Preface
There’s a particular relief in being busy. At the end of every day, you rest knowing that you completed a particular set of tasks; your inbox clears up, your calendar crosses out completed events, you merge a pull request, etc.
But have you ever stopped and asked yourself what any of it actually was?
If you’re anything like me, you’ll have started to ask that question more and more as AI has become a permanent fixture of our lives.
This short essay is a philosophical discussion on what that means for us as individuals, and groups at large.
I.
To ask about the quiddity of an object is to ask about its whatness. Namely, after you strip away everything it happens to be and the implications it asserts, what is that object? A particular chair is brown and it has 4 legs, but none of these attributes actually define a chair’s chairness. The objective of asking about something’s quiddity is always, once its contingent features are removed, does there exist something essential, or does the thing dissolve into its properties with no core beneath them.
Before November 30, 2022, this question was never asked of work. I argue it’s because the attributes and accidents of work and its quiddity were fused. But AI has since driven a wedge between them. We can now suddenly accomplish a task without doing it, and the aforementioned question suddenly holds much more merit.
The claim this essay defends is that work’s essence was never the doing. The doing was just an accident of work, an erroneously time-consuming accident, but an accident nonetheless. What survives is not a residue or a remainder. It is what work essentially was, now visible because the AI removed the accidents that obscured it.
II.
To see the salient part of my claim, start with the ordinary conception and watch where it breaks. The ordinary conception identifies work with activity. With this view work is a set of performances: the writing, the calculating, the assembling, the corresponding. To work is to expend effort across time toward some output.
This framework is largely intuitive to us because for all of history this was the case. And it was the case because the execution of valuable tasks was scarce, difficult, slow, and requiring trained human labor, and so execution and value moved together. Valuable things were always harder to do and hard things were usually valuable to some degree.
Unfortunately, our first disservice to our future (aka today) was to mistake that correlation for an identity; to treat the doing and the value as one thing and call the union work. But correlation is not identity, and the test is whether the two come apart. The answer is that for all of history they could in principle, and now can in fact.
Consider what is wanted at the end of any piece of work. No one wants the report; they want the decision which the report informs. No one wants the contract drafted; they want the agreement it secures. The artifact is a means, and the doing is the cost of producing the means. The end–the decision, the agreement, the cured patient, the working bridge–was never the activity itself. This is not a modern observation. Aristotle distinguished praxis, action whose end is internal to itself, from poiesis, making whose end lies in the product. Most of what we call work is poiesis: its point is outside it. And if the point is outside the doing, then the doing is not the point.
III.
It’s at this point that you might identify an objection. The objection is that for some work the doing genuinely is the value, a type of work which is both praxis and poiesis, where the activity is not a mere means but is constitutive of the good produced. The craftsman at the bench, the surgeon’s hand, the writer in the sentence: surely here the execution is not separable cost but the very substance of the achievement. To strip the doing from these is to destroy the thing, not to reveal it, and so its quiddity is well defined.
However, I argue that there’s some conflation happening. Take the craftsman. What we admire is said to be the work of the hands, but watch closely and the hands are in service of something prior: a discrimination, repeated continuously, between the cut that is right and the cut that is wrong. The skill of the hand is the transmission of a judgment of the eye. Where the craftsman’s hand is steady but their judgment is poor, we get competent, lifeless work; technically executed, badly judged. Where the judgment is keen, we forgive a great deal of unsteadiness in the hand. The doing, even here, is the medium through which a judgment is expressed; it is not itself the judgment. What is rare and what is essential is the discrimination, which foreshadows the core argument I make in the next section.
IV.
Now that the claim is sufficiently supported, let’s move to its implications. Namely, if the doing is accidental, what is essential? I want to propose that work’s quiddity, once the executable accidents are removed, resolves into three core subjects, none of which is doing in the ordinary sense.
Before the work: the choosing of the end
The election of ends composes the first part of the essence of work. Execution selects among given options; determination invents the option and draws the frame the options will appear in. It can’t be specified in advance, because to specify an end is already to have determined it. And it answers to nothing above itself: when you execute an instruction, the instruction can absolve you, but the one who sets the end stands at the head of the chain with nothing upstream to absorb the error. This is the part of work that bears the most and conceals the least, which is exactly why we have so little appetite for it. Much of what we feel as busyness is the comfort of having ends handed to us.
Within the work: the discernment of its quality
Knowing, when set before a result, whether it is good. One expects that if execution is cheap, the value moves to specification: write the brief well enough and the right thing comes out. But no specification is complete. Any description detailed enough to follow still needs interpretation, which needs further description, without end. We missed it as a separate thing only because, doing our own work, we exercised it silently and called the whole event skill. Take the doing away and it stands alone, the thing the skill was for.
After the work: the bearing of the answer
The assumption of responsibility for it — standing behind what was made, vouching for it before those it touches. This one differs in kind. The other two are capacities: improvable, in principle imitable. This is not a capacity but a relation — the standing of a self that can be held to account, that bears the cost if the work fails. A machine may one day match us at choosing ends and judging quality. It cannot match us here, not for want of skill but because it has nothing at stake.
V.
These three acts share a feature that explains why they were so long concealed, and it is the crux of why we ask the aforementioned question.
None of them has duration. The determination of an end can happen in an instant. The recognition of something as good or bad is nearly immediate. The act of standing behind a piece of work takes no time at all — it is complete the moment one commits. They are, in a word, durationless, and this is precisely the property that made them invisible in the old arrangement.
A job, in the economic and social sense, had to fill time. It had to occupy a day, a career, a life. The essential acts of work, being durationless, could not by themselves constitute a job; there is nothing to fill with an instantaneous judgment. So they were necessarily embedded in the doing, which supplied the duration the judgments lacked. The hours of execution gave judgment a body to inhabit and a surface to show. We only measured the hours because the hours were measurable, and we came to believe the hours were the work, when the hours were only obscuring the work.
VI.
When the accidents are stripped away, does a thing dissolve, or does an essence remain? Well… it depends.
As we’ve shown, for work, an essence remains. But the more honest question is why we hid it so well, and why its uncovering feels more like a loss than a discovery.
Here is the answer: we did not bury work’s essence by accident. The burying was the point. The three acts that make up the real work are also the three that expose us most: to choose is to have no one to blame; to judge quality is to risk being the one with bad taste; to answer for the work is to be the one a grievance lands on. In probably one of the most ironic ways, the reason we felt relief doing work was to not have to answer for it. And busyness was the cover. The endless executable doing (the inbox, the tickets, the meetings) was never just how the work got made. It was where we went to avoid the part of the work that exposes us.
This is exactly why AI does not feel like relief, whatever it promises. By taking the doing, it takes the hiding place. It returns us to the exact thing the doing let us avoid: the choice with no cover, the judgment with no rule, the answer no one else will bear.
This is not only a private reckoning. Whole organizations were built on the doing — headcount measured in hours, value inferred from activity, the busy ones presumed the useful ones. When the doing is cheap, those proxies stop working, and a team is left facing the same exposure an individual is: what was actually worth doing, was it any good, and who stands behind it. The hiding place was load-bearing for institutions too, and it is being pulled out from under all of us at once.
I’ll admit the personal version of this took me longer to see than the argument did. Writing this, I realized how much of my own sense of being useful had quietly become a sense of being busy — and that the two had almost nothing to do with each other. The work was never the busywork. It was the part I’d been using the busywork to avoid.
If you made it this far and liked it, come say hi at adam.cm, or email me at adamghaida [at] cmu [dot] edu !
comments