The more I use AI, the more human I have had to become.
That sounds paradoxical, and I suppose it is. But after working with AI tools intensively over the past two years, and watching senior leaders across Hong Kong and Singapore navigate the same shift, I keep returning to the same strategic frame: the barbell.
In finance, the barbell strategy means concentrating your exposure at the extremes and leaving the middle deliberately thin. Maximum safe assets on one end, maximum high-risk bets on the other, very little in the conventional middle ground. Nassim Taleb made it famous as a way of thinking about asymmetric risk.
I think it maps cleanly onto how individuals should position themselves in the age of superhuman AI. And most leaders are getting it exactly backwards.
The Two Ends Worth Owning
On one end of the barbell: get as close to AI as you can. Not just use it, but embed it into the grain of how you work. Play with the models. Use them to think through hard problems and teach you things you do not yet understand. Build with them, write with them, understand what makes them tick. Run experiments. Have fun. Let AI become a genuine colleague in your daily practice, not a productivity add-on you reach for occasionally.
On the other end: go as deep as possible into the things that genuinely resist automation. Learn to tell stories that land in a room full of skeptical people. Build a team that trusts you through real difficulty, not just the easy runs. Lead with courage when the outcome is unclear and nobody is watching you for cues. Confess your fears to someone who matters. Debate persuasively for a position most people will not touch. Write something by hand and actually mean it. Fall in love with an idea that takes years to develop. These are not soft skills. They are the things that compound most powerfully when everything else can be replicated at zero marginal cost.
The Middle Is Where I Would Be Careful
The conventional middle is the dangerous ground: the mechanical-but-cognitive work that once signalled effort and intelligence. Building spreadsheets. Following playbooks. Searching for needles in haystacks. Balancing accounts. Writing functional code by hand. Summarising meeting notes. Producing the first draft of every document from scratch.
None of this is worthless. But AI does it faster, better, and at effectively zero marginal cost. That is not a threat to manage with anxiety; it is a redirect worth acting on. The time and energy you free from those tasks has somewhere better to go.
What This Looks Like in Practice
The leaders I see thriving right now are doing exactly this: intensely curious about AI, genuinely invested in the human skills that do not compress. They are experimenting with agentic systems and autonomous workflows on Monday, and coaching a difficult team conversation with nuance and presence on Wednesday. Both investments compound.
The ones struggling are frozen somewhere in the middle. Half-protecting old workflows because they built careers on them. Half-experimenting with AI tools without committing deeply to either end of the barbell. Hoping the disruption slows down enough for them to catch up at a comfortable pace.
It will not slow down. But the response to that is not panic. It is positioning.
The Practical Question
The barbell works because it forces a choice. You cannot protect the middle and own the extremes at the same time. The question worth sitting with is not “how do I stay relevant?” but “where on this barbell am I actually spending my time and credibility this quarter?”
The interesting work has always lived at the extremes. AI has just made the cost of staying in the middle much higher than it used to be.
Venture to the extremes. That is where all the interesting work has always been anyway.