Recently, I was listening to a podcast1 with Mark Bertolini, the CEO of Oscar Health and former CEO of Aetna. His story is wild. He saved his son’s life from a rare form of lymphoma. He survived a ski accident that left him in a coma for five days. After the accident, he was pumped full of narcotics that dulled the pain but didn’t erase it. At one point, he planned to drive into a bridge and end his life, only to be pulled over by the cops. For the next 18 years, he slept one hour a night, pacing his room just to pass the time.
He describes the aftermath like this: “The right side of my brain literally shut down. So my level of empathy was pretty low, and the left side of my brain was a machine. It was like a supercomputer.” He became a “business Terminator”, still human, but emotionally disconnected, with analytical horsepower cranked to the max. During those years, he climbed the ranks at Aetna and eventually sold the company to CVS for $69 billion.
As I listened, I started to see his story as a metaphor for where we are right now with AI.
The Left/Right Brain Framework
The human brain can be thought of as split into two hemispheres. The left half is adept at language, logic, analysis, and reasoning. Its right counterpart handles creativity, emotions, embodiment, and contextual meaning. Author Iain McGilchrist describes the left hemisphere as “using narrow-beam attention to one detail after another, sees what is familiar, certain, static, explicit, abstract, decontextualized, disembodied, categorized, general in nature, and reduced to its parts.”
The right hemisphere is the exact opposite: “Bringing broad, open, sustained, vigilant attention to bear on the world, it sees what is fresh, unique, never fully known, never finally certain, but full of potential. It understands all that is, and must remain, implicit: humor, poetry, art, narrative, music, the sacred, indeed everything we love; it understands that nothing is ever static and unchanging, that everything is flowing and interconnected.”
Of course, both hemispheres are involved in almost everything. But McGilchrist’s framing is useful. In The Master and His Emissary, he describes the left brain as the emissary and the right as the master. The left works in parts and components; the right sees depth and the whole. His central point is critical to understand right now: the analytic power of the left should serve the integrated vision of the right, not the other way around.
Information Processing, Not Intelligence
I’m not trying to come off as a boomer Luddite who prefers to communicate via pigeon mail. I use AI tools every day. When it comes to writing code, parsing PDFs, or extracting insights from dense documents, LLMs are awesome.
However, I think we’re conflating information processing with true intelligence. It’s like sitting in a sauna at 190°F and convincing yourself you’re working out. You might be sweating, but you’re not getting any fitter. The same goes for AI. It moves fast and works efficiently, but that doesn’t mean it actually understands anything.
With tools like Suno, I can generate a song in a chosen genre with custom lyrics, but the result is a novelty to show friends, not something I’d add to my playlist. I can talk to ChatGPT about personal problems, but reasoning through anxiety with an LLM feels like running my thoughts through the wash with no detergent. Lots of motion, little cleansing.
My coaching clients notice the same. Using AI as a therapist feels innovative at first, but it doesn’t lead to meaningful transformation. And in companionship, the gap is even bigger. I value in-person presence far more than a call or FaceTime, and texting barely carries any emotional weight. We keep assuming LLMs can deliver embodied empathy and connection, when all they’re doing is processing inputs and predicting the next token.
AI as the Hyper-Left Hemisphere
Large language models reflect the same cognitive imbalance. They behave like hyperactive left brains by breaking wholes into parts, extracting patterns, and predicting what comes next without any sense of meaning.
The process begins with scraped internet text, stripped of context and lived experience. GPT-22 was trained on 40GB of Reddit posts and GPT-33 on 570GB from broader sources, but even at that scale, it’s still a map of abstractions, not the actual terrain of reality. That text is then split into tokens, disembodied fragments that the model learns to predict in sequence. The goal isn’t to understand. It’s to guess what comes next.
Then comes fine-tuning, where human experts label data, write evals, and reinforce preferred outputs. What wasn’t scraped during the internet sweep gets hunted down like rare loot, like Anthropic buying out-of-print books and then slicing pages up to feed to the models.
Not all information is documented online though. An indigenous healer’s tacit knowledge can’t be scraped and labeled. A professional athlete’s mastery is cumulated from thousands of hours of practice, coaching, and feedback. And someone like Rick Rubin seems to operate more off of vibes and taste than theory or technique. You could ask any of them how they do what they do, and you wouldn’t get a logical answer. Not because it’s not possible to learn, but because some forms of knowing aren’t transferable through indirect experience (or LLMs). And maybe that’s not a bad thing.
In deployment, the same constraints persist. LLMs rely on prompts and context windows, not situational awareness. They generate responses fluently and obediently, but never pause to resist or invite exploration. Whether it’s text, image, voice, or video, the interaction remains flattened. It’s still just high-speed information prediction severed from lived context.
Hemispheric Imbalance in Humans
Mark Bertolini’s 18-year ordeal as the “business Terminator” illustrates left-brain dominance: intensely detail-oriented while emotionally disconnected. He isn’t the only case.
Schizophrenics, thought to be in left-brain overdrive, perceive a fragmented world, often imagining people have turned into machines or zombies. Stroke victims with right-brain damage show a similar pattern, with flattened speech, monotone delivery, and a loss of irony, humor, metaphor, and emotion. With less embodied awareness, they end up forgetting their own limbs and miss social cues. As the left brain takes over, attention narrows into tunnel vision aperture, focusing only on what is explicit and graspable.
But right-brain dominance is just as limiting. Without the left brain’s language, logic, and control, people lose the ability to form coherent sentences even though they still understand tone, emotion, and metaphor. The right hemisphere processes richly and deeply, but without the left hemisphere’s capacity for expression and control, perception is rich while articulation and action remain stunted.
One hemisphere is not better than the other.
Humans (and the right hemisphere) as the Master
In the modern context, McGilchrist’s idea that the master must guide the emissary becomes a nested metaphor.
Originally, the right hemisphere, with its sense of the whole and its awareness, was meant to guide the left hemisphere, which specializes and works in parts. With AI, the parallel is clear: humans must guide the machine, not the other way around. Instead of being glued to technology like attention-sucking parasites, we should repurpose our relationship with AI more like an exoskeleton, a tool that multiplies our intentions and amplifies our power.
Seen this way, it makes sense why AI’s greatest value shows up in domains aligned with left-hemispheric strengths: programming, research, data analysis, and automation4. We already see products guiding programmers through nearly every step of development, as well as advances in drug discovery, where AI accelerates biology research, and in law, where it speeds up contract review. But why these areas? Is it simply that practitioners tend to be early adopters, or is it that the work itself is inherently suited to the strengths of LLMs? If it’s the latter, slow adoption elsewhere may reflect the natural limits of where AI has leverage.
A different perspective might help. In An Immense World, Ed Yong shows how every animal perceives reality through its own configuration of senses. A mantis shrimp, for example, has up to sixteen photoreceptor types compared to our three, which allows it to perceive 100–200 times more colors than we can. That range is not just invisible to us, it is inconceivable. We can’t even begin to fathom what it means to see colors outside the palette of human experience.
The same bias shows up with AI. What we call “superhuman” is always relative to human capacities. Because AI can generate words effortlessly, we mistake its fluency for depth, overlooking the resonance and sparks of novelty that emerge only through the slower, organic process of human writing.
The better analogy is a car. A car is not a genius because it outruns us on the highway. It is superhuman in one narrow sense: speed. AI is superhuman in information processing, but intelligence is more than pattern-matching or fluency. True general intelligence would require the right-brain capacities of context, empathy, embodied presence, and meaning-making—areas where AI consistently falters. In that sense, AGI as evocatively envisioned may not be possible. At best, we are building a savant emissary, not a master. The real leverage comes not from the raw horsepower of a model, but from the context and domain in which it is applied.
The top beneficiaries will be those who treat AI as the emissary, not the master. Those who recognize the irreplaceable qualities of human creativity and judgment and how to direct the machine’s horsepower will reap the rewards.
Creators will use AI to brainstorm and plan, researchers to accelerate discovery, and entrepreneurs to extend their productivity without large teams. In every case, the human sets the vision and directs the AI towards specific, discrete tasks.
Speculation is heated within the creative industries, but I don’t think movies will be replaced by entirely automated productions. Instead, AI will be folded into the workflow, taking on laborious tasks like storyboarding, special effects, or editing. Early Disney films required thousands of hand-drawn frames, until digital tools streamlined the process. The medium improved, but gradually, not through an overnight Cambrian explosion. I think AI will follow a similar path, enabling creators5 to focus more on vision and meaning while the machine carries the weight of execution.
The Path of Integration
Pre-AI, the world was already tilted toward left-brain thinking: optimization, fixation, endless measurement in the economy, policy, and culture. With the rise of what feels like “a country of geniuses in a data center,”6 it has never been more crucial to restore something more holistic and connected. That work begins at the individual level, as in each and every one of us.
You probably know someone (or are someone) who lives almost entirely in left-brain mode. The calendar is time-blocked and every minute is accounted for. Meals are tracked by macros and workouts are logged on Strava. Even something as simple as sleep is reduced to a single score. Relationships are managed through a personal CRM that reminds who to reach out to. Meditation is collapsed into a productivity tool. Every part of life becomes a system to optimize and a problem to solve. Life begins to resemble a factory. Hyper-efficient and structured, but hollow and bland.
Right-brain presence feels different. It relies more on energy level than amount of time available to make decisions. It’s when you lose track of time in conversation with a friend. It’s noticing the gut feeling that, if it had a voice, would say “I don’t know why, but this feels right” and trusting that. It’s when you pause mid-walk just to stare at a tree because something about it feels alive. It shows up as emotional awareness, connection to one’s body, creativity, imagination, and intuition. You can’t optimize your way into it. You have to feel your way there. Cultivating that presence isn’t just a luxury. It might be the only real antidote to the disembodied logic loop we’re building.
As technology advances, it becomes ever more vital to remember what it means to be human. As AI enters more of our livelihood, we need to cultivate embodiment, relationality, emotional awareness, and creativity with even greater care. The irony is that the very things that once looked like unproductive downtime, like playing outside, moving our bodies, and sitting in silence, are becoming essential if we want to feel human at all.
Most people assume adopting new technology simply means replacing the old. But we have to go further and ask what else must change if we want balance. The more I use AI, the more I value yoga, meditation, sports, and in-person hangouts. What once felt like optional hobbies, now feels necessary. For me, playing basketball with a bunch of dudes is a countermeasure against being sucked into a techno-void that looks like some AI-generated mishmash of WALL-E and Ready Player One. It’s how I preserve the human parts of me that machines will never be able to replicate.
Whatever your stance on AGI, the deeper point is that these mediums change us as users. Social media, smartphones, and now AI have nudged us toward disembodiment and over-fixation. Everyone is free to choose how to respond. I’m choosing to lead with the master, not the emissary.
Return of the Master
Mark Bertolini’s story fascinates me because not only does he recover from extreme left brain dominance to a more balanced way of being, he also happened to be quite financially successful both during the 18 years of one hour sleep a night and after his recovery.
In his “business Terminator” years he drove Aetna’s turnaround, producing a 652% shareholder return. Starting in 2015, he introduced yoga and mindfulness to the workplace, raised wages, paid down student loans, installed gyms and healthy food, and even rolled out pet therapy. In total, the company was investing 130 million dollars annually into employee well-being. Instead of profits plummeting, the opposite happened. Over his entire tenure, shares went from $9 to $208 before the business was sold to CVS for $69 billion.
Eastern practices became central to his life. On the podcast, I heard him mention yoga, Reiki, craniosacral therapy, and Samadhi - blissful concentration. He even has Soham, the Sanskrit phrase for “I am that,” tattooed across the back of his neck to remind himself that everything is one. Not exactly what you’d expect from an old white guy from Michigan who runs a multibillion health insurance company. Mark is part of a growing collective of heart-led business leaders proving that compassion and wholeness don’t have to be at odds with profits and growth.
The opportunity in front of us is to make the same shift as Mark, but at scale.
AI may free up time, but that doesn’t dictate how we’ll spend it. We can drown in shallow content and chase material distractions, or we can reorient toward presence, discernment, and deeper connection—with ourselves and each other.
McGilchrist reminds us that “Machines serve us well when they relieve us of drudgery, but human affairs must remain human.” He warns of the simulacrum, a hollow, unsatisfying veneer of life, a representation that resembles reality but leaves us empty. The danger isn’t that AI takes over, but that we forget what’s worth preserving. We pour our resources into making machines better. But rarely stop to ask what ‘better’ even means, or what we risk leaving behind.
As for myself, I’m paying closer attention to what feels natural. I’ve cut back on social media to reclaim my attention and remember what it’s like to feel grounded. Yoga, jiu jitsu, and pick-up basketball pull me out of my head and back into my body.
The choice is simple: use the machine like a tool, or let it use us. What I’m learning, and what I’m inviting you to consider, is that feeling fully alive doesn’t come from the sheen of social media or simulated conversation. It comes from presence, movement7, and connection. From doing real shit with real people.
P.S. I’m a coach.
I help ambitious people find work that actually matters and design their version of the good life. If that sounds interesting to you, learn more here.
specifically automation in simple domains that have objective outcomes that can be labeled good or bad, and not in complex domains that are more subjective and more creative
here I’m using creators quite broadly, I think everyone is a creator, whether they’re fully aware of that and doing something with it or not
This is a reference to Anthropic CEO Dario Amodei’s essay Machines of Loving Grace. Now, my take is that yes, it does seem like there’s a ton of geniuses packed into Claude or ChatGPT, but… it’s a very specific type of genius. It’s like a SAT wizard or human calculator. We still haven’t seen LLMs create in a generative, creative, novel way… yet.
I recently read this article on the healing powers of surfing. It brought up the point of just because something hasn’t been measured, doesn’t mean that it’s not real or true. We have yet to hook up a bunch of sticky pads to a surfer’s brain mid-barrel, but that doesn’t make the experience any less therapeutic, or any less awe-inducing. That’s been true in my own life too.