Why I'm Writing a Book About Identity in the Age of AI

The Question

Here is something that happens more often than I'd like: Someone asks me what I do. I say I'm a technical writer. Or I say I work in developer experience. Or I say I build AI documentation systems. And in each case, the response is some version of: "Isn't AI going to do that?" The question used to annoy me. Now it interests me. Not because the answer has changed, but because the question reveals something the asker doesn't realize they're asking. They're not really asking whether AI can do my job. They're asking whether the things I define myself by are still mine. Whether the skills I spent years building still mean anything when a model can approximate them in seconds. Whether "I'm a writer" means the same thing it meant five years ago. That's not a question about automation. That's a question about identity. And almost nobody is treating it that way.

How I Got Here

I've spent six years building an identity around going deep. It started in technical writing at Weesho Lapara, where I wrote documentation for clients in Australia from an apartment in Kathmandu. Then ComplianceQuest, where I wrote compliance-grade docs for Salesforce products and quietly replaced a workflow of zipped Word files with actual version control. Then Logpoint, where I covered six product areas (SIEM, SOAR, UEBA, Director, Integrations, Security Research), not because anyone assigned them to me, but because I was the person who stepped in when things needed doing. Along the way, I got an MSc in Data Science. My thesis was on insider threat detection: four unsupervised anomaly detection models on the CMU-CERT dataset, best AUC-ROC 0.799. I learned Python, built evaluation pipelines, spent months thinking about how to detect malicious behavior in organizational data. The work was rigorous and isolating and deeply satisfying. Then I joined EkLine, and everything shifted. EkLine builds AI tools for documentation. My job (first as a consultant, then as Customer Success and Growth Manager) involved designing AI agent scaffolding, running evaluations across 90 Docker-isolated benchmark runs, and figuring out what to encode into a system versus what to let the model infer. In other words: I spent years building the skills that AI is now learning to approximate. And then I spent a year building the AI that approximates them. That's a strange position to be in. Not threatening, exactly. But clarifying. When you build the thing that does what you used to do, you learn very quickly which parts of "what you do" were actually you, and which parts were just the task.

The Pattern Collector's Dilemma

I don't have a single expertise. I have a mode of thinking. Technical writing. Security products. Data science. AI systems. Product marketing. Customer success. Developer experience. These aren't chapters of a career. They're data points in a pattern I'm still figuring out. The thread connecting them is this: I go into a domain, I learn its structure one layer deeper than the role requires, and I make that understanding legible. The domain changes. The mode doesn't.

This is what makes the identity question so personal for me. If my value were tied to a specific skill (writing docs, say, or running evaluations), then yes, AI threatens that value directly. Models can write docs. Models can run evaluations. The skill is automatable. But the mode of thinking isn't a skill. It's a perspective. It's the ability to see structural similarities between domains that look nothing alike on the surface. It's knowing which questions to ask, not because you've seen the answer, but because you've been wrong in enough adjacent fields to recognize the shape of the problem. Can AI do that? I genuinely don't know. But I know that the conversation about AI and jobs almost never gets to this layer. It stays at the surface: which tasks can be automated? Which roles will disappear? How many jobs will be lost? Those questions matter. But they're not the deepest question. The deepest question is about what happens to the people whose sense of self is woven into the work that's being automated. Not their income. Their identity.

What the Book Is About

The book is about identity displacement, not job displacement. Job displacement is an economic problem. It has economic solutions: retraining, UBI, new industries, new roles. We can argue about the details, but the framework is understood. People lose jobs, people find new jobs, the economy adjusts. It happened with manufacturing. It happened with agriculture. It'll happen with knowledge work. Identity displacement is a different kind of problem, and it doesn't have economic solutions. When a factory worker loses their job to a robot, they lose their income. That's a material problem with material solutions. But when a writer loses their writing to an AI, when the thing they spent decades getting good at becomes trivially reproducible, what they lose is harder to name. It's the sense that their skill meant something. That the years of practice had a purpose beyond the output. That being good at this thing made them a particular kind of person. That loss doesn't show up in employment statistics. It shows up in a quiet erosion of meaning.

The book explores this through several lenses: Skill as identity. We define ourselves through what we can do. "I'm a writer." "I'm a developer." "I'm a musician." When the thing we can do becomes automatable, the statement doesn't just lose economic value. It loses ontological weight. What does "I'm a writer" mean when everyone is? The taste hypothesis. If every skill becomes a commodity, what's left? I think the answer is taste: the judgment about which skills to apply, when, and why. Taste is formed through friction, failure, and accumulated experience. It can't be automated because it's not a capability. It's a disposition. It's the residue of having been wrong in useful ways. Expertise without credentials. The autodidact's dilemma. When I taught myself Linux to write better documentation, no one credentialed that. When I learned prompt engineering by building evaluation frameworks, no certificate appeared. The expertise is real. The legibility is low. AI tools lower the expertise bar while raising the legibility bar. That creates a specific kind of identity crisis for people who define themselves by depth rather than credentials. The observer problem. Here's the part that concerns me most: the changes to identity happen slowly enough that we might not notice them. A writer who gradually relies more on AI suggestions doesn't wake up one day and think "I've stopped being a writer." The shift is incremental. The skills atrophy quietly. The sense of self adjusts. By the time the displacement is visible, it's already complete.

Why in Public

I could write this book privately. Spend a year in a room, emerge with a manuscript, submit it to publishers. That's the traditional approach, and there's nothing wrong with it. But I'm choosing to write in public for two reasons. First, because the ideas need friction. I don't want to develop these arguments in isolation and discover they're wrong only after publishing. I want to write essays, get pushback, refine my thinking, and let the book emerge from the conversation. The notes on this site are part of the book. The blog posts are part of the book. The thinking is happening in real time, and you're watching it. Second, because writing in public is itself an act of identity. It's saying: this is what I think, this is how I think it, and I'm willing to be wrong about it publicly. In a world where AI can generate convincing prose on any topic, the thing that makes writing matter is the person behind it. Their perspective. Their specific history of being wrong. Their willingness to put their name on an idea before it's fully formed. This site is not a portfolio in the traditional sense. It's a demonstration of a mode of thinking. The essays are evidence that a particular human, with a particular history and a particular pattern of failures and a particular way of seeing, is doing the thinking. That's what the book is about. And that's why it can't be written by an AI.

The Question Behind the Question

When someone asks "Isn't AI going to do that?" about my work, there's a question behind their question. They're really asking: what makes you you if the things you do can be done by a machine? I used to not have an answer for that. Now I think the answer is something like this: I am not the things I produce. I am the perspective that shapes what I choose to produce, and why, and for whom. That perspective was formed by six years of going deeper than the role required, across domains that look nothing alike, accumulating a library of patterns and failures that no model has access to. Because they didn't happen to a model. They happened to me. The skills are reproducible. The perspective isn't.

The question isn't whether AI will change what we can do. It's whether we'll notice what changes in who we are.

That's what the book is about. I'm writing it because I want to notice.