Briefing #3: The Last Human Advantage
Why the most valuable AI skill isn't technical.
Note: This briefing was originally published on LinkedIn on August 8, 2025. It has been migrated to our new home on Substack to create a complete archive. Multi-format features like video and audio commentary are available for all new briefings published from April 2026 onwards.
The headlines are unsettling. We’re seeing layoffs in data science and software engineering — the very jobs we were told represented the future of work. A quiet anxiety is rippling through the knowledge economy: if even the technologists aren’t safe, who is?
This isn’t just fearmongering. It’s a rational response to a profound market shift. A recent, extensive study from Microsoft Research analyzing 200,000 real-world AI conversations confirmed that the highest “AI applicability” is in highly technical and knowledge-based roles, including CNC Tool Programmers, Mathematicians, and Data Scientists, suggesting such roles may well be among those most affected by AI in the long term.
This creates a stunning paradox. The very technology that was supposed to create a new generation of technical experts is now advanced enough to automate many of their core tasks. The “AI skills gap” we’ve been panicking about for years is being closed not by human training, but by the technology itself.
But this “AI culling” of technical tasks doesn’t mean human value is obsolete. It means the source of defensible, long-term value has shifted from the hands to the head.
When the machine can handle the technical execution — writing the code, analyzing the data, drafting the report — the human’s primary, irreplaceable role becomes the strategic direction and critical evaluation of that work.
This is the last human advantage.
The new, more urgent skills gap isn’t about knowing how to use the tools; it’s about the uniquely human ability to question them. It is the capacity for:
AI Judgment: The ability to critically evaluate an AI’s output, spot subtle biases, and know when a technically correct answer is strategically or ethically wrong.
Workflow Design: The wisdom to design new processes that artfully blend human insight with machine efficiency, rather than just automating broken ones.
Ethical Reasoning: The foresight to ask “Should we do this?” not just “Can we do this?”
For years, we’ve focused on teaching procedures: how to operate the machine. We must now shift to teaching principles — how to think alongside it. The goal for leaders is no longer to build a technically-literate workforce, but to cultivate an “AI-wise” one.
This requires a fundamental change in how we hire, train, and measure performance. It means valuing the employee who asks the uncomfortable question over the one who simply accepts the first answer. It means rewarding strategic skepticism over blind adoption.
It reminds me of a meeting I once had as a graduate student with my supervisor. He was reviewing my undergraduate transcript and noticed I had taken a philosophy course on “Critical Thinking.” He mused casually, “Critical thinking. Is there any other kind?”
At the time, it seemed “no” was the obvious answer.
But the age of AI is teaching us that the real answer is, “Yes, there absolutely is.”
The future of knowledge work will not be defined by those who can simply operate algorithms, but by those who have the wisdom to direct and doubt them — those who take critical thinking to the next level. That is the uniquely human advantage no AI can (yet) replicate.
Hands, to head to heart.