the developmental logic applies to learning a skill. it does not apply to a 34-year-old trying to get work done.
there's a specific genre of post that's become common over the last year. it takes different forms, but the core argument is usually some version of: "stop using ai for this, it's making you worse at it." sometimes it's about writing. sometimes it's about coding. sometimes it's a vaguer claim about "thinking" or "creativity." the post usually goes viral in a certain demographic, generates a lot of agreement, and changes almost nothing about how people actually behave.
i've read dozens of these posts. i think most of them are wrong, and the ones that are right are wrong about why.
the structure is consistent. using ai for task x prevents you from developing the skill required to do task x. writing your own emails makes you better at communication. solving your own code problems makes you a better programmer. doing your own research makes you better at thinking. therefore, using ai for these things is making you worse.
this argument sounds right. it has an intuitive logic. it's the same logic behind "don't use a calculator until you understand the math." and it's correct in the specific, limited context it usually applies to — learning. if you're a student and you're trying to develop a skill, having ai do it for you does slow that development.
but most people being told to stop using ai are not students trying to develop a skill. they're adults who have already developed skills and are now trying to work.
the people posting these arguments are usually conflating two different situations: learning and working. the developmental logic applies to learning. it does not apply to a 34-year-old project manager who already knows how to write emails and is using ai to write them faster so they can spend more time on the parts of their job that actually require their judgment.
"telling someone who can already drive that using gps is making them worse at navigation is technically true and practically irrelevant."
the question is never "can this person do this without ai?" the question is "what is the best use of this person's time and cognitive resources?" if the answer to the second question is "write a meeting recap manually," something has gone wrong in the analysis.
there is a real concern underneath the bad argument. it's not about individuals making tool choices for their own work. it's about what happens to institutions and systems when certain cognitive skills stop being practiced at the population level.
if ai does enough of the writing, the editing, and the synthesis for enough people, there's a plausible future where the number of humans who can evaluate the quality of a piece of writing shrinks significantly. that's a different problem than "this individual is getting worse at writing." it's a structural problem about maintaining the evaluative capacity to know whether the tools are producing something good.
that's worth thinking about. but it's a very different argument than "you should write your own emails." it's an argument about education, institutional investment in human capability, and what skills we decide matter enough to maintain deliberately.
i think most of the "stop using ai" posts are less about the tools and more about a specific kind of status anxiety. the people writing them are often people for whom a particular skill is central to their identity — writing, coding, research. ai makes that skill more accessible, which changes what it means to have it. a skill that used to differentiate you no longer does. the post is a way of defending the value of the skill by arguing that the shortcut is cheating.
that's understandable. it's also not a good basis for advice. the tool exists. the people who learn to work with it will move differently than the people who don't, for reasons that have nothing to do with whether the skill was "earned."