For years, the discourse surrounding artificial intelligence has been dominated by a singular, anxiety-inducing question: Who will AI replace? It is a framing that focuses on displacement, obsolescence, and the shrinking footprint of human labor in the face of machine efficiency. Yet, as the technology matures and integrates into our daily workflows, a far more compelling—and perhaps more optimistic—question is emerging: Who will AI release?
This question concerns the person whose mind produces ten ideas before breakfast only to lose eight of them by lunch. It refers to the employee who sees brilliant, non-linear connections that others in the room miss, but who remains unable to synthesize those insights into a memo that anyone will actually read. It speaks to the professional who has heard the refrain, "You have so much potential, if only you could get organized," in nearly every performance review of their career.
For the past century, the modern workplace has functioned as a rigid sorting machine. It did not prioritize intelligence, creativity, or raw judgment. Instead, it sorted people based on one dominant, quantifiable trait: the ability to convert complex thought into tidy, linear output, delivered on schedule and in a specific, standardized format. If an employee could master the packaging of their work, they advanced. If they could not, the quality of their underlying thinking often became irrelevant because, quite simply, no one ever saw it.
In decades of building software companies, the pattern has been consistent. The most original, innovative minds are rarely the most organized in the traditional sense. They think in complex webs rather than straight lines. They can hold an entire system in their heads yet consistently miss the status meeting regarding its status. Historically, promotions and professional validation have flowed toward those who write clean, predictable documents and manage their email inboxes with rhythmic, linear efficiency.
This reality has profound implications for entrepreneurship, a field that has long served as a refuge for these unconventional thinkers. Michael Freeman, a clinical professor of psychiatry at the University of California, San Francisco, has highlighted this phenomenon, noting that 29% of the entrepreneurs he studied reported having ADHD. With the CDC estimating that approximately 6% of U.S. adults—roughly 15.5 million people—live with an ADHD diagnosis, it is clear that many founders turned to company-building precisely because it was one of the few professional structures that did not actively punish the way their minds function. For these individuals, AI has arrived as a long-awaited intervention, removing the structural obstacles that have hindered their professional progress for years.
The Packaging Problem
Before an idea is ever treated as a serious contribution, it must be "packaged." This process involves hours of sequencing, structuring, formatting, and polishing. It is, in effect, an "organization tax"—a heavy toll levied between the moment of genuine insight and the moment of professional recognition.
This tax has never been applied equally. For a linear, methodical thinker, the distinction between thinking and packaging is negligible; the two motions are nearly synonymous. However, for a fast, associative thinker—of which the ADHD mind is a primary example—the packaging process is inherently punishing. Their insights arrive in fragments, often arriving in three threads at once and entirely out of sequence. Consequently, these ideas often die in a neglected notes app, arrive weeks past the deadline, or manifest as a rambling, incoherent contribution in a meeting, only to be restated by a more "organized" colleague who then receives the credit.
The workplace has historically pathologized this difficulty with packaging, framing it as a character flaw. Performance reviews have been littered with labels like "lacks follow-through," "needs to prioritize," or "not detail-oriented." For a century, organizations have essentially been grading the efficiency of packaging and calling it potential.
However, the lens is now shifting. If a career has been built primarily on the ability to package—defined by faithful execution, clean formatting, and the polished delivery of ideas that originated elsewhere—then AI is aimed directly at that skill set. The "Prompt Test" serves as a vital indicator here: if the instructions for a task can be dropped into an AI tool to produce the same quality of output, then that role is, by definition, automatable.
AI has absorbed far more than simple instruction-following; it has commoditized the very act of organizing. Structuring, summarizing, sequencing, and formatting are now available for the price of a standard software subscription. The trait that the old system selected for the most aggressively is now the first to be fully commoditized by technology. Those who struggle in this new era will not fail because AI has outthought them; they will fail because packaging was their primary product, and now that the product is free, they chose to defend it rather than evolve past it.
The Scattered Mind Gets a Translator
On the other end of the spectrum, a profound shift is occurring. For the first time, individuals whose minds operate in fragments possess a tool that can assemble those fragments into a coherent whole. One can speak to an AI in ten different directions, and the machine will synthesize those inputs into a clear, linear through-line. Three weeks of half-formed, chaotic notes can be fed into an AI interface to produce a structure that the creator recognizes as their own, yet finally sees as usable.
For many people with ADHD, this provides a rare opportunity to see the "bigger picture" of their own cognition. The intellectual dots were always there, but connecting them was the insurmountable hurdle. AI provides the bridge.
The evidence supporting this shift is not merely anecdotal. When the U.K.’s Department for Business and Trade conducted a three-month evaluation of Microsoft 365 Copilot across 1,000 employees, neurodivergent staff reported significantly higher satisfaction levels than their neurotypical counterparts, with many noting that the tool effectively leveled the playing field.
What the Research Shows
The implications of this shift extend well beyond neurodivergence. Research conducted by Stanford economist Erik Brynjolfsson, along with MIT’s Danielle Li and Lindsey Raymond, studied over 5,000 customer support agents using generative AI. Productivity rose by 14% on average, but the gains were not distributed evenly: the newest, least-skilled agents saw a 34% increase in productivity, while the most experienced agents saw almost no change.
Similarly, a study conducted by Harvard Business School and the Boston Consulting Group involving 758 consultants yielded a comparable distribution. The bottom half of performers improved by 43% when utilizing GPT-4, a gain more than double that of the top performers, who improved by 17%.
These findings suggest that AI is effectively compressing the "execution gap" while simultaneously widening the "judgment gap." Whatever factors previously separated the bottom of the distribution from the top—namely speed, polish, and adherence to format—AI is closing. What AI cannot supply is the quality of the raw, original thinking, the refined taste required to discern which of ten threads actually matters, and the critical judgment to realize when the underlying assignment itself is fundamentally flawed.
Ultimately, the question of success is no longer whether an individual uses AI—that will soon be a universal baseline. The true question is what remains of the individual once the labor of packaging is removed. If an employee’s thinking was always superior to their output, then AI represents the greatest career accelerator they have ever encountered. Conversely, if an employee’s output was always superior to their thinking, they face a difficult transition and have less time to adapt than they might imagine.
Every major technological shift acts as a re-sorting mechanism for the workforce. The assembly line rewarded punctuality and rote motion. The modern corporation rewarded administrative organization. The era of AI rewards original judgment and the relentless curiosity to keep feeding that judgment with new data. This transition should worry those whose professional identity was tied to the labor of packaging, but it should liberate those who have long been held back by it. For a century, the professional world sorted people by their ability to organize their thoughts; it is now beginning to sort them by whether those thoughts were worth organizing in the first place. Some are about to be found out, while others are finally about to be found.
