Why willpower isn't the problem

TL;DR. Modern packaged food is engineered to bypass the hunger and fullness signals that regulate eating. The strongest evidence is Kevin Hall's 2019 NIH metabolic-ward trial: people eating ultra-processed food ate about 500 extra calories a day versus a whole-food diet matched on sugar, fat, salt, and fiber. That changes what works. Trying harder is not the lever. Changing the food around you is.

What you'll learn

• Explain, in one sentence, what the Hall 2019 trial actually proved about ultra-processed food.
• Name three concrete ways modern food is engineered to override appetite.
• Tell the difference between "wanting" and "liking" — and why that distinction matters for self-blame.
• Spot the conditioned-hypereating pattern (loss of control, no satisfaction, preoccupation) without diagnosing yourself.
• Describe two specific changes to your food environment that work better than willpower.

Why this matters

The average US supermarket carried about 6,000 items in 1980. Today it carries around 33,000. A can of Coca-Cola delivers its sugar payload to the brain in roughly 600 milliseconds — about twenty times faster than the nicotine in a cigarette. By the five-marker definition (waist, blood pressure, fasting glucose, triglycerides, HDL), 93.2 percent of US adults are metabolically unhealthy. That is not a generation that forgot how to be disciplined.

Something else changed. The food did. Roughly sixty percent of US calories now come from ultra-processed products that didn't exist in their current form a hundred years ago — extruded, reconstituted, flavor-paired, brand-able formulations built to be eaten in volume. The people designing them are smart, well-funded, and explicitly optimizing for "craveability" and "share of stomach."

If you have ever felt like your appetite has a mind of its own — like the bag of chips empties faster than you can decide whether you wanted it — you are not weak. You are downstream of an industry that has spent decades figuring out how to make that exact thing happen. This isn't a willpower problem. It's a food problem, and that means there are different things to do about it.

What did the Hall 2019 trial actually show?

The standard rebuttal to "ultra-processed food makes people overeat" used to be that it just has more sugar, fat, and salt — so of course people eat more. Kevin Hall, an NIH physicist and outspoken skeptic of the ultra-processed framing, set out to test that directly.

Hall ran 20 adults through the NIH Clinical Center's metabolic ward for four weeks: two two-week phases, one ultra-processed, one minimally processed, menus matched for calories offered, sugar, fat, protein, fiber, and sodium. Food was ad libitum.

The results, published in Cell Metabolism in 2019, were unambiguous. On the ultra-processed diet, participants ate about 500 extra calories per day and gained roughly a pound a week. On the matched whole-food diet, the same people lost weight. Same humans, same nutrient profile on paper, two different outcomes. Processing itself drove the gap — even when the nutrition label barely moved.

How modern food is engineered

Three mechanisms do most of the work. Each has a name and a paper trail.

1. The bliss point. In the 1980s, Howard Moskowitz formalized a technique from his time at the US Army's Natick labs: run hundreds of recipe variations past taste panels and find the precise concentration of sugar (or salt, or fat) that maximizes liking without triggering the brain's "enough" signal. Moskowitz used it to rescue Maxwell House, redesign Prego around a sugar load most people would never put in homemade marinara, and develop Cherry Vanilla Dr Pepper through 61 formulas and 3,904 tastings. The bliss point is a measured number on a curve, and almost every major shelf-stable product has been tuned to it.

2. Speed of delivery. The faster a reward arrives, the harder the brain pulls toward it. Crack hits harder than snorted cocaine for the same reason an iced Coke hits harder than a piece of fruit: glucose dissolved in cold acidic water reaches the brain's reward circuitry in roughly 600 milliseconds, before the "stop" circuits in the orbitofrontal cortex have time to weigh in. Liquid sugar bypasses chewing, fiber, and the time-delayed fullness hormones your gut releases when food arrives in normal form. By the time the brake engages, the calories are already in.

3. Variety and ubiquity. Your body has a built-in brake called sensory-specific satiety: get bored with one flavor and you stop. In the late 1980s, Kroger ran a "Variety Research Program" with Nabisco, Frito-Lay, Kellogg's, Coca-Cola, and General Mills to study which shoppers — "variety seekers" — bought and ate the most when novelty was on offer. The line-extension supermarket (twelve flavors of Doritos, nine variants of Oreo, an aisle of yogurts) is the direct output. Smorgasbord effect, industrial scale.

These levers are not secrets — they are line items in trade publications. The food works the way it was designed to.

Wanting vs. liking — the addiction signature

Neuroscientist Kent Berridge, at the University of Michigan, spent decades showing that two brain systems we usually conflate are actually separate. "Liking" is the pleasure of a reward — opioid-driven, fleeting. "Wanting" is the urge to pursue a reward — dopamine-driven, motivational, prone to escalation. In a healthy system, they track together. In an unhealthy one, they pull apart.

Eric Stice's fMRI work at Oregon Research Institute found the signature plainly: as people gain weight, they want palatable food more without liking it more. Pursuit grows; payoff doesn't. That decoupling is the neurological fingerprint of addiction — what cocaine looks like in the brain, what nicotine looks like, what the sugar-fat-salt combinations on supermarket shelves look like.

David Kessler, the former FDA commissioner, gave the resulting behavior a clinical name: conditioned hypereating. Three symptoms: loss of control around certain foods, inability to feel satisfied, preoccupation with food between meals. In his analysis of the Reno Diet Heart Study, about half of overweight participants and seventeen percent of lean ones met the pattern. This isn't gluttony. It's a learned response to engineered cues — circuitry doing exactly what evolution built it for, applied to the wrong stimulus.

If you have noticed this in yourself, you are noticing a real thing. You did not invent it.

What this means for what you can actually do

This is the practical part. Three moves, in order of leverage.

1. Change the food around you, not the willpower inside you. Hall's trial worked because the food in the ward was simply different. You can't lock yourself in a metabolic ward, but you can do a smaller version: stop bringing the most engineered products home. The decisions you make once at the grocery store have outsized leverage over the dozens you would otherwise make at 10 p.m. on the couch. White-knuckling past a bag of Doritos in your kitchen is the worst possible terrain for that fight. Not having the bag is.

2. Use the NOVA-4 markers as your one filter. Module B2 goes deeper, but the short version: NOVA Group 4 is anything formulated from ingredients you don't own — modified starches, refined seed oils, protein isolates, emulsifiers, gums, low-calorie sweeteners, "natural flavorings." If it's made in a plant, not from a plant, that's the tell. Read the ingredient panel. If three or more lines look like a chemistry textbook, it's NOVA 4.

3. Slow the delivery. Speed is the single biggest lever the industry pulls. You can pull it the other way. Cook for yourself even imperfectly. Eat at a table without a screen. Drink water before a sweet drink. Chew. None of this is moral — it's mechanical. Every step that puts time between you and the calories gives your body's stop circuits a chance to fire before the go circuits have won.

None of this requires you to be a better person. It requires you to make the food around you a worse opponent.

Frequently Asked Questions

So is food addictive? What does "addictive" actually mean here?

The DSM-5 definition of addiction is "a repetitive behavior that some people find difficult to quit." Ultra-processed food clears that bar: roughly 15 percent of US adults meet the Yale Food Addiction Scale's criteria, comparable to alcohol. "Addictive" doesn't mean everyone who eats a Pop-Tart is hooked. It means the substance can produce compulsive use in a substantial minority, with the same brain signature as other addictive substances.

If willpower doesn't work, what about people who do successfully change their eating?

They almost never describe it as "trying harder." They describe restructuring their environment — what's in the house, what they buy on Sunday — so fewer decisions get made in cue-rich moments. They use specific if-then rules ("I don't drink soda with meals") rather than vague intentions. The work is real; it just happens upstream of temptation.

Are all processed foods bad? What about frozen vegetables?

No. NOVA Group 1 is whole foods. Group 2 is culinary ingredients like oil and salt. Group 3 is foods made by combining the two — canned beans, frozen vegetables, real bread, smoked fish. None of this is the problem. Group 4 — ultra-processed — is the category Hall's trial isolated. Frozen broccoli is fine. Frozen broccoli "alfredo cups" with modified starch and natural flavor are a different food.

Why does this affect some people more than others?

Genes load the gun; environment pulls the trigger. Clare Llewellyn's UK twin studies found obesity heritability runs around 40 percent in food-secure households and over 80 percent in food-insecure ones — same genes, very different outcomes. Stress, sleep loss, and childhood food exposure all change how strongly a person responds to engineered cues.

Doesn't this just give people an excuse to overeat?

The opposite, in practice. People who believe overeating is a moral failure cycle through shame, restriction, and rebound binges — what dietitians call the what-the-hell effect. People who treat it as a food-environment problem make structural changes that hold. Shame is not a useful fuel. Accurate causation is.

What about GLP-1 drugs — does this mean those just work because they kill appetite?

Roughly, yes. Semaglutide and tirzepatide quiet the reward-driven "wanting" circuit — patients describe food going quiet in their heads, the same clinical signature phen-fen produced in the 1990s. The drugs are doing pharmacologically what changing the food environment does mechanically. They are not magic. They are evidence the problem was never willpower.

Where do I start tomorrow?

One swap, not ten. Pick the single ultra-processed product you eat most days and replace it with a NOVA 1–3 version of the same role. Soda becomes sparkling water. Cereal becomes yogurt and fruit. One swap held two weeks beats an overhaul that lasts three days.

Sources

1. Hall, K.D. et al. (2019). "Ultra-Processed Diets Cause Excess Calorie Intake and Weight Gain: An Inpatient Randomized Controlled Trial of Ad Libitum Food Intake." Cell Metabolism 30(1): 67–77. DOI: 10.1016/j.cmet.2019.05.008.
2. van Tulleken, C. (2023). Ultra-Processed People: The Science Behind the Food That Isn't Food. W.W. Norton.
3. Moss, M. (2021). Hooked: Food, Free Will, and How the Food Giants Exploit Our Addictions. Random House.
4. Moss, M. (2013). Salt Sugar Fat: How the Food Giants Hooked Us. Random House.
5. Kessler, D.A. (2009). The End of Overeating: Taking Control of the Insatiable American Appetite. Rodale.
6. Berridge, K.C., Robinson, T.E. (2016). "Liking, wanting, and the incentive-sensitization theory of addiction." American Psychologist 71(8): 670–679.
7. Araújo, J., Cai, J., Stevens, J. (2019). "Prevalence of Optimal Metabolic Health in American Adults: NHANES 2009–2016." Metabolic Syndrome and Related Disorders 17(1): 46–52.

Related modules

• B2: Real food vs. edible foodlike substance →
• B5: Hunger, fullness, and the no-shame default →

Related glossary terms

• Bliss point
• NOVA classification
• Hyperpalatable
• Conditioned hypereating
• Ultra-processed food
• Sensory-specific satiety
• Wanting vs. liking