It was 3:47 a.m. when I found myself scrolling through Twitter again, phone brightness turned down low, lying on my side in bed. My wife was asleep next to me. I'd been awake since 2:15. This wasn't unusual. This was Tuesday.
I couldn't remember the last time I'd slept through the night. Not an exaggeration — I genuinely could not recall a single unbroken night of sleep in the past four years. I'd wake up around 2 or 3 a.m., completely alert, mind racing about work emails or something I'd said three days ago or absolutely nothing at all. Then I'd lie there, getting progressively more frustrated, reaching for my phone because at least scrolling felt like doing something.
The Ambien had stopped working months ago. Well, it worked, but only if I kept increasing the dose, and my doctor wasn't comfortable going higher. I didn't blame her. I could feel myself depending on it in a way that scared me. On nights I didn't take it, I'd lie awake convinced I'd never fall asleep, my heart rate climbing just from the anxiety of not having taken the pill.
When Exhaustion Becomes Normal
By year three, I'd stopped talking about it. Friends would ask how I slept, and I'd say "fine" because what else was there to say? Everyone's tired. Everyone complains about sleep. I wasn't special. I just felt like I was moving through life at 70% capacity, all the time, forever.
I'd tried everything I could Google. Melatonin — didn't touch it. Magnesium supplements. A weighted blanket that made me sweat. I stopped caffeine after 2 p.m., then after noon, then cut it entirely for two months. No difference. I bought blackout curtains, a white noise machine, a sleep tracker that told me what I already knew: my sleep was terrible.
The worst part wasn't the tiredness. It was the feeling that my brain had betrayed me. Sleep is supposed to be automatic. Animals do it. Babies do it. But somewhere along the line, I'd forgotten how.
The Protocol That Sounded Crazy
I came across CBT-I — Cognitive Behavioral Therapy for Insomnia — in an article that mentioned research from Dr. Michael Perlis at the University of Pennsylvania. The success rates were remarkable, better than medication for long-term outcomes. But when I read what it actually involved, I almost closed the tab.
Sleep restriction. Stimulus control. The idea was to completely reset my relationship with bed and sleep by initially limiting my time in bed to match my actual sleep time — even if that meant only allowing myself five or six hours in bed total. It sounded like torture. I was already exhausted. Now I was supposed to sleep less?
But I was desperate enough to try. I found a structured guide and committed to six weeks.
Week One: The Brutal Part
I tracked my sleep for a week first — when I got into bed, when I actually fell asleep, when I woke up, when I got out of bed. Average actual sleep time: 5.5 hours, spread across 8.5 hours in bed. That meant I was spending three hours per night just lying there, not sleeping, training my brain to associate bed with wakefulness.
The protocol said to restrict my sleep window to match my actual sleep time, plus 30 minutes. So I set my window: midnight to 5:30 a.m. No getting into bed before midnight, no matter how tired I felt. Alarm at 5:30 a.m., no matter what.
The first night, I sat on the couch until midnight, exhausted and angry, watching the clock. I fell asleep within ten minutes of getting into bed. Then woke at 2:45 a.m. Here we go again. But this time, instead of lying there scrolling, the protocol said to get up. Leave the bedroom. Come back only when sleepy.
I read on the couch for forty minutes, eyes burning. Went back to bed. Slept until the 5:30 alarm. Total sleep: maybe four hours. I felt like death.
The Shift I Didn't Expect
By day four of week one, something strange happened. I got into bed at midnight and fell asleep almost immediately. Woke once around 3:30, got up for fifteen minutes, went back to bed, slept until the alarm. It wasn't perfect, but it was different. My body was so tired that it stopped fighting.
The protocol tracks something called sleep efficiency — the percentage of time in bed that you're actually asleep. When I started, mine was around 65%. By the end of week one, it was 83%. Once you hit 85% or higher for three consecutive nights, you're allowed to expand your sleep window by fifteen minutes.
Week two: 11:45 p.m. to 5:30 a.m. I protected that window like it was sacred. If I woke up at 3 a.m., I got out of bed. Every time. Bed was for sleep only, not for scrolling or worrying or lying awake. I was retraining my brain to expect that bed meant sleep.
Week Six: The Breakthrough
By week four, my sleep window had expanded to 11:00 p.m. to 6:00 a.m. — seven hours. I was sleeping about 6.5 of those hours, and the middle-of-the-night wakings were getting shorter. Ten minutes instead of an hour. Sometimes I'd wake up and fall back asleep without even getting out of bed.
Week six, I slept through the night for the first time in four years. 11 p.m. to 6 a.m., straight through. I woke up and actually didn't believe it. I checked my watch to make sure it was really 6 a.m., that I hadn't just fallen asleep.
I didn't cry, but it was close. I didn't realize how much of my identity had become "person who doesn't sleep" until I suddenly wasn't that person anymore.
What Actually Changed
It wasn't magic. It was consistency and structure applied to something I'd been approaching with desperation and chaos. The sleep restriction built up sleep pressure — my body's actual need for sleep — to the point where it overrode the anxiety and the racing thoughts. The stimulus control retrained my brain to associate bed with sleep instead of wakefulness.
I also worked on the cognitive part — noticing the catastrophic thoughts ("I'll never sleep again," "This will ruin my whole day") and testing them against reality. Most of the time, I functioned fine even on bad sleep. The anxiety about the sleep was often worse than the actual sleep deprivation.
If you're dealing with chronic insomnia and you're tired of bandaid solutions, I put together everything I learned into Rest →, a self-administered CBT-I protocol. It's the same structure I followed, with sleep logs, week-by-week guidelines, and the specific techniques that reset my sleep after years of struggling.
I still have bad nights occasionally. Last week I woke up at 4 a.m. and couldn't fall back asleep. But it was one night, not a pattern. And I knew what to do — get up, leave the bedroom, come back when sleepy. I didn't spiral. I didn't catastrophize. I just followed the protocol, and the next night I slept fine.
That's the real difference. I'm not afraid of my own bed anymore.
— Simon