeps the day’s receipts
And still, with the cold, dark, quiet, well-cued room, some nights didn’t take. That’s when I found the thing I actually came here to tell you.
The night is downstream of the day.
Start with the most embarrassing example. I’m a coffee person, and I’d have my last cup mid-afternoon, sure it was harmless. It wasn’t. In a tidy little 2013 study in the Journal of Clinical Sleep Medicine, researchers gave people caffeine at bedtime, three hours before, and six hours before. Even the six-hours-before dose stole more than an hour of sleep, and here’s the part that rearranged my head: the sleepers couldn’t feel it. Caffeine has a half-life of five to six hours; a quarter of that 3 p.m. cup is still circulating near midnight, quietly sanding down your deep sleep while you swear the coffee “doesn’t affect you.” Your afternoon was writing checks your night had to cover, invisibly.
There’s a companion to that, and it’s the one I get asked about most, so let me report it straight. A lot of people now reach for cannabis, THC, to get to sleep, and the first half of the folk wisdom holds up: THC does tend to shorten the time it takes to drift off, and can even deepen slow-wave sleep for a night. But the research keeps returning the same catch, and it’s the mirror image of the caffeine problem. With coffee, you can’t feel the harm. With THC, you can’t feel the cost.
The cost is paid mostly in REM, the dreaming sleep where the mind does its emotional filing, and THC suppresses it. A 2026 randomized trial in the Journal of Sleep Research found a nightly cannabinoid dose cut REM by roughly half an hour and shoved it later into the night. Use it regularly and two more things happen. You build tolerance, so the sedative magic fades. And when you stop, REM returns with interest, which is the part that quietly traps people: a 2025 systematic review of cannabis and sleep found that quitting reliably brings a stretch of worse sleep, insomnia and vivid, sometimes lurid dreams, as the brain floods back the REM it was denied. That rebound is unpleasant enough that it drives a lot of people straight back to the thing they were trying to leave.
I’m not here to tell you what to put in your body; you’re an adult with your own judgment, and the science here is genuinely still arguing with itself about dose and formulation. I’ll only point at the shape of it, because it’s the shape of this whole essay: a thing that seems to give you sleep may only be lending it, at an interest rate you don’t notice until you try to close the account. (The clinicians’ quiet consensus, for what it’s worth, is to exhaust the un-sexy, well-evidenced tools first, the light, the cold, the cognitive-behavioral ones, before the pharmacology.)
Then the bigger one. The racing mind I mentioned at the top, the ceiling negotiations, turns out to be the most measurable saboteur of all. A 2020 study led by David Kalmbach found that nocturnal cognitive arousal, the technical name for lying there ruminating and worrying, predicted objectively disturbed sleep more reliably than an actual insomnia diagnosis or depression did. Earlier work by Allison Harvey and Colin Espie noticed something even more useful: poor sleepers, asked what fills their heads at lights-out, describe “rehearsal and planning,” what happened today and what I’ve got tomorrow. Good sleepers, asked the same thing, mostly report thinking about nothing in particular.
Sit with that. The difference between a wrecked night and a clean one is often not a chemical or a mattress. It’s whether the day left something unfinished, unsaid, unforgiven: something the mind insists on litigating the moment you go still.
Which means the most powerful sleep intervention I know isn’t in the bedroom at all. It’s living the waking hours so that fewer of them come to collect at midnight. Make the apology in the afternoon. Close the loop before dinner. Say the true thing while the sun’s up. I didn’t fix my sleep by managing my thoughts at 2 a.m., which is like trying to fold a map in a windstorm. I fixed it upstream, by leaving fewer regrets lying around for my brain to trip over in the dark.
There’s real evidence you can move the whole system this way. When researchers took confirmed night owls and changed not their bedtime but their days (earlier light, fixed mealtimes, caffeine cutoffs, timed exercise), the group shifted their internal clock nearly two hours earlier in three weeks. The bonus was the tell: they also reported less depression and stress, and tested better on reaction time and even grip strength. Live the day on purpose, and the night reorganizes itself around it.
The symbiosis
Here’s the frame that finally made it all cohere for me, and it comes from the most vivid image in the whole science.
A decade ago, Maiken Nedergaard’s lab at Rochester found that the sleeping brain does something the waking brain essentially can’t: it opens a plumbing system (she named it the glymphatic system) and flushes out the metabolic junk that piles up all day, including the amyloid proteins tied to Alzheimer’s. It runs up to ten times harder in sleep than in waking. She compares it to running the dishwasher overnight and waking to a clean kitchen. Her deeper point is the one I keep returning to: the brain has two states, and it more or less can’t do both at once, awake and gathering the world, or asleep and clearing it.
That’s not a switch. That’s a partnership. The day fills the sink; the night empties it. A day lived well (moved through, resolved, paid up) is what gives the night something clean to restore you toward. And a night that actually restores you is what hands you back a day clear enough to live well. Less regret in the day means less to ruminate on at night; deeper sleep at night means a sharper, kinder day; a sharper, kinder day means fewer regrets to carry back to bed. The loop runs in whichever direction you feed it.
I spent years trying to win the night as if it were separate territory. It never was. It was always the second half of the same day.
Marcus Aurelius, filling a private notebook by lamplight nineteen centuries ago, kept circling one idea: that a good life is a life lived in agreement with itself, the governing part of us not at war with the rest, the person the same on the inside as the face he shows the day. He wasn’t thinking about thermostats and melatonin. But he was describing exactly what I felt when my days and nights finally stopped fighting each other. The cold dark room was never really the point. The point was a kind of inner harmony that a divided life can’t fake and a well-lived day doesn’t have to. You sleep like someone who isn’t at war with himself. Which means that on some nights the most honest sleep aid is simply this: to arrive at evening a little more at peace with how you spent the light.
Hold it loosely
One more thing, and I say it precisely because I now have the training to know better than to skip it. Be careful with sleep experts, including the famous ones. The best-selling sleep book of our era, Why We Sleep, was later shown by the independent researcher Alexey Guzey to contain real errors, enough that even a university review conceded corrections were owed. I’m not telling you to distrust the science. I’m telling you the opposite of what a guru would: I could be wrong about pieces of this, the field is still arguing with itself, and you should treat everything above as an instrument to test, not a rule to obey.
So I’m not handing you a protocol. I’m handing you a lamp. Black out your room for a week and watch. Move the coffee earlier and see if the night notices what you can’t feel. And on the evening a good sleep won’t come, resist the urge to reach for another gadget, and ask, gently, what the day is trying to tell you.
Your night has been keeping honest records this whole time. It might be the most truthful thing you own.
Sources
• Chang, A.-M., Aeschbach, D., Duffy, J. F., & Czeisler, C. A. (2015). Evening use of light-emitting eReaders negatively affects sleep, circadian timing, and next-morning alertness. PNAS. https://www.pnas.org/doi/10.1073/pnas.1418490112
• Tähkämö, L., Partonen, T., & Pesonen, A.-K. (2019). Systematic review of light exposure impact on human circadian rhythm. Chronobiology International. https://www.tandfonline.com/doi/full/10.1080/07420528.2018.1527773
• Okamoto-Mizuno, K., & Mizuno, K. (2012). Effects of thermal environment on sleep and circadian rhythm. Sleep Medicine Reviews (as reviewed). https://www.sleepfoundation.org/bedroom-environment/best-temperature-for-sleep
• Drake, C., Roehrs, T., Shambroom, J., & Roth, T. (2013). Caffeine effects on sleep taken 0, 3, or 6 hours before going to bed. Journal of Clinical Sleep Medicine. https://sleepeducation.org/?p=708
• Kalmbach, D. A., Buysse, D. J., Cheng, P., Roth, T., Yang, A., & Drake, C. L. (2020). Nocturnal cognitive arousal is associated with objective sleep disturbance and indicators of physiologic hyperarousal in good sleepers and individuals with insomnia disorder. Sleep Medicine. https://pmc.ncbi.nlm.nih.gov/articles/PMC8212183/
• Suraev, A., et al. (2026). Acute effects of oral cannabinoids on sleep and high-density EEG in insomnia: a pilot randomised controlled trial. Journal of Sleep Research. https://onlinelibrary.wiley.com/doi/10.1111/jsr.70124
• Cannabis and sleep architecture: a systematic review and meta-analysis (2025). Sleep Medicine Reviews. https://www.sciencedirect.com/science/article/pii/S1087079225001170
• Facer-Childs, E. R., et al. (2019). Resetting the late timing of “night owls” has a positive impact on mental health and performance. Sleep Medicine. https://www.sciencedirect.com/science/article/abs/pii/S1389945714004936
• Xie, L., et al. / Nedergaard, M. (2013). Sleep drives metabolite clearance from the adult brain: the glymphatic system. Science / University of Rochester Medical Center. https://www.urmc.rochester.edu/news/story/to-sleep-perchance-to-clean
• Guzey, A. (2019). Matthew Walker’s “Why We Sleep” Is Riddled with Scientific and Factual Errors. https://guzey.com/books/why-we-sleep/
• Marcus Aurelius. Meditations (2nd century AD; public domain): on living in agreement with oneself and the retreat into one’s own soul, esp. Books 4 and 6.