A recent modeling study suggested that reducing ultra-processed food consumption could prevent thousands of heart-disease deaths, with researchers arguing that public education and individual advice are not enough when the food environment itself keeps making degraded inputs easy, cheap, branded, and normal.

As always, the debate immediately became more complicated than the headline. Modeling is not proof. Not all processing is the same. Not every packaged food deserves the same judgment.

Food choices are shaped by money, time, access, culture, stress, labor, and the realities of modern life. That complexity matters, but so does the pattern.

A few weeks ago, at a networking event in New York, someone used a word I had heard before, but not often enough to feel how much work it was doing.

Enshittification.

The word is ugly, which is partly why it works. It does not try to make degradation sound strategic. It does not hide decline behind language like optimization, personalization, monetization, engagement, growth, convenience, or user experience. It says what many people feel when something that used to work gradually becomes worse while still demanding their participation.

At first, I felt amused by the honesty of it. Then relieved. Then concerned.

Relieved because I realized I was not the only person noticing the pattern. Concerned because once we have a name for this level of deterioration, we might start treating it as a structural condition rather than something that should still offend us.

That is the part I cannot quite shake.

As a former software engineer and technology leader, I find some modern software failures strangely personal. Not because I expect perfection. Software has always shipped under pressure. Trade-offs are real. Technical debt is real. Teams make mistakes.

But some of what makes it into the market now feels different: obvious bugs, broken flows, confusing interfaces, features designed around internal metrics rather than user clarity, apps that add friction while claiming to simplify, changes that break familiar workflows without improving the underlying experience. Products often seem to have shipped because someone needed a release, a growth story, an AI feature, or proof that the roadmap was moving.

There are moments when it feels offensive. Not because the app inconvenienced me, but because the standard feels lower than it should be, and because the user is quietly asked to absorb the difference.

That is where software and food start to look less separate than they appear.

In both cases, the surface story is convenience. The hidden story is the transfer of cost into the user.

The user becomes the patch

Cory Doctorow popularized “enshittification” to describe a particular pattern of platform decay. In his formulation, platforms first serve users, then shift value toward business customers, and eventually extract more value for themselves at the expense of both. It is a platform-capitalism argument, not simply a complaint about annoying apps.

But the word has traveled because people recognize the broader feeling. Things that once seemed useful, elegant, trustworthy, or clear gradually become more extractive, more cluttered, more fragile, more manipulative, or simply worse.

In software, degradation often shows up as extra cognitive labor.

The product still exists. The interface still loads and the feature still technically works, but the user now has to double-check, distrust, repeat, navigate, supervise, and remember more. A form does not save properly, so the user screenshots it. Search gets worse, so the user learns workarounds. The interface changes, so the user relearns a task that was already solved. Settings move, so the user hunts for control. Automation makes a mistake, so the user becomes the auditor. Support disappears, so the user negotiates with a chatbot.

The user becomes the patch.

Bad software does not only waste time. It trains a state. It trains suspicion, impatience, monitoring, interruption, and low-grade vigilance. It teaches the body that the tool cannot fully be trusted, even as the person remains dependent on it.

A broken interface is input. A confusing flow is input. A dark pattern is input. A tool that keeps moving the controls is input.

From a Human OS perspective, software is not outside the person. It enters the system through attention, expectation, frustration, decision load, and trust.

The app is not just a tool. It is part of the operating environment.

Convenience hides the cost

The reason this matters is that degradation rarely arrives as obvious loss. It often arrives as convenience.

I am not against convenience. Good convenience reduces friction without degrading the underlying quality of the experience. A well-designed tool can remove unnecessary labor. A good interface can protect attention. A thoughtful product can make a complex task easier without making the user weaker, more dependent, or less perceptive.

The problem is not convenience itself.

The problem is convenience that hides the cost by moving it somewhere harder to measure.

Sometimes the cost moves into attention. Sometimes into trust. Sometimes into the body, into culture, into a loss of standards so gradual that people stop noticing the loss.

This is not only a software story. We have made this trade before.

When music became more portable, access exploded and entire catalogs moved into pockets. That was extraordinary, but it also changed the ritual of listening, the relationship to albums, the meaning of ownership, and for many listeners, the expectation of fidelity. Moving from vinyl to compressed digital formats was not simply decline. It was a real gain in access and convenience. But it also trained a different relationship to sound.

The question is not whether the trade was good or bad. It is what the trade trained us to stop noticing.

Ultra-processed food is the enshittification of nourishment

Food may be the clearest example.

Ultra-processed food is not just junk food with a scientific name. In the NOVA classification, ultra-processed foods are industrial formulations made through processes and ingredients that are not typically part of home cooking. They are often designed to be convenient, shelf-stable, highly palatable, branded, easy to distribute, and easy to consume.

This does not mean all processing is bad.

Cooking is processing. Fermentation is processing. Freezing can preserve food. Canning can make food accessible. Olive oil, yogurt, cheese, coffee, bread, and many traditional foods involve forms of processing. Human beings have always transformed food through culture, technique, preservation, and care.

The problem appears when nourishment is replaced by engineered consumption.

The food-like product becomes easier to store, sell, transport, market, and consume. It becomes more profitable at scale. It becomes more intense to the palate and less demanding to prepare. It solves real problems of time, price, portability, and access.

But the body absorbs the hidden cost.

Satiety changes. So does appetite regulation, texture, fiber content, and eating speeds. The relationship to cooking changes. The meaning of a meal changes. The expectation of what food should provide changes too.

The product gets easier to consume while becoming harder for the system to trust.

That is the same operating pattern.

Bad software makes the user the patch.

Bad food makes the body the patch.

Not all degradation feels like loss

This is why the problem is hard to see clearly.

A degraded input may feel better at first. New app features ship faster. Photos and videos is more portable. The service is cheaper. The interface is more automated. The answer arrives instantly. The product is always available. The food is easier to consume.

Something is gained.

But the gain can obscure what has been removed. A product can become more available while becoming less trustworthy. A food can become more stimulating while becoming less nourishing. A tool can become more powerful while making the user more dependent. A platform can become more engaging while becoming less human. A system can reduce friction while increasing dysregulation.

This is the pattern worth naming.

Modern markets are very good at producing convenience. They are less reliable at protecting coherence.

Convenience usually measures the immediate transaction.

Coherence asks what happens to the system over time.

That distinction matters because the body lives with the aftereffects of inputs long after the transaction is complete. The user clicks, eats, scrolls, buys, downloads, replies, accepts, skips, and adapts. The market records engagement. The body records the cost differently.

The body knows the difference before the market admits it

The market can describe ultra-processed food as affordable, tasty, convenient, fortified, low-fat, high-protein, sugar-free, plant-based, energy-boosting, kid-friendly, performance-oriented, or lifestyle-compatible.

The body receives the whole input. Not only the label.

The same is true with software. A company can describe a product as AI-powered, intuitive, frictionless, personalized, productivity-enhancing, or user-centered. The nervous system receives the actual experience.

Did the tool reduce uncertainty? Did it protect attention? Did it make the task more legible? Did it create trust? Did it help the person complete the loop? Or did it add monitoring, friction, interruption, and dependence?

Quality is not aesthetic. It is regulatory.

A well-made tool changes the user’s state differently from a careless one. A nourishing meal changes the body differently from a food product engineered primarily for shelf life, margin, and repeat consumption. A quiet room changes the system differently from a noisy one. A relationship built on trust changes state differently from one built on performance.

Inputs are not morally neutral just because they are common.

They train the system.

Adaptation can mean lowering standards

One of the most dangerous things humans do well is maladapt.

We get used to the buggy app. We learn the workaround. We accept the degraded interface. We stop expecting customer support. We tolerate the chatbot. We stop noticing that food no longer feels like food. We forget what deeper listening felt like. We assume that worse is simply modern.

Adaptation is not always progress. Sometimes adaptation means lowering standards.

When a word like enshittification appears, it can create relief because it gives language to a shared perception. We are not imagining it. Something has degraded, there is a pattern, and other people see it too.

But language can also become resignation.

Once a decline has a name, people can start treating it as weather.

That is just how platforms work now. That is just how apps are now. That is just how food is now. That is just how customer service is now. That is just how modern life is now.

That is the surrender I worry about.

The point of naming a pattern should not be to normalize it, but to recover agency.

The Input System is not neutral

From a Human OS perspective, the deeper issue is input quality.

Input is not only information. It is everything that enters the system and shapes what happens downstream.

Food is input. So are the music we listen to, the places we inhabit, noise, and the software interfaces we use. Advertising is input. Social cues are input. Default settings are input. Customer service is input. Packaging is input. A broken system that the person has to use every day is input.

These inputs shape state. State shapes attention, behavior, recovery, output, pattern, and adaptation.

So when the quality of inputs degrades, the human system does not remain untouched and make rational choices from a neutral place. It compensates. It works around, lowers standards, burns attention, consumes what is available, and adjusts expectations. It becomes more vigilant or more numb.

That is the cost. Not only worse products: worse operating conditions.

This is why degraded inputs eventually become a health issue, even when they do not look like one at first. The body is not only affected by what we call health. It is affected by everything it has to process.

Convenience without care becomes extraction

The issue is not technology or food processing.

The issue is not scale or even profit by itself.

The issue is what happens when convenience is separated from care.

Good design cares about the user’s state. Good food culture cares about the organism, the meal, the land, the cook, the rhythm, and the context of eating. Good technology reduces unnecessary burden without increasing hidden dependence. Good markets can expand access without degrading what is being accessed.

But when convenience is optimized without care, the user eventually pays with attention, trust, health, or agency.

That is why this pattern matters beyond consumer frustration.

A society that accepts degraded tools, degraded food, degraded attention, degraded service, and degraded trust is not simply becoming less refined. It is asking human beings to regulate themselves inside lower-quality environments.

Then it calls the consequences personal failure.

Why are people tired? Why are people reactive? Why are people distracted? Why are people eating too much? Why are people mistrustful? Why are people overwhelmed by simple tasks? Why are people losing patience with systems?

Part of the answer may be that the systems are worse than they need to be, and the human organism has been absorbing the difference.

What refusing degradation requires

The answer is not nostalgia. I do not want to return to a world where everything was slower, more expensive, less accessible, or less portable. I do not want software to stop improving, food to become precious and inaccessible, or technology to serve only people with time and money.

The answer is not how to go backwards. It is how to stop confusing convenience with improvement.

That requires a higher standard.

Does the product make the human system more coherent? Does the tool reduce unnecessary load? Does the food nourish more than it stimulates? Does the service preserve agency? Does the interface create trust? Does the shortcut actually shorten the path, or does it move the cost into attention, recovery, and the body?

These are not romantic questions. They are design, health, and business questions. They are Human OS questions.

Because when enough inputs degrade, personal discipline cannot compensate forever.

A person can make better choices. Yes. But the environment still matters. If the environment keeps offering degraded inputs while calling them progress, convenience, personalization, or innovation, the burden quietly shifts onto the individual to notice, resist, repair, and recover.

That is not sustainable.

Worse should still offend us

This is what I want to preserve from my first reaction to the word enshittification. The amusement is useful. So is the relief and the shared language.

But the offense matters too.

Not outrage for its own sake. Neither nostalgia, nor moral panic. Offense as a signal that something in the system still recognizes a standard.

A product that used to respect the user should not degrade without resistance. A food system should not replace nourishment with engineered consumption and then frame the resulting dysfunction as lack of discipline. A tool should not make the user more vigilant and call that productivity. A market should not move cost into the body and call that convenience.

Bad software is not only annoying.

Ultra-processed food is not only a dietary issue.

Both reveal a larger pattern: inputs degraded by systems optimized for scale, margin, retention, and ease, while the human organism absorbs the hidden cost.

That is the part we have to keep seeing.

Because the real danger is not only that things become worse. It is that they become worse slowly enough that we adapt, and then forget that better was ever a reasonable expectation.

If this helped you name something you have been feeling but had not quite put into words, please share it with someone who might need the same language.

The Human OS Manual is now available for pre-order. If this work resonates, you can reserve your copy and follow the ideas here as they continue to develop.

Feel free to leave a comment. I read and respond to every thoughtful note.

Sources and notes

Cory Doctorow, “The ‘Enshittification’ of TikTok,” Wired https://www.wired.com/story/tiktok-platforms-cory-doctorow/

The Guardian, “‘What many of us feel’: why ‘enshittification’ is Macquarie Dictionary’s word of the year” https://www.theguardian.com/science/2024/nov/26/enshittification-macquarie-dictionary-word-of-the-year-explained

Food and Agriculture Organization of the United Nations, Ultra-processed foods, diet quality and human health https://www.fao.org/family-farming/detail/en/c/1735103/

Carlos A. Monteiro et al., “Ultra-processed foods: what they are and how to identify them,” Public Health Nutrition https://www.cambridge.org/core/journals/public-health-nutrition/article/ultraprocessed-foods-what-they-are-and-how-to-identify-them/E6D744D714B1FF09D5BCA3E74D53A185

Virginie Hamel et al., “Estimating the preventable burden of cardiovascular disease attributable to ultra-processed dietary patterns in Canada: A modeling study,” American Journal of Preventive Medicine https://pubmed.ncbi.nlm.nih.gov/41966464/

The Guardian, “Reducing ultra-processed foods could prevent thousands of heart disease deaths, study suggests” https://www.theguardian.com/society/2026/jul/15/reducing-ultra-processed-foods-could-prevent-thousands-of-heart-disease-deaths-study-suggests

Kevin D. Hall et al., “Ultra-Processed Diets Cause Excess Calorie Intake and Weight Gain: An Inpatient Randomized Controlled Trial of Ad Libitum Food Intake,” Cell Metabolism https://pmc.ncbi.nlm.nih.gov/articles/PMC7946062/

Melissa M. Lane et al., “Ultra-processed food exposure and adverse health outcomes: umbrella review of epidemiological meta-analyses,” The BMJ https://www.bmj.com/content/384/bmj-2023-077310