Something landed differently this week.

GPT-5.4 Thinking launched on Monday, fusing reasoning, coding, and agentic workflows into a single model. Two days later, Anthropic announced plug-and-play AI agents that departments can install like software. The Wall Street Journal ran a profile of “Stoic CEOs” navigating the transition. Tim Cook and Satya Nadella both cited Stoic principles, which would be encouraging if it didn’t feel like reading about how the people holding the levers are coping while the levers get pulled.

If you’re a writer, a designer, a developer, a strategist, a therapist, a teacher (if your sense of self is tangled up with what you create or what you know), this week probably hit a nerve. Not the abstract “AI is changing everything” nerve you’ve been managing for two years. The specific, sharper nerve. The one that asks: what am I for, if not this?

That’s not a productivity question. That’s a philosophy question. And it’s old.

The Quick Version

Epictetus divided everything in the universe into two categories: what’s up to us and what isn’t. AI development, automation trajectories, which jobs get displaced: firmly in the second category. Your character, your contribution, your relationship to your own creativity: firmly in the first. The dichotomy of control isn’t a cope. It’s a precise instrument for separating the terror from the actual problem, so you can work on the problem.

The Particular Sting of This Kind of Threat

Job loss has always been frightening. But previous waves of automation targeted physical repetition (factory lines, data entry, routine processing). The current wave is different in a way that deserves honest acknowledgment, because pretending it isn’t different doesn’t help anyone.

What agentic AI systems do well is reason, generate, synthesize, and produce. These are the things that many knowledge workers and creatives built their identities around. Not just their income. Their sense of being someone particular, with a specific mind, capable of something that mattered.

A 2026 Springer Nature paper on Stoicism and AI singularity risk made a point that’s been circulating in philosophy circles since: the threat isn’t only economic displacement but what researchers call epistemic redundancy: the feeling that your judgment, creativity, and knowledge are no longer rare or necessary. That the thing you thought was essentially you turns out to be replicable.

That stings differently than “the factory moved overseas.” It touches something closer to the center.

The Stoics were, as it happens, specifically interested in that center. They called it hegemonikon (the ruling faculty, the seat of reason and judgment and character), and they had very clear things to say about what can and can’t touch it.

What Epictetus Actually Said

Epictetus was a freed slave who taught philosophy in a converted house in Nicopolis. He never wrote anything. His student Arrian took notes, which became the Discourses and the short Enchiridion. His life had included no creative autonomy, no career to speak of, no professional identity to protect. He developed his philosophy from that ground.

His central claim, stated in the first paragraph of the Enchiridion, is this:

“Some things are in our control and others not. Things in our control are opinion, pursuit, desire, aversion, and, in a word, whatever are our own actions. Things not in our control are body, reputation, command, and, in a word, whatever are not our own actions.”

Notice what he puts in the “not in your control” column: reputation, command, body. Things you might have staked your identity on. Things that can be taken.

The Stoics weren’t naive about what this costs. Epictetus isn’t saying “therefore nothing matters.” He’s saying: if your purpose is lodged inside the things that can be taken, your purpose is contingent. It will be threatened every time circumstances change, which they always do. Real stability requires anchoring purpose somewhere else.

Applied to 2026: if your sense of meaning lives primarily in being the one who writes the copy, codes the feature, designs the system, you’ve anchored it in a “not up to us” thing. The skills themselves can be supplemented or replaced. What can’t be replaced is your character, your judgment about what to make and why, your relationship to your own creative instinct.

That’s not consolation. It’s a reorientation.

The Dichotomy of Control as a Diagnostic Tool

Most people use the dichotomy of control as a coping mechanism, a way to feel better about things they can’t change. Epictetus intended something more active.

He described it as a melete — a practice, a discipline. Something you apply repeatedly until it becomes your default mode of processing experience. The goal isn’t to shrug at AI displacement. It’s to get precise about where your energy actually belongs.

Try this, and be specific about it.

Write down what you’re actually afraid of. Not a general “AI taking over.” Something specific: I’m afraid the content strategy work I’ve spent eight years learning will be automated, and with it my professional identity and income. That specificity matters.

Now sort it.

What’s in your control: The quality of your judgment about what’s worth making. Your relationships with the people you work with and for. Whether you develop skills in directing and evaluating AI output versus only producing output yourself. How you respond when a client says they’re switching to AI tools. Whether you let fear make you brittle or push you toward figuring out what your actual value is.

What isn’t: Whether GPT-6 arrives in six months. Whether the market for your specific skill set shrinks. What your industry does or doesn’t do with these tools. What Tim Cook decides to deploy.

The second list is not yours to solve. The first list is. And the first list is longer than it usually feels when you’re inside the anxiety.

Identity Attached to Output

Here’s the specific philosophical problem that AI displacement surfaces for knowledge workers. It’s worth pausing on rather than rushing past.

Most of us have built our sense of self partly around what we produce. Writers feel like writers when they write. Developers feel like developers when they build things. Designers feel like designers when they solve visual problems. This isn’t neurotic. It’s natural. Humans have always found meaning in craft and creation.

But the Stoics would flag something here. They distinguished between what you do and what you are. Between the external expression of character and the character itself. Marcus Aurelius wrote to himself: “You have power over your mind, not outside events.” He wasn’t discounting the events. He was tracking what they actually reach.

If your identity is “I am someone who writes” and AI can now write, you face an existential problem. If your identity is “I am someone with judgment, values, specific experience of being human in this particular body with these particular relationships, who has chosen to express that through writing,” you face a different problem. A more tractable one.

The Stoics weren’t anti-craft. Seneca, who was a brilliant prose stylist, clearly took pleasure in the quality of his sentences. Marcus was a committed student of philosophy for its own sake. Epictetus taught with obvious passion. The point isn’t to become detached from your work. It’s to locate your identity at a level deep enough that what happens to the work doesn’t hollow you out.

What’s at that level? The Stoics called it virtue — not in the moralistic sense, but in the Greek arete: the excellence of a thing at being what it is. For humans, that means the exercise of reason, the cultivation of character, and the orientation toward contribution, in ways that no labor market shift touches.

The Springer Nature Paper Worth Reading

In February 2026, Springer Nature published a peer-reviewed paper on Stoicism as a philosophical resource for navigating AI singularity risk. The paper, by researchers in applied philosophy and AI ethics, argues something that sounds radical but is actually quite Stoic: the existential threat of AI isn’t primarily economic but epistemological.

What gets threatened isn’t just your job. It’s your episteme, your status as a knower, a reasoner, a contributor of genuine insight.

The paper draws on Epictetus’s concept of prohairesis (the faculty of choice, the capacity to assent or dissent to impressions, to form judgments about what’s worth pursuing). The researchers argue this faculty is precisely what AI cannot replicate: not creativity in the generic sense, but the specific act of a particular person, in a particular life, choosing what matters and why, taking responsibility for those choices.

That’s not just philosophical comfort. It’s pointing at something real. The agentic AI systems launched this week are extraordinarily capable at producing output given direction. They remain dependent on someone providing the direction: the judgment about what to make, for whom, toward what end, within what constraints. That judgment belongs to the person whose life is at stake.

Three Practices for Right Now

These aren’t exercises for “when you’ve stabilized.” They’re for this week, which is when it matters.

The Impression Audit

When the dread arrives (when you read about GPT-5.4 or see your colleague’s anxiety on a call), Epictetus would say: pause before assenting to the impression. The impression might be “this ends me” or “I’m becoming obsolete.” Before accepting it as true, examine it.

Concretely: write it down. “The impression is: my value is disappearing.” Then ask what’s actually being claimed. Is it that your income might shift? That some specific skills are becoming less scarce? That your industry is reorganizing? These are different claims with different practical implications. The diffuse terror tends to collapse multiple distinct problems into one overwhelming one. Separating them makes each one workable.

The “What Remains” Question

This one is harder and more useful. Ask yourself: If I couldn’t do the specific work I do (not for a week but for good), what would remain of my sense of who I am?

This isn’t a thought experiment about career transition. It’s a diagnostic about where your identity is actually anchored. If the answer is “not much,” that’s information. Not a verdict, but information. The Stoics would say this is the precise moment to do the work of building identity at a deeper level. Not because disaster is coming, but because contingent identity is always fragile, AI or no AI.

If the answer surprises you, if there’s more there than you expected, that’s equally useful. It means you already have the resource. You just haven’t been drawing on it.

The Contribution Frame

Epictetus identified three disciplines for the Stoic life: the discipline of desire, the discipline of action, and the discipline of assent. The middle one he described specifically as doing your work in relation to others, for the community of rational beings, not for recognition or outcome.

Applied here: ask not “is my work still valuable in the market?” but “am I still contributing something real to actual people I care about?” These questions have different answers, and the second one tends to be more stable.

A developer who frames their work as “I build things people need” has a more durable sense of purpose than one who frames it as “I am a developer.” The first framing survives market shifts. It even survives AI, because people still need things, and the judgment about what to build remains human even when the building gets automated.

What the “Stoic CEOs” Are Actually Doing

The Wall Street Journal profile of Cook and Nadella is worth thinking about critically, not just as validation that Stoicism is in vogue. What they’re doing (what the article credits to Stoic practice) is maintaining a quality of decision-making under uncertainty that doesn’t depend on having certainty.

That’s genuine Stoic practice. The dichotomy of control applied at scale. But there’s an asymmetry that deserves naming: they’re deploying these tools. Most readers aren’t in that position.

The Stoic framework doesn’t change based on whether you’re directing automation or being displaced by it. Epictetus was enslaved. He didn’t have executive leverage. His philosophy wasn’t for people with power. It was for people without it. The dichotomy of control is, if anything, more relevant when you’re not the one making the decisions about which jobs get restructured.

Amor fati — the Stoic concept of accepting and even embracing what comes — isn’t about liking what’s happening. It’s about not wasting your inner life fighting the fact that it’s happening, so you have energy left to respond to what you can actually affect.

The Honest Limits of This Framework

The Stoic framework doesn’t pay your rent. If your income is genuinely threatened, the answer is practical action: upskilling, diversifying, building relationships, exploring how to position yourself in an AI-assisted economy rather than against it. The philosophy addresses the psychological layer on top of that problem, not the problem itself.

And the philosophical tools work better when the material situation isn’t in genuine crisis. If you’re facing immediate income loss, these practices are most useful for the second-order questions: how to avoid making the situation worse through fear-driven decisions, how to maintain your character under pressure, how to stay oriented toward what actually matters while you navigate the immediate practical problem.

There’s also the community piece, which the Stoics consistently emphasized. Seneca had Lucilius. Marcus had his philosophy teachers. The anxiety around AI displacement isn’t only individual. It’s shared across a generation of knowledge workers sitting with the same uncertainty. Finding people thinking through these questions honestly, rather than either dismissing the concerns or catastrophizing, makes the thinking better.

The earlier post on Stoic wisdom for general AI anxiety covers the three core Stoic practices in more depth. And for the foundational question of where purpose lives when it’s unhooked from career and productivity, the post on Stoic purpose beyond career is the best starting point, especially the section on telos, which is exactly what’s being threatened here.

What Doesn’t Change

Epictetus made a list. He put property and reputation in the “not in our control” column. He put prohairesis (the faculty of choice, the capacity for judgment, the ability to decide what matters) in the other column.

AI doesn’t change that list. The faculty of choice, the specific experience of being this person in this life making these judgments: that remains yours. It’s not a consolation prize. The Stoics thought it was the whole game.

This week’s launches are significant. The economy of creativity and knowledge work is genuinely shifting. Some of what took years to learn will take months to automate.

And: your character isn’t in that column. Your care for the people you work with isn’t in that column. Your judgment about what’s worth making and why (the specific thing that comes from your particular life) isn’t in that column either.

What’s in your control has always been smaller than you’d like and more important than it seems.

Start there.

---

If the existential weight is genuine rather than manageable anxiety, please consider talking with a therapist. Philosophy can clarify what’s at stake. It isn’t a substitute for support when the ground is actually shifting under you.