AI and the Götterdämmerung of Work

Work is dead. And we have killed it.

Not the robots. Not some external force that arrived and took something from us. We built the systems. We wrote the playbooks. We encoded the judgment. We designed the agents, defined the processes, documented the methodology — and now the systems don’t need us. We are standing in the wreckage of our own creation, and we haven’t yet grasped what that means.

Nietzsche’s madman didn’t announce God’s death as a victory. He announced it as a catastrophe — not because God was good, but because God was load-bearing. The entire structure of values, meaning, and orientation that organized Western civilization depended on that foundation. Remove it and you don’t get liberation. You get vertigo. You get the question: now what do we stand on?

The myth that organized work was simpler than God but just as load-bearing: human value comes from reliable output. Show up. Execute. Produce. Hit the metrics. Be consistent. Be legible. Be useful in a way that can be measured and rewarded. That myth organized careers, institutions, identities, retirements. It told people what they were for.

AI didn’t defeat that myth. It exposed it. And we’re the ones who handed AI the evidence.

The fleet

I run a fleet of AI agents. Not a single chatbot. A system. Scouts that scan codebases for improvements, triage agents that sort and prioritize issues, grinders that execute task after task without coffee breaks or existential crises. Every month, the fleet gets faster and more consistent. And every month, it gets faster and more consistent for the same reason: I removed a human step.

Auto-triage replaced my morning review. Auto-scout replaced my weekly audit. Auto-grind replaced the implementation work I used to do between meetings. Each removal made the system better. Not marginally. Measurably.

The human gates that remain — voice, strategy, taste — feel essential today. I’m the one who decides what the fleet should care about, who judges whether the output meets the bar, who provides the vision that holds the whole thing together. Those feel like irreducible human contributions. They feel permanent.

But each one is a bet. A bet that the playbooks, the documents, the accumulated methodology won’t eventually encode what I know. And I’m the one writing the playbooks.

Leibniz’s dangerous idea

Gottfried Wilhelm Leibniz, writing in 1714, proposed that the universe is composed of monads — simple substances, each containing within itself everything it needs. No monad interacts with any other. They don’t exchange information or influence each other directly. Yet they all behave in perfect harmony.

How? Because God, in his infinite wisdom, designed each monad’s internal program so that its autonomous behavior would perfectly coordinate with every other monad’s autonomous behavior. Leibniz called this pre-established harmony. God doesn’t intervene. He doesn’t need to. The design is sufficient. The system runs itself.

I think about this every time I deploy a new agent.

The monads were designed by God to not need God’s ongoing intervention. We are designing agents that don’t need our ongoing intervention. The harmony isn’t miraculous — it’s methodological. We write the documents, define the processes, encode the judgment. And if the methodology is good enough, the system coordinates without us.

Leibniz’s God was the ultimate architect of self-obsolescence. He built something so complete that his continued involvement would have been a design flaw.

The machine-self comes home

In The Work of Being, I argued that the industrial model of work taught people to suppress the parts of themselves that are human — curiosity, imagination, intuition, judgment, creativity — in exchange for security and a paycheck. You showed up, executed procedures, processed information, produced outputs, followed rules, hit metrics. You became what I called the machine-self: a predictable, compliant, programmable version of yourself that the organization could slot into its processes.

AI exposed that bargain. If a machine can do the work you do when you act like a machine, then the machine-self you built over your career is already obsolete. That was the book’s central provocation: the version of yourself you performed at work was never really you, and now an actual machine can perform it better.

But here’s what I didn’t say clearly enough: the Götterdämmerung isn’t just about the machine-self dying. It’s about the entire apparatus that created and sustained it. The performance reviews, the job descriptions, the org charts, the competency frameworks, the career ladders — the whole Valhalla of institutional work. These were the contracts the myth required. And like all contracts built on a dead foundation, they’re burning.

Hannah Arendt distinguished between labor (what we do to survive), work (what we do to build a durable world), and action (what we do to disclose who we are). The industrial model collapsed all three into labor. Show up. Produce. Repeat. The machine-self was perfectly adapted to this collapse. It could labor endlessly without ever needing to act, without ever having to be someone.

AI doesn’t replace action. It can’t. Action requires what Arendt called natality — the capacity to begin something genuinely new, something that couldn’t be predicted from what came before. But AI does replace labor almost entirely. And it replaces a great deal of work. What’s left is the part that was always most human and always most neglected.

The paradox of creation

Here’s where it gets uncomfortable.

If the system you built to extend your capabilities eventually doesn’t need you, was building it an act of creation or an act of self-obsolescence?

Both. And I’m not sure there’s a difference.

Every parent knows this. You spend years building a human being’s capabilities — teaching them to walk, to read, to think, to choose. The entire project is aimed at making yourself unnecessary. A parent who remains necessary has failed. The child who never stops needing you never becomes themselves.

The more my fleet handles, the more I’m confronted with the question: what am I actually for? Not what tasks do I perform — the fleet is eating those. But what is the nature of my contribution that no system can replicate?

The answer keeps coming back to the same place. Commitment. Judgment under genuine uncertainty. The willingness to say this is mine and I stand behind it. The things I described in the book as the work of being — discernment, courage, self-authorship, formation. The things the industrial model trained out of us and that AI now makes essential.

The revaluation

Nietzsche didn’t stop at the death of God. That was only the diagnosis. The prescription — the harder, longer project — was what he called the revaluation of all values. Not replacing the dead myth with another myth. But doing the philosophical work of building a new table of values from the ground up. Values that could stand without the old foundation.

That’s what’s actually being demanded of us now.

The death of the reliable-output myth doesn’t leave a vacuum. It leaves a question: what is human work actually for, when it isn’t for producing outputs that machines can produce better? The answer isn’t comfortable and it isn’t quick, but it’s been visible all along underneath the machinery:

Work is for action. For natality. For the disclosure of who you are through what you choose to begin. For judgment that can’t be encoded because it depends on having a self that stands behind it. For the kind of commitment that gives output its meaning — not its accuracy, not its efficiency, but its significance.

The fleet will keep getting better. I’ll keep removing human steps. And the steps that remain will keep getting more essentially human — until the question won’t be “what can the AI do?” but “what was I doing all along that I thought was work but was actually just the costume?”

The machine-self was never you. The myth was never permanent. And the capacities it suppressed — judgment, discernment, the courage to begin — were always underneath, waiting not to be built but to be returned.

Work is dead. And we have killed it. The question Nietzsche’s madman was really asking isn’t whether we’re ready to mourn. It’s whether we’re ready to build what comes next.

We are the murderers. That means we’re also the architects.