What Did I Actually Do?
AI expands our capabilities while making the evidence of intelligence ambiguous. When tools produce sophisticated results, what do we know about our own competence—and what have we actually demonstrated?
By Vishnu Rajkumar · Marginalia
I find myself increasingly capable of producing things whose production does not necessarily make me feel more intelligent. I experience the satisfaction of accomplishment without the corresponding certainty of competence.
What troubles me isn't that AI contributed to the work. It is that I can no longer reliably distinguish my own intelligence from the intelligence I've learned to command. The result is mine in some meaningful sense, but its sophistication cannot be attributed entirely to me.
Perhaps the deepest disruption of AI is not that it has made intelligence abundant, but that it has made the evidence of intelligence ambiguous. I can now do more than I could before, while knowing less about what that increased capability says about me.
I'm making more while becoming less.