He Tried to Delete Sleep. Eleven Days Later, His Body Sent the Bill.

It starts, like most extreme experiments do, with a spreadsheet.

A remote software contractor runs the numbers on his own life and doesn't like what he sees. Eight hours a night is roughly 2,920 hours a year — about 122 full days spent unconscious. He has a side project he never touches. So he decides to buy the time back.

Read nextPetri Dish or Profit Engine? Inside the Wild Business of Gym Cold Plunges and Extreme Fasting Resets

The plan is called Uberman: six 20-minute naps, spaced every four hours, and nothing else. Two hours of sleep a day. Twenty-two hours awake. On paper, it hands him 42 extra hours a week.

He sets six alarms — 2, 6, 10, 14, 18 and 22 — buys blackout curtains and a sunrise lamp, quits caffeine cold to keep his adenosine signaling "clean," and starts logging everything publicly.

The first three days are pure misery, and he expects that. He cannot actually fall asleep inside the 20-minute windows. His wearable records about 40 minutes of real sleep across the whole of day two. "Gravity is increasing," he writes.

Then things get strange in a way he reads as progress.

By day four he's unconscious within 90 seconds of lying down. He's dreaming — full, narrative, plot-having dreams — inside naps that last twenty minutes. He takes this as proof the protocol is working, that his brain is learning to skip straight to the good stuff.

His output tells a different story. His own productivity scores fall from 8 out of 10 to 4. He notes reading the same paragraph four times before it registers.

Day seven is the turn. At 3 a.m. he gets a surge of clarity, writes 1,200 lines of code, and announces to the community that adaptation is complete. He has, he says, broken sleep.

He goes back to that code later. A large chunk of it is unusable.

Days eight through ten unravel fast. He sees a cat that isn't in the room. Text slides sideways on his monitor. He sleeps through two alarms and wakes with no idea which nap he's in. His resting heart rate climbs about nine beats above baseline. He's cold all the time and eating an extra 700 calories a day, almost all of it sugar.

On day eleven he falls asleep at a stoplight. He drives home, sleeps for eleven hours, and posts that it's over.

The comments split immediately, and the split is the interesting part.

"You quit right before the adaptation locked in," one commenter said, insisting week three is where the schedule stabilizes. Another argued the opposite — that everything in the log is a known deprivation curve with a biohacking label stapled on.

The science leans hard toward the second camp, and the reason is unsettling.

In a landmark two-week laboratory study on sleep restriction, people limited to six hours in bed kept getting measurably worse on attention testing, day after day, ending up comparable to people who'd been awake for two straight nights. But their own ratings of how sleepy they felt flattened out early. They stopped feeling worse while continuing to perform worse.

Your brand, right here.Reach story-obsessed listeners in 45+ languages → advertise on Avonetics

That is day seven. The euphoria isn't adaptation. It's the gauge breaking.

The dream detail is worse than it sounds too. Deep slow-wave sleep is front-loaded into the early hours of a normal night; REM is back-loaded into the final third. A 20-minute nap structurally shouldn't reach REM at all. When it does, sleep physicians have a name for it — a sleep-onset REM period — and they see it in narcolepsy and in severe deprivation.

So the exact symptom the community celebrates as mastery is, in a clinic, a flag.

"You didn't teach your brain efficiency, you built up so much REM pressure it started cutting the line," one commenter said.

There's a real steelman here, and it deserves airtime. Sleeping in one unbroken block isn't sacred history. Historians have documented hundreds of pre-industrial references to a "first sleep" and a "second sleep" with a calm wakeful gap between them, and lab subjects placed in long dark periods have drifted into exactly that bimodal pattern on their own.

Naps work, too. A well-known aviation study found a planned 26-minute nap improved performance by about 34% and alertness by roughly 54%.

And a tiny number of people genuinely are built different — mutations in genes like DEC2 and ADRB1 produce natural short sleepers who run fine on four to six hours. Those families are real. The catch is that the trait is inherited, not earned. You discover it; you don't unlock it.

Which leaves the honest version of the finding: segmented sleep has support, strategic naps have support, and six-nap schedules have testimonials.

There's one number that ends most of the argument. Roughly 17 to 19 hours awake degrades driving performance about as much as a blood alcohol level of 0.05%. Push to 24 hours and you're near 0.10%.

A 22-hour waking day puts you in that territory every single day — not as a failure of the protocol, but as the protocol working exactly as designed.

He reclaimed 42 hours a week. He just spent them at roughly half capacity, and paid the difference at a stoplight.

The hosts take this one apart on the podcast — the day-seven euphoria trap, the REM rebound argument, and which parts of polyphasic sleep are actually worth stealing.