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

AI Is Making Us More Capable. But Not More Coherent.

What DES Málaga clarified about productivity, digital health, future skills, and the missing human operating layer

Core EssayJune 14, 2026On Substack

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

You can also read this essay on Substack.

I have been using AI heavily enough now to notice a strange inversion.

For most of my professional life, the constraint was often capability. Could I write the piece, design the solution, analyze the problem, organize the project, conduct the research, build a prototype, compare the options, find the angle, or create the next asset? Sometimes the answer was yes, sometimes no, and often the answer depended on time, skill, budget, software, collaborators, or the energy available to carry the work from idea to execution.

Now, very often, the answer is yes.

Not perfectly. Not without exercising judgment, verifying information, applying taste, making corrections, or maintaining editorial discipline. But yes more often than before, and much faster than before. That changes the problem. When a tool removes friction from almost every productive impulse, work becomes easier to start, easier to expand, easier to justify, and harder to stop.

A new question appears: what happens to the human system when doing more becomes almost frictionless?

That was the most useful thing DES clarified for me.

I did not arrive at the conference as someone looking at technology from the outside. I have spent more than two decades working inside technology, advertising, entrepreneurship, sales, cybersecurity, digital systems, and business transformation. I am not suspicious of technology by default. I am a technology person by training, career, and temperament, but my interest has shifted over time. I no longer look only at what a system can do. I look at what it asks of the human being inside it.

On the surface, DES was a major digital transformation conference, covering AI from a variety of angles including digital health, future of work, talent, hospitality, human-centered technology, predictive health, data, trust, interoperability, and enterprise transformation. Those are all familiar to me. I know the language, the incentives, the optimism, the vendor logic, the genuine usefulness, and the blind spots.

What interested me was not the fact that AI is changing work. That is already obvious. What interested me was the pattern underneath the sessions and hallway conversations: AI is increasing capability faster than organizations, workers, and health systems are able to redesign the human operating conditions required to absorb that capability.

That is where the conversation still feels incomplete.

AI expands capacity, but most systems convert extra capacity into more output. The missing conversation is recovery, trust, agency, attention, coherence, and adaptation.

The productivity surplus does not automatically become recovery

One of the most useful conversations I had at DES happened away from the stages, with people from the local startup community. The topic was familiar: AI and productivity. The tone was optimistic, but not naïve. AI helps us do things that used to sit outside our immediate skill set. Research, drafting, strategy, design, automation, analysis, prototyping, publishing, iteration. A founder or operator can now move across functions with much less friction than before.

That is genuinely useful. It is also dangerous in a very specific way.

When friction disappears from work, stopping becomes the new skill.

For founders, builders, and people with ADHD-like attention patterns, AI does not only make work easier. It removes friction from starting, expanding, iterating, chasing adjacent ideas, and producing the next version. The constraint is no longer only time, skill, or execution capacity. The constraint becomes stopping capacity.

That was the line that stayed with me from the conversation: I do not need an alarm to wake up, I need one to stop working.

That is not a half-joke about discipline. It is an operating-condition problem.

Productivity tools help people do more. But if they do not also help people downshift, recover, and close loops, they can turn capacity into compulsion. AI can make a person more capable while making the day less complete.

This is the part most productivity conversations miss. They ask what the tool can produce. They ask how much time can be saved. They ask which tasks can be automated, accelerated, or delegated to a machine. They ask how organizations can increase efficiency and how workers can keep up.

They don't often ask what happens to the human system after the friction is removed.

Where does the saved time go? Does it become recovery, reflection, better judgment, more time with family, more sleep, more movement, more time outside, more integration, or more human contact? Or does it simply become more work?

In many organizations, and in many ambitious people, the productivity surplus rarely becomes recovery by default. It becomes additional output, higher expectations, faster response cycles, more projects, more surface area, and less tolerance for delay.

This is how capacity becomes a trap. Not because AI is bad, but because the surrounding system is still organized around extraction.

Human-centered AI is not just a better interface

Another phrase that kept appearing at DES, explicitly or implicitly, was human-centered AI. It is a good phrase. It points in the right direction. It is also at risk of becoming another layer of acceptable language around unchanged operating models.

In many technology conversations, “human-centered” still means easier to use, less intimidating, more intuitive, more ethical, more transparent, more empathetic in tone, better designed, or more adoptable. Those things matter, but they are not enough.

A tool is not human-centered because people use it. It becomes human-centered only if it improves the operating conditions people live and work inside.

That distinction changes the test. The test is not only whether AI improves the process for the company. The test is whether it improves the experience and agency of the human being inside the process. Does it protect attention? Does it reduce unnecessary load? Does it make decisions clearer? Does it increase trust? Does it make recovery more possible? Does it give people better context, or simply make them responsible for acting on more signals? Does it make work more humane, or just make pressure more efficient?

This matters because “human in the loop” is also not enough.

A human somewhere in the system does not automatically make the system human. The question is whether the human understands the loop, has power inside the loop, knows what data is being used, understands what decisions are being shaped, and has enough context to remain accountable without becoming a decorative checkpoint.

Otherwise, “human in the loop” becomes a compliance phrase. The loop still belongs to the system. The human simply absorbs the ambiguity.

Employee experience can become the emotional wrapper around efficiency

The hospitality and talent conversations at DES revealed the same pattern from a different angle.

AI is being introduced into recruitment, customer experience, personalization, back-office operations, and service environments. In hospitality and food service, this is especially interesting because the product is not only the room, the meal, the app, the reservation, or the brand. The product is also the state of the people delivering the experience.

Guest experience is not created by technology alone. It is created through attention, timing, emotional availability, judgment, care, and the ability of employees to respond well under pressure. That means the operating condition of the employee becomes part of the product.

This is where “employee experience” can become slippery. At its best, it is a serious recognition that people cannot create good experiences for others while being poorly supported themselves. At its weakest, it becomes the emotional wrapper around operational efficiency.

A company says employee experience matters while the actual system asks people to move faster, absorb more change, manage more tools, respond across more channels, and remain emotionally available to customers through increasing complexity.

The language softens. The load increases.

This is not a critique of hospitality alone. It appears across corporate wellbeing, healthcare, education, management, and any service environment where humans are expected to create trust, warmth, judgment, and care while the system around them becomes more instrumented and less forgiving.

The people expected to create restorative environments need restorative operating conditions themselves. Without that, AI does not humanize the system. It optimizes the strain.

People are not upgradeable skill stacks

A future-of-work session brought another version of the same issue.

The language of skills over jobs is useful. Job titles have always been imperfect containers for what people actually do, and AI is making that even more obvious. Roles will change. Required skills will change. Career paths are already becoming less linear. People will need to keep learning, keep adapting, keep moving across functions, and keep making their relevance visible.

That is true. It is also incomplete.

Conversations about the future of work keep asking whether people have the right skills. They ask less often whether people have the operating conditions required to keep changing.

Many people will become portfolio workers, and nonlinear careers may be more realistic than the old ladder. But that also demands more self-regulation, identity flexibility, recovery, and social support than most organizations are designed to provide.

There is a hidden burden inside the skills conversation: keep learning, keep signaling, keep reconfiguring, keep demonstrating relevance, keep absorbing uncertainty, keep proving that you are still useful. That may be the reality of the AI era, but if the human operating layer is ignored, the result is not adaptation. It is chronic activation with better language.

People are not upgradeable skill stacks. They are biological systems operating under load.

A person can learn a new tool and still be exhausted. They can update a profile and still be disoriented. They can reskill and still lose coherence if the rate of change exceeds their ability to integrate it.

This is the distinction many transformation conversations still miss. Responsiveness is not the same as adaptation. A system adapts when it integrates experience over time and emerges with greater range, coherence, and capacity. A person who is constantly responding may look adaptive from the outside while becoming more brittle inside.

That matters for work, leadership, AI adoption, and every organization that thinks the human side of transformation can be solved by more training.

Detection is not prevention

The digital health conversations at DES added another layer.

The promise of AI in health is real. Earlier signals, better diagnostics, more personalized care, telemedicine, connected patient experiences, more efficient clinicians, smarter workflows, and better use of data are all important developments, especially in systems where access is uneven, clinicians are overloaded, and legacy infrastructure makes care harder than it needs to be.

But the same distinction appeared again.

Detection is not prevention. Detection tells you something may be happening. Prevention asks whether the person has the support, trust, agency, and daily conditions required to change the trajectory.

That gap matters.

Digital health is getting better at making the clinical system smarter. That is useful. The deeper opportunity is to make the patient's life more intelligible to the patient.

If data only helps the clinician decide what to do to the patient, the model remains paternalistic, even if the technology is advanced. If data helps the patient understand the operating conditions shaping their health, then the technology can become agency-building.

This is especially important in workplace health initiatives and predictive employee health systems. Health data is not the same category as industrial data. A company may be comfortable managing large datasets in business operations, energy, logistics, or production. Human biological data is different. It carries privacy, consent, trust, fear, identity, liability, and power.

The challenge is not only whether AI can identify risk. It is whether people trust the system enough to act before the risk becomes visible as disease.

That requires more than detection. It requires context, physician oversight, clear boundaries, human judgment, ethical data handling, and, most importantly, operating conditions that make change possible after the signal appears.

A cardiovascular risk flag does not change a life by itself. A sleep score does not protect sleep by itself. A brain scan does not redesign the day. A wearable can reveal stress, but it cannot decide whether the workplace, relationship, commute, schedule, food rhythm, alcohol pattern, loneliness, financial pressure, or recovery deficit is the part that needs attention first.

The signal matters. The life receiving the signal matters more.

The missing layer is not information

This is the thread that connected the conference for me.

AI adoption, human-centered technology, future skills, employee wellbeing, digital health, trust, and prevention can sound like different conversations. At DES, they were happening in different rooms, with different speakers, different commercial incentives, and different vocabularies.

Underneath, they were circling the same missing layer: the human operating layer.

Not the human as a user, employee, patient, talent, consumer, or data subject. The human as a living system that has to metabolize input, maintain state, recover from load, produce output, stabilize patterns, adapt over time, and remain coherent inside increasingly accelerated environments.

That is not a soft concern. It is the operating condition underneath every transformation project that depends on human judgment.

AI can produce more information, but more information does not automatically create better decisions. AI can reduce friction, but less friction does not automatically create more freedom. AI can improve detection, but earlier detection does not automatically create prevention. AI can personalize experiences, but personalization does not automatically create agency. AI can make work faster, but speed does not automatically create coherence.

A system can become more intelligent and still make the people inside it less capable of living well.

That is the risk.

What DES clarified

DES confirmed something important for me.

The Human OS conversation does not sit beside the AI conversation. It belongs inside it.

I say that as someone who has spent most of his professional life around technology and digital systems, not as someone arguing from outside the room. I still believe in tools. I still believe in intelligent systems. I still believe software can remove unnecessary friction, increase access, improve decisions, reduce waste, and expand what individuals and organizations are capable of doing.

But a serious technology conversation cannot stop at capability. It has to ask what happens to the human being once capability increases.

The market is talking about AI adoption, future skills, digital health, employee wellbeing, predictive health, trust, human-in-the-loop systems, and human-centered technology. Much of that conversation is still fragmented because it lacks a coherent model of the human being as an operating system.

That is the opening.

Not productivity, but capacity. Not wellbeing, but operating conditions. Not prevention as earlier detection, but prevention as changed trajectory. Not skills alone, but adaptation under load. Not human-centered AI as interface, but human-centered AI as environmental redesign.

The future will not be shaped only by organizations that adopt AI faster. It will be shaped by organizations and individuals that understand the human operating conditions required to use acceleration without becoming less coherent, less resilient, and less alive.

That is not anti-technology. It is the next level of serious technological advancement.

Because if the tool increases capability while the person loses attention, agency, recovery, trust, judgment, or adaptive range, the system has not become more advanced in any meaningful human sense. It has simply become better at converting life into output.

That is the test I want to keep applying as this conversation accelerates.

Does the technology make the human system more coherent? Does it help people think, recover, decide, relate, and adapt with more range? Does it make the day more inhabitable? Or does it make the old mismatch faster, smarter, and harder to question?

Those are not side questions. They are the questions that will determine whether AI becomes a genuine support layer for human life, or the most powerful productivity engine ever built on top of an already strained system.

Sources and notes

Some observations in this essay come from my field notes at DES Málaga 2026, including sessions and hallway conversations. External references below are included to verify event context and support the broader claims around digital transformation, AI, digital health, cardiovascular disease, AI regulation, and healthcare interoperability.