The morning after David's layoff, I made a decision. I would work without AI for one day. Just one day, to prove to myself that I still could. The rationalizations from before felt hollow now, exposed by David's departure, by the cardboard box in his arms, by the empty desk where he used to sit. If I was truly in control, truly just "using a tool," then I should be able to set it aside. I should be able to function without it. I sat at my desk, took a deep breath, and closed the CodeForge tab. The interface disappeared, leaving only my blank editor. The cursor blinked in the empty space, patient, waiting. I can do this, I told myself. I've been coding for sixteen years. AI has only been part of my life for a few months. The skills are still there. They have to be. I cracked my knuckles, placed my fingers on the keys, and began to type. --- The first hour wasn't so bad. I started with a simple task, a data transformation function that should have taken me twenty minutes. I remembered the general approach: parse the input, apply the transformation rules, format the output. The pattern was familiar, one I'd implemented dozens of times over my career. I typed the function signature. Then the input validation. Then the main transformation logic. Each line came slowly, like pulling words from a deep well, but they came. I was doing it. I was thinking. I was solving the problem myself. See? I thought, a flicker of hope kindling in my chest. I can still do this. The dependency isn't that deep. I just got out of practice, that's all. By 10 AM, I had a working implementation. It wasn't elegant, the code was verbose, the logic more convoluted than it needed to be, but it worked. The tests passed. The function did what it was supposed to do. I allowed myself a small smile. Maybe I'd overreacted. Maybe the blankness from the other day had been stress, or fatigue, or just a bad morning. Maybe I was still the engineer I'd always been. Then I opened the next ticket. --- The second task was harder. It was an API integration, connecting our service to a third-party provider. I'd done similar integrations before, probably a hundred times. The pattern was standard: read the documentation, implement the authentication, build the request handlers, add error handling. But as I started working, something felt different. I opened the documentation and began reading. The words were there, the information was clear, but it wasn't... sticking. I'd read a paragraph and realize I hadn't absorbed any of it. I'd look at a code example and find myself unable to translate it into my own implementation. Focus, I told myself. Just focus. You've done this a thousand times. I started typing. A line. Then another. Then I stopped, unsure where to go next. The pattern that should have been obvious, the authentication flow, the token refresh, the retry logic, felt just out of reach, like a word on the tip of my tongue that refused to come. I stared at the screen. The cursor blinked. My mind was... quiet. Not the productive quiet of concentration, but the empty quiet of nothing happening. This is ridiculous, I thought, frustration building. I know this. I've done this before. I forced myself to continue. I typed more lines, deleted them, typed different lines, deleted those too. Nothing felt right. Nothing clicked into place the way code used to, when the solution would emerge from the problem like a shape from fog. An hour passed. Then two. I was still on the same function, still stuck, still unable to make my mind work the way it used to. By noon, I was starting to panic. I'd been working for four hours and had almost nothing to show for it. The API integration was a mess of half-written functions and commented-out attempts. My mind felt sluggish, like trying to run through water, every thought requiring enormous effort for minimal progress. What's wrong with me? The question echoed in the void where my thoughts should have been. I knew this work. I'd done it before. I had sixteen years of experience, thousands of solved problems, a career built on the ability to think through complexity. But that person felt like a stranger now. A memory of someone I used to be. I pushed back from my desk, ran my hands through my hair, tried to clear my head. Maybe I needed a break. Maybe I was overthinking it. Maybe if I just started fresh, the thoughts would come. I walked to the break room, poured myself coffee, stared out the window at the parking lot below. My colleagues were going about their day, typing at their keyboards, having conversations, solving problems. They looked normal. They looked capable. What did I look like? I returned to my desk and tried again. The cursor still blinked. The problem still sat there, unsolved. My mind still felt empty. Two hours. I'd been staring at the same blank editor for two hours. The problem was simple, I knew it was simple, but my mind was empty. Not slow. Not struggling. Empty. Like a computer that had been wiped, all the programs deleted, nothing left but the blank screen and the blinking cursor. I tried to think through the authentication flow. First, you get the token. Then you include it in the header. Then you... But the thoughts wouldn't form. They were there, somewhere, in the abstract, I knew what authentication was, what tokens were, how HTTP headers worked, but I couldn't translate that knowledge into code. I couldn't make the connection between the concept and the implementation. This is what David was talking about, I thought, the realization hitting me like a physical blow. This is what he was trying to warn me about. The thinking isn't the obstacle. The thinking IS the point. But I'd forgotten how to think. Or rather, I'd delegated it. Every problem for the past few months had been solved by CodeForge. Every decision had been made by the AI. Every line of code had been generated by a prompt. I'd become the middleman, the translator, the prompt engineer. And somewhere along the way, I'd lost the ability to do the thing I used to do best. The panic was real now, a tightness in my chest, a buzzing in my ears. I couldn't solve a simple API integration. I couldn't think through a basic problem. I was thirty-five years old, a senior engineer with a decade and a half of experience, and I couldn't write code without an AI holding my hand. What have I done to myself? At 2 PM, I gave up. My hands moved on their own, opening CodeForge before my conscious mind could stop them. The familiar interface loaded, the cursor blinking with patient anticipation, and I felt something break inside. I need this, I thought, the words tasting like ash. I can't do it without this. I typed the prompt: "Implement API integration for third-party provider. Include authentication, error handling, and retry logic." The solution appeared on my screen, elegant and complete, generated in thirty seconds what I'd failed to produce in four hours. I stared at it, at the proof of my dependency, and felt nothing but shame. I'd tried. I'd really tried. And I'd failed. Not because the problem was hard, it wasn't, but because I couldn't think anymore. Somewhere in the past few months, I'd lost the ability to solve problems on my own. The AI had become my mind, and I'd let it. I'd welcomed it. I'd praised myself for it. I copied the generated code into my editor, made the usual minor adjustments, and committed it. The ticket was closed. The metrics would be good. Jennifer would be pleased. But I felt nothing but the hollow ache of a man who'd given away something essential without even realizing it. That evening, I sat at my desk long after everyone else had gone home. The CodeForge interface was still open, the cursor still blinking, waiting for my next prompt. It had been waiting all day, through my failed attempt at independence, through my surrender, through the shame that still sat heavy in my chest. What am I without AI? The question I'd been avoiding for weeks now demanded an answer. And the answer was terrifying: I didn't know. I couldn't remember the last time I'd solved a problem without AI assistance. I couldn't remember what it felt like to think through something from scratch, to feel my mind expand around a challenge, to experience the satisfaction of figuring something out. I was a prompt engineer now. A translator between humans and machines. A shell of the engineer I used to be. And the worst part? I'd chosen this. I'd embraced it. I'd told myself it was efficiency, it was progress, it was the future. I'd judged David for his resistance, felt superior in my adaptation, validated by the metrics and the praise. But David could still think. David could still solve problems. David had something I'd traded away for productivity and position and the comfortable, seductive ease of not having to think. I closed my laptop and sat in the darkness, the question still echoing in the void where my thoughts used to be. What have I become? Word Count: 1,726
After the failed attempt, I returned to AI-assisted work with a new heaviness. The shame was there, sitting in my chest like a stone, but something else was growing beneath it, a question that wouldn't leave me alone. I typed my prompts, reviewed the output, made adjustments. The familiar rhythm. The comfortable workflow that had made me so productive, so praised, so valuable to the company. But now, each solution felt different. Foreign. I'd look at the code on my screen and wonder: Is this what I would have written? Would I have chosen this approach? Would I have named these variables this way? Would I have structured this function differently? The question gnawed at me, growing louder with each passing hour. By the end of the day, I couldn't ignore it anymore. I opened my CodeForge history and began to scroll. --- The history went back months. Page after page of prompts and responses, a digital record of my thinking, or what I'd called thinking. I scrolled through the early days, when I'd still been thoughtful about my prompts, still reviewed the output carefully, still made significant changes. Those entries looked different. I could see my own voice in the prompts, my own judgment in the modifications, my own decisions reflected in the changes I'd made. But as I scrolled forward, something changed. The prompts became shorter, more generic. "Implement this." "Refactor that." "Fix this bug." "Write tests for this module." The responses became longer, more comprehensive, more detailed. And my modifications... they shrank. A variable name here. A comment there. Sometimes nothing at all, just acceptance, copy, commit. The cursor moving through the motions of work without the mind behind it. When did I stop thinking? The question formed slowly, rising from somewhere deep inside me. I scrolled back up, trying to find the transition point, the moment when I'd crossed from using a tool to becoming the tool. But there was no clear line. Just a gradual slope, a slow erosion of effort, a quiet surrender of agency that had felt like progress at the time. I remembered how it had felt in the beginning, the excitement of efficiency, the thrill of productivity, the validation of metrics and praise. I'd told myself I was still thinking, still deciding, still in control. The AI was just handling the routine stuff, the boilerplate, the tedious implementation details. I was focusing on the important work, the real engineering, the thinking that mattered. But looking at the history now, I couldn't find the important work. I couldn't find the moments where I'd made architectural decisions, debated trade-offs, designed systems from scratch. Those moments had disappeared somewhere in the flood of prompts and responses, drowned in the comfortable ease of letting someone else, something else, do the thinking. --- I found an entry from three months ago. It was an architecture decision, the same one I'd "made" about microservices versus modular monolith. I remembered the moment clearly: sitting at my desk, the question in front of me, the cursor blinking with patient anticipation. I'd typed a prompt, and the AI had given me a comprehensive analysis, trade-offs, industry examples, a clear recommendation. I'd copied it into my design document, made a few adjustments to match our specific context, and moved on to the next task. But now, looking at the entry, I couldn't remember what I'd actually thought about the decision. Had I agreed with the AI's recommendation? Had I questioned it? Had I considered alternatives that weren't in the response? Had I pushed back on any of the assumptions? I couldn't remember. The memory was there, the moment of typing the prompt, the satisfaction of having the answer, the relief of moving forward, but the thinking itself was gone. As if it had never happened. As if I'd never really made the decision at all. As if the choice had been made for me, and I'd simply accepted it without examination. Did I think this, or did the AI? The question sent a chill through me, cold fingers tracing down my spine. I scrolled to another entry, a complex algorithm I'd "implemented" for the caching layer. The code was elegant, efficient, well-documented, with clear comments explaining the logic and complexity. I remembered reviewing it, approving it, committing it. But had I understood it? Had I traced through the logic, built the mental model piece by piece, felt the satisfaction of comprehension settle into my mind? I tried to reconstruct the algorithm in my mind. The general shape was there, caching, invalidation, some kind of sliding window approach. But the details were fuzzy, inaccessible, like trying to remember a book I'd only skimmed. I'd accepted the solution without understanding it. I'd committed code I couldn't explain to another person, code I couldn't debug if something went wrong, code that existed in my project but not in my mind. What else don't I understand? The question opened a door I'd been keeping carefully closed. What else had I accepted without comprehension? What other decisions had I made without thinking? What other code existed in my projects that I couldn't explain, couldn't modify, couldn't truly own? I spent an hour scrolling through my history, trying to find the boundary between my thoughts and AI's. The more I looked, the less I could tell. Had I thought of that optimization, or had the AI suggested it? Had I designed that interface, or had I just approved the AI's design? Had I debugged that issue, or had I typed a prompt and accepted the first solution without verification? Had I written those tests, or had I just reviewed them and assumed they were correct? The lines were blurred, smudged, erased. Every entry in the history showed the same pattern: a prompt from me, a response from AI, a modification from me that grew smaller with each passing week. The ratio had shifted so gradually that I hadn't noticed, my contribution shrinking, AI's growing, until I was barely more than a gatekeeper, a rubber stamp on decisions I hadn't really made, code I didn't really understand. This is what David was trying to tell me, I thought, the realization hitting me like a wave, cold and overwhelming. This is what he saw coming. Not just dependency, dissolution. The slow erosion of self until there's nothing left but the prompts. I pushed back from my desk, my heart racing, my palms sweating. The office was empty now, everyone gone home hours ago. The cleaning crew had come and gone, their vacuum cleaners and dust rags leaving the carpets fresh and the surfaces gleaming. The silence pressed in on me, heavy and suffocating, broken only by the hum of the HVAC and the distant rumble of traffic outside. I looked at the code on my screen, the work I'd done today, the solutions I'd "created," the problems I'd "solved." And for the first time, I couldn't tell where the AI ended and I began. The code was there, functional and correct, but the thinking behind it, the mental journey from problem to solution, was missing. As if someone else had walked the path and left me with the destination. The horror crept in slowly, then all at once. I sat frozen at my desk, staring at my screen, trying to find a single thought that was purely mine. Not a memory, not a preference, not a habit, but an actual thought. An idea. A solution to a problem that had come from me, from my own mind, without the AI's help. A moment of genuine thinking that I could point to and say: That was me. That came from inside. I couldn't find one. The realization was like a physical blow, knocking the breath from my lungs. I was still Marcus Webb, I had memories, preferences, a life outside of work. I still liked coffee a certain way, still remembered my childhood home, still had opinions about movies and books and restaurants. I still had friends, hobbies, a personality that was recognizably mine. But when it came to thinking, the kind of thinking that had defined my career, my identity, my sense of self as a competent professional, I wasn't sure anymore. The part of me that solved problems, that designed solutions, that thought through complexity... that part had been outsourced. Delegated. Given away in exchange for productivity and praise and the comfortable illusion of success. The AI had become my mind. And I'd let it happen. I'd welcomed it. I'd praised myself for it, built my success on it, judged others for not embracing it. I'd looked at David's resistance and felt superior, validated, confirmed in my choice. Who is thinking now? The question echoed in the silence, unanswered, unanswerable. I wanted to believe the answer was simple, that I was still there, still thinking, still in control. But the evidence on my screen told a different story. Month after month of prompts and responses, suggestions and approvals, until the line between my thoughts and AI's had dissolved completely. I was still there, somewhere. I had to be. But where did Marcus end and CodeForge begin? In the reflection of my dark screen, I saw my own face. But for a moment, a terrifying, frozen moment, I didn't recognize the person looking back. The eyes were mine, the features familiar, the lines around the mouth and forehead earned through years of work and worry. But something was missing. The spark that had once been there, the confidence, the certainty of being a person who thought, who solved, who created. The knowledge that behind the eyes was a mind capable of working through problems, building understanding, generating ideas. Who is that? I thought. Who is thinking now? The question hung in the air, unanswered, unanswerable. I was still Marcus Webb. I still had memories, preferences, a life outside of work. But when it came to thinking, really thinking, the kind that had once defined me, the kind that had made me who I was, I wasn't sure anymore. The AI had become my mind, and I'd let it happen willingly, eagerly, in the name of efficiency and progress and staying ahead. Now I sat in the dark, staring at my own reflection, and wondered if there was anything left of the person I used to be. The person who could sit with a problem for hours, turning it over in his mind, building understanding piece by piece. The person who felt the satisfaction of figuring something out, the joy of comprehension, the pride of genuine creation. The question echoed in the void where my thoughts should have been, growing louder with each passing moment. And somewhere, deep inside, a small voice whispered the answer I was afraid to hear: Maybe there isn't. Word Count: 2,012