The Reverse Aging Challenge, a live application of the Human OS framework. Boost your healthspan using science-backed, natural tools.
See upcoming editions

The Trap Is Also the Business Model

Why seeing the AI productivity trap is not the same as being free from it

Core EssayJune 28, 2026On Substack

This essay is part of an ongoing exploration of the human operating system.

You can also read this essay on Substack.

The most interesting reaction to my last essay on AI was not disagreement. It was recognition followed by resignation.

People understood the argument quickly. AI is making us more capable, but it is not automatically making us more coherent. It helps us produce more, move faster, generate more options, draft more things, summarize more inputs, and multiply what one person can touch in a day. That part no longer feels surprising to most people paying attention.

The more revealing signal was that almost nobody seemed surprised by the cost.

People could already see the trap. They could see that more output creates more output to review. They could see that faster tools raise expectations. They could see that automation does not always create freedom. Sometimes it creates more supervision, more decisions, more open loops, and more uncertainty about what actually matters.

That is where the harder truth begins. The trap is also the business model. It is the paycheck, the KPI, the client expectations, the founder's runway, the manager's quarterly targets, the creator's visibility, the consultant's responsiveness, the employee's perceived relevance, and the operator's survival strategy.

That is why awareness alone does not solve it.

A person can recognize the trap and still keep walking into it because the surrounding system rewards participation. The trap is not hidden. It is incentivized.

Awareness is not enough when incentives remain intact

This is where most productivity conversations become too narrow. They talk as if overload is mainly a personal boundary problem. Use the tool better. Turn off notifications. Block the calendar. Protect deep work. Build better habits. Be more intentional.

Some of that is useful. I am not dismissing personal responsibility. But personal responsibility becomes dishonest when it ignores the incentive system surrounding the person. If an organization rewards visible speed, the person who moves slowly may be wiser, but may also look less valuable. If clients reward instant responsiveness, the person who takes time to think may produce better work, but may also seem less committed. If a market rewards volume, the person who chooses higher quality may have a stronger signal, but a weaker algorithmic footprint.

Systems do not need to fool people in order to shape behavior. They only need to control incentives.

AI did not create this pattern. It accelerates it. The productivity rat race did not begin with AI; AI is making the pattern harder to ignore. Before AI, modern work was already full of contradictions. We said we valued strategy, but rewarded responsiveness. We said we valued creativity, but filled calendars until no one had space to think. We said we valued wellbeing, but treated recovery as an individual responsibility after work had already consumed the person.

AI enters that environment and removes friction. At first, that sounds unambiguously good. Less friction means less wasted effort, less waiting, less manual work, less repetitive tasks, and less blank-page resistance. But when a system has no clear mechanism for stopping, friction is not only a constraint. Sometimes it is the last remaining brake.

Remove the brake and the system does not automatically become wiser. It usually becomes faster. If the underlying operating logic remains "more output, more responsiveness, more visibility, more growth," AI will not correct that logic by itself. It will amplify it.

The surplus rarely becomes recovery. It becomes more work.

The supervision load is real

One of the clearest patterns in the AI conversation is that work does not simply disappear. It changes shape.

AI can draft the document, but someone still has to decide whether it is true, useful, appropriate, legally safe, strategically aligned, and worth sending. AI can summarize the meeting, but someone still has to know what mattered. AI can generate ten options, but someone still has to choose. AI can produce the first version, but someone still has to hold the standard. AI can automate parts of a workflow, but someone still has to supervise the system, detect errors, notice missing context, and understand when a clean answer is quietly wrong.

That is not trivial.

A lot of AI adoption is sold through the language of time savings, and some of those savings are real. But saved time often returns as monitoring load. The person is no longer only doing the work. They are also supervising the output of a system that can produce work faster than the person can fully metabolize.

That may still be useful, valuable, and commercially necessary. But it is not the same as freedom.

It is a different form of load.

This is why the AI productivity conversation needs more precision. If a tool reduces execution time but increases review burden, decision fatigue, option volume, and responsibility for errors, the human operating system has not been liberated. It has been moved into a different role in the production chain. The work may become more abstract, more leveraged, and more scalable, but it can also become more cognitively and ethically demanding.

The question is not only whether AI saves time. The question is what kind of work replaces the time it saves.

Marketing already showed us the pattern

I have seen parallels in B2B marketing.

Most good marketers already know that trust matters. Customer research matters. Positioning matters. Brand memory matters. Community matters. Timing matters. Language matters. The work that creates real demand is often slower, more contextual, and harder to defend in a short reporting window.

But many marketing teams are still forced to defend themselves through the narrow language of pipeline, attribution, MQLs, conversion metrics, and quarterly performance. They know the deeper work matters, but the system funds what can be measured quickly.

People do not always surrender to shallow work because they are shallow. They surrender because the deeper work is harder to protect inside a system that rewards immediate proof.

AI may do something similar to knowledge work. Many people already know that judgment, context, trust, craft, recovery, and human understanding matter. But if the organization only rewards visible speed, the wiser thing becomes harder to choose. Coherence starts to look like a luxury.

This is how incentives quietly train culture. People may privately value depth, but publicly perform speed. They may understand that the good work requires space, but still keep producing what the dashboard recognizes. Over time, the system does not only change what gets done. It changes what feels responsible.

Founders know this from the inside

This is especially obvious for founders.

Most founders do not need another article telling them they are the bottleneck. They know. They know they are in too many conversations. They know too many decisions depend on them. They know the business has adapted around their availability. They know the system is too dependent on their nervous system, attention, urgency, and ability to absorb exceptions.

But being the bottleneck often feels safer than redesigning the business while it is still moving. Dependency is expensive, but redesign feels risky. So the founder stays inside every Slack thread, every client exception, every sales conversation, every operational fire, and every decision that "only takes five minutes," until the entire business becomes a machine for converting founder capacity into forward motion.

AI can make this worse. Not because the tools are bad, but because they can give the founder more capacity before the business has learned how to use capacity wisely. Now there are more drafts, more options, more automations, more content, more experiments, more channels, more data, and more things that could be done.

The question becomes less "Can we do this?" and more "Should this enter the system at all?"

That question is harder than it sounds because founders are rewarded for possibility. They are trained to see openings, seize leverage, move quickly, and find ways through constraints. AI expands the field of possible action. Without restraint, that expansion can look like opportunity while functioning as overload.

The body knows before the business admits it

At the individual level, this pattern is familiar. When capacity is low, people often add more systems: more lists, more dashboards, more automations, more calendars, more trackers, more prompts, more productivity methods. Some of that structure helps. Some of it simply conceals the fact that the system is no longer recovering.

This is where the body becomes inconvenient.

The body does not care that the dashboard looks cleaner. It does not care that the workflow is more efficient. It does not care that the calendar is color-coded, the CRM is updated, the project board is organized, and the AI assistant is producing clean summaries.

The body asks a simpler question: can this system restore itself?

If the answer is no, the system is not coherent. It may be productive, impressive, profitable for a while, and even mistaken for growth. But it is not coherent.

That distinction matters because many people do not collapse dramatically. They keep functioning. They answer the messages, attend the meetings, maintain the appearance of control, and continue producing. The cost appears as a narrowing of range: less patience, less depth, less recovery, less play, less embodied presence, less tolerance for ambiguity, and more reliance on urgency to feel organized.

The business may not notice at first because the output continues. The body notices earlier.

The professional cost of restraint

This is why the "just slow down" answer is incomplete.

People are not afraid of work. They are afraid of what slowing down might cost. Someone at the brink of burnout may know that another yes is expensive. Their body may remember the last time they overcommitted. Their sleep may already be unstable. Their attention may already be fragmented. Their relationships may already be absorbing the overflow.

But the professional environment often gives them bad options: keep up and pay later, or slow down and risk being seen as less committed.

That is not only a personal dilemma. It is a design problem.

When work systems reward output but externalize recovery, people learn to treat their biology as a private constraint rather than a shared operating condition. AI intensifies this because it raises the ceiling of possible output. The more that becomes possible, the more restraint has to be designed. Otherwise the system will keep expanding until the person becomes the bottleneck, the shock absorber, or the failure point.

This is one reason many people recognize the trap but do not resist it. They are not irrational. They are responding to the incentives around them. The behavior may be costly, but within the current system it can still look professionally rational.

That is the deeper problem.

Stress is not the enemy

This is not an argument against pressure. Humans have adaptive systems. We need challenge, load, friction, and meaningful demand. Capacity grows through stress and recovery, not through comfort alone.

The issue is not stress. The issue is chronic mismatch: demand without recovery, speed without integration, output without closure, information without meaning, technology without downshift, growth without redesign, capability without coherence.

That is where the Human OS lens matters. A human being is not a productivity stack. A person is a biological system operating inside conditions.

Change the tools and you change the conditions. Change the pace and you change the conditions. Change the incentives and you change the conditions. Change the expectations and you change the conditions.

The question is not only what AI allows us to do. The question is what kind of organism, team, company, and culture it trains us to become.

The opposite of the AI rat race is not refusal

Refusing AI is not a serious option for most people. That may be satisfying as a posture, but it is not realistic for many operators, founders, teams, marketers, consultants, researchers, creators, and professionals whose fields are already being reshaped.

The better question is not whether to participate. It is how to participate without surrendering the entire operating system to speed.

That requires designed restraint, not vague balance.

Designed restraint asks where AI creates useful leverage and where it creates unnecessary output. It asks where AI reduces cognitive load and where it increases supervision load. It asks where AI improves decision quality and where it floods the system with options. It asks where AI helps people recover time and where it silently converts recovered time into more work. It asks where AI supports judgment and where it replaces the conditions that judgment requires.

These are not philosophical questions. They are operating questions.

For an individual, this may mean deciding which tasks should not be accelerated because the slowness is part of the thinking. For a founder, it may mean refusing to add another channel just because AI makes content easier to produce. For a team, it may mean designing explicit stopping rules, review standards, decision rights, and recovery windows before automation increases surface area. For an organization, it may mean measuring whether AI adoption actually reduces unnecessary load, or merely increases throughput.

Restraint is not resistance to technology. It is the discipline that allows technology to become useful without becoming totalizing.

The collective trap

The AI productivity trap becomes collective when everyone can see it privately but keeps participating publicly.

That is where we are heading. Many people know that more does not always mean better. They know that faster does not always mean clearer. They know that constant responsiveness is not the same as commitment. They know that productivity can become a socially acceptable form of compulsion. They know that tools designed to save time often raise expectations for how much can be done. They know the system is speeding up.

They also know that opting out can carry a cost.

So the public behavior continues, even when private awareness has already changed.

That is what makes the next phase so important. The question is not whether AI will make us more capable. It already is. The question is whether we can build the individual, organizational, and cultural restraint required to keep capability from becoming compulsion.

Because the trap is not hidden. It is incentivized.

And the first real act of intelligence is not seeing it. It is redesigning the conditions that make surrender feel rational.

Sources and signals