The Modern Sleep Problem Is Bigger Than Your Phone

Source: krishashok, "Guess what is destroying your sleep? | It’s not your phone" · Video ID: cu8DUB4GDXI · Published August 5, 2026

Poor sleep is not a single bad habit. It is the collision between a biological clock built for dark, cool, quiet nights and a world that made night brighter, warmer, louder, scheduled, monetized, and always reachable.

Who Is Krish Ashok?

Krish Ashok is a science communicator known for making everyday biology, food, health, and technology intelligible without flattening the science. His strength is not merely explaining mechanisms; it is connecting mitochondria, city heat, coffee, school start times, aircraft, and WhatsApp irritation into one coherent operating model for modern life.

Sleep Runs on an Hourglass and a Clock

The most useful frame is simple: sleep is governed by two interacting systems. One is an hourglass that measures how long you have been awake. The other is a clock that estimates what time it is on a spinning planet. Modern life is unusually good at making those two systems disagree.

The hourglass is built partly from adenosine. Cells use ATP—adenosine triphosphate—as rechargeable energy currency. The brain, though only about 2% of body weight, consumes roughly 20% of resting energy, continuously maintaining electrical gradients, sending chemical messages, and preparing for thoughts that may land at any moment. As ATP and related molecules are used and recycled, adenosine accumulates around wake-promoting circuits. Attention wanders. Eyelids gain mass. Sleep pressure rises.

Sleep reverses that hourglass. A nap empties some sand, which is why a two-hour nap at 6 p.m. can feel wonderful and still leave you staring at the ceiling at midnight: you paid part of the EMI before it was due.

The clock is the suprachiasmatic nucleus, or SCN, a small cluster of neurons near the crossing of the optic nerves. Its timing comes from genes and proteins cycling through feedback loops that take roughly 24 hours. On average, under carefully controlled conditions without sunlight or social cues, the human circadian period runs slightly longer than an Earth day—around 24 hours and 11 minutes. Light corrects that daily drift.

Blue Light Is the Least Interesting Part of the Light Story

The panic around blue light is based on a real mechanism, but often collapses it into a bad consumer-grade myth. The retina has rods, cones, and a third light-sensitive system: retinal ganglion cells containing melanopsin. These cells are less interested in images than in ambient light. They tell the SCN how bright the world is.

Melanopsin responds strongly to blue-rich wavelengths, but color is only one variable. Brightness, duration, distance, and timing matter too. A dim phone is not biologically equivalent to the morning sky merely because both contain blue light. Warm mode on a laptop or phone may help a little, but it does not solve the larger problem of bright light and stimulating content arriving at exactly the wrong biological time.

Morning light usually pulls the clock earlier. Evening light pushes it later. As evening light fades, the SCN signals the pineal gland to release melatonin. Melatonin is not a sleeping pill; it is closer to a chemical announcement that biological night has begun. Anxiety, caffeine, a brightly lit room, or the wrong circadian timing can easily override that announcement.

Caffeine Does Not Create Energy; It Hides Tiredness

Caffeine works because it resembles adenosine closely enough to occupy adenosine receptors without delivering adenosine’s message. The hourglass keeps filling, but the doorway is blocked. The tiredness signal is masked, not erased.

That distinction matters because caffeine lasts. Its half-life is commonly around five to six hours, with large variation from genetics, pregnancy, smoking, medicines, age, and liver function. A 100 mg caffeine dose at 4 p.m. can leave roughly 50 mg circulating at 9 p.m. Some people can drink coffee and still fall asleep, but falling asleep and sleeping well are not the same thing. Controlled studies have found measurable sleep disruption even when caffeine was taken six hours before bedtime.

The loop is easy to recognize: coffee delays or lightens sleep; the next day feels foggy; more coffee arrives, often later. Coffee can be useful and contains antioxidants, but if sleep is poor, moving it earlier or testing decaf for a couple of weeks is a cleaner experiment than changing six variables at once.

Sleep Quality Means Architecture, Not Just Hours

Sleep is not one continuous low-power mode. It alternates between non-REM and REM sleep in roughly 90-minute cycles. Non-REM begins with lighter stages and moves into deeper slow-wave sleep, where large groups of cortical neurons synchronize into alternating activity and silence. The brain has not shut down; it has synchronized.

Lighter non-REM includes sleep spindles, brief bursts shaped by interactions between the thalamus and cortex. These help reduce sensory flow into conscious awareness and participate in memory processing. The hippocampus acts like a desk where today’s files have landed; during non-REM sleep, recent patterns are replayed and stabilized. That is why an all-night study session is self-defeating: it buys extra exposure by disabling part of the machinery that stores what was just learned.

REM sleep is stranger. Brain activity becomes faster, vivid dreams become common, and most voluntary muscles are temporarily paralyzed so dreamed actions do not become physical actions. If waking consciousness returns before REM paralysis lifts, sleep paralysis can result: dream imagery leaks into the room as a presence, a figure, or pressure on the chest. Ancient cultures named these demons and spirits because “REM atonia persisting into wakefulness” was not available as an explanation.

Deep slow-wave sleep is concentrated earlier in the night; REM becomes more abundant toward morning. Losing the final two hours is therefore not an even slice of all sleep. It disproportionately removes REM-rich sleep.

There Was Never One Ancestral Sleep Protocol

The claim that humans naturally sleep in one uninterrupted eight-hour block is too tidy. Historian Roger Ekirch found hundreds of European references to “first sleep” and “second sleep,” with people waking calmly in the middle of the night to pray, talk, tend fires, have sex, visit neighbors, and then return to bed. In the 1990s, Thomas Wehr placed people in roughly 14 hours of darkness daily; after adjustment, several shifted into two sleep blocks with a quiet wakeful interval between them.

But the opposite simplification is also wrong. Hunter-gatherer communities such as the Hadza, San, and Tsimane do not reveal one universal pre-electric template. Many sleep six or seven hours, do not necessarily sleep at sunset, and do not all wake for long periods every night. Sleep varied with latitude, season, temperature, danger, childcare, status, and culture. There was no single ancestral sleep protocol, just as there was no single ancestral diet.

One useful throughline is temperature. People without electric lighting often fall asleep as the night cools, not immediately when darkness begins. Human core temperature usually falls before sleep; blood flow increases near the hands and feet so heat can leave the body. A warm shower can help for this reason: it brings blood toward the skin, after which heat escapes more easily.

Modernity Turned Night Into a Management Decision

Krish Ashok’s historical chain is the real argument. Joseph Swan demonstrated a working incandescent bulb in Newcastle in early 1879 before Edison out-patented and out-marketed him. Levi Hutchins built an American alarm clock in 1787 that rang only at 4 a.m. because that is when Hutchins wanted to wake up. Industrial Britain employed “knocker uppers” to tap workers’ windows with long poles. In 1806, a Manchester cotton mill was illuminated by coal gas, and the working day no longer had to be decided by a thermonuclear explosion 150 million kilometers away. It could be decided by management.

Then aircraft learned to outrun the sun. The Concorde could leave London, fly west at twice the speed of sound, and land in New York at a local time earlier than its departure. The phone then placed every argument, crisis, scare-mongering clip, and infinite scroll beside the bed. The phone is not innocent, but blaming it alone is like blaming the waiter because the dish was bad. It is merely the last person in a long chain of decisions.

Cities added heat and noise. Concrete and asphalt absorb heat during the day and release it after sunset. In a Chennai bedroom under a hot terrace, the body is trying to unload heat into a building radiating heat back. Noise need not fully wake someone to disturb sleep; the sleeping brain keeps monitoring sound because total sensory indifference would have been a terrible survival strategy. A horn or aircraft can cause a brief arousal, faster brain activity, and a rise in heart rate with no memory the next morning.

Teenagers, Shift Workers, and Jet Lag Expose the Same Bug

Puberty shifts the circadian clock later. Melatonin rises later, wakefulness persists into the evening, and teenagers still need roughly eight to ten hours of sleep. A teenager waking at 5:45 a.m. for school would need to be asleep by 8:45 p.m. to get nine hours—exactly when the adolescent clock may still be promoting wakefulness. Later school starts reliably increase sleep. Calling early starts “discipline” often disguises chronic sleep loss.

Night shift work pushes the conflict harder. Between roughly 2 and 6 a.m., core temperature is near its low point, melatonin is high, attention is unstable, and sleep pressure may be heavy. Then the worker goes home and tries to sleep while the clock promotes wakefulness, sunlight enters the room, the neighborhood warms, and pressure cooker whistles announce lunch. Rotating shifts are worse because the body starts adjusting just as the roster changes again.

Jet lag is the same disagreement with a passport. The phone updates time zones instantly; the SCN does not. The master clock responds to light, while peripheral clocks in the liver, pancreas, gut, muscle, and fat also respond to meals and activity. Midnight biryani sends digestive organs a daytime signal while the master clock still says night. Different organs adjust at different speeds.

“I Have a Sleep Problem” Is Too Vague to Fix

“I have a sleep problem” is like saying a vehicle is making a noise. The battery may be dead, the tire flat, or the engine broken. Sleep complaints often collapse distinct failure modes into one label.

These problems can overlap. Someone with apnea may use caffeine to survive the day and then struggle to sleep. A shift worker may develop insomnia around daytime sleep. The same sentence—“I’m always tired”—can have entirely different causes.

What Sleep Loss Actually Breaks

Sleep deprivation does not make the brain uniformly slower. It makes performance unstable. A person may respond normally several times and then suddenly miss a signal. That instability is especially dangerous in long quiet tasks with one critical event: highway driving, patient monitoring, factory supervision at night.

People are also poor judges of their own decline. After several short nights, tiredness begins to feel normal while objective performance keeps worsening. Emotional regulation weakens: a mildly irritating WhatsApp message at midnight can provoke anger that would feel harmless in the morning. Sleep restriction can reduce insulin sensitivity, alter stress and appetite systems, make calorie-dense food more attractive, and create more hours in which eating is possible.

After roughly 17 to 19 hours awake, performance on some tests can resemble a blood alcohol concentration of 0.05%. Sleep deprivation and alcohol are not identical, but at 80 km/h, a three-second lapse moves a car about 67 meters. The danger is real even when the law treats it differently.

Key Lessons

Why This Matters for Diffie

For Anand and Diffie, the useful analogy is not “sleep better so you can work harder.” It is systems diagnosis. Krish Ashok’s strongest point is that a visible culprit—the phone—often distracts from the larger architecture: light, heat, schedules, caffeine, noise, anxiety, physiology, and institutions all pushing the system out of sync.

That maps directly to frontend testing. A flaky browser test is rarely just “the selector broke.” It may be timing, network jitter, hydration, browser state, async rendering, animation, third-party scripts, CI load, or a product change that made an old assertion obsolete. The high-value product move for Diffie is to make the hidden systems legible: not merely report that a test failed, but identify which clocks are disagreeing.

The article suggests three concrete product and GTM moves:

• Lead with root-cause framing. Position Diffie as the tool that distinguishes “insufficient sleep opportunity” from “insomnia” from “apnea” for frontend QA: different failure modes require different fixes.
• Design for clean experiments. Krish’s advice—do not change six things at once—is exactly the workflow engineers need when stabilizing tests. Diffie can emphasize controlled replay, isolated environment changes, and evidence trails.
• Sell operational alignment, not just bug detection. Modern sleep fails when the human clock, work clock, city clock, and device clock conflict. Frontend quality fails when design, product, engineering, CI, and browser reality disagree. Diffie’s ICP-building should target teams already feeling that cross-functional mismatch.

The practical goal is not perfect sleep, and the practical goal for frontend engineering is not perfect tests. It is to give the system clearer signals, enough time to stabilize, and better diagnostics when the machinery still fails.