Aimphetamine
You're not more productive. You're dependent.
Artificial intelligence is the modern workplace’s performance drug. The comparison is not loose; it is alarmingly precise.
Within a single year, AI has spread through the knowledge economy with unprecedented speed. Employers do not merely permit its use; hoping for productivity gains, they fund it and institutionalize it, erecting digital leaderboards to celebrate employees who most effectively hollow themselves out for the company. Those who hesitate face a direct ultimatum: adapt, or be replaced by someone less encumbered by caution and doubt. Users report compressing days of cognitive drudgery into minutes—working longer, harder, and with a sudden, unearned confidence that allows them to perform beyond what their abilities and experience would allow.
The immediate dividends are visible. The long-term toll on the psyche is not. Heavy users already report a distinct mental exhaustion, a state described in the Harvard Business Review as “brain fry.” Alongside it, a more insidious pattern is emerging, now documented in research published in Nature: AI-induced deskilling. Physicians who had each performed over 2,000 colonoscopies saw their diagnostic accuracy drop by six percentage points within three months of using an AI tool—but only on days it was unavailable. The more one uses it, the less one can perform without it. Cognitive capacity, like muscle, requires resistance to develop—and like muscle, it atrophies without it.
For the young, the implications are grimmer. By eliminating the friction of difficult assignments, AI invites students to bypass the struggle that constitutes learning. You do not build strength by avoiding the weight and cheating on the reps.
Now the last exits are closing. As AI embeds itself not only in professional tools but in the search engines, communication platforms, and consumer software that mediate daily life, avoidance has become its own statement: AI-Minimalism. In today’s workplace, it is treated as an eccentricity at best, a career-killer at worst.
Now hold that picture and run it through a different frame. Suppose this were a pharmaceutical compound—spreading through the workforce, producing these documented effects, including dependency and atrophy. We would not allow employers to mandate it. We would not let it migrate unrestricted into children’s classrooms. We would demand restricted access for minors, strict guardrails for workplace use, and international agreements limiting its deployment before the harms compound. We would treat it, in short, as what it is: a powerful substance with real benefits and documented risks, requiring governance commensurate with its reach.
AI is that substance. The side effects are here. The regulation isn’t.



I'm a heavy AI user, so your pharma comparison lands because the side effects aren't theoretical in my own workflow. The guardrail that seems to matter is doing the hard version by hand often enough that your taste and error-sense stay alive. Where would you draw the line between using AI for leverage and outsourcing exactly the friction that keeps someone sharp?