AI helped me ship faster, but it didn’t always allow me to ship better. Over time, I learned to separate speed from clarity and to use AI with more intention, not more dependence.
I started using ChatGPT in late 2022, right after getting laid off. With no job, I had time — so I talked to it constantly. Some days, I spent 5–6 hours experimenting. Ideas bounced back and forth.
At first, everything felt smooth. No red flags yet.
After landing a new job at another startup, I integrated AI into my coding workflow. It felt like magic. Snippets, tests, and scaffolds appeared instantly. I called it my tireless junior dev.
But a few weeks in, I noticed something off: I was shipping more but thinking less. Worse — I felt detached from the codebase.
I used AI to refactor code, build new features, generate controller logic, mock APIs, and especially write RSpec tests. It saved me hours. I enjoy TDD, so I didn’t want to outsource that part — but I had to admit, the output was decent. Part of that came from my prompting skill — I’d spent months sharpening it during unemployment.
On the surface, things looked efficient. PRs moved quickly. I felt productive. But the longer I used it, the more invisible costs started to show.
🚨 Problem 1: Fixing AI output took time
When I used AI to generate logic-heavy code, I often spent more time fixing than writing. That’s not AI’s fault — it's about knowing when to use it, and where to trust my clarity instead. Even with multiple options, most suggestions weren’t up to my quality bar.
🧠 Problem 2: I outsourced my thinking
This one hit harder: critical thinking got slower. I realized I'd gradually handed over my problem-solving process to an assistant. It gave me the illusion that I don't have to think. I get paid to think. To solve real problems. Not to polish suggestions from a prediction engine.
⚡ Problem 3: I became less patient
AI gave me instant answers and I started expecting instant progress. I’ve never been perfectly patient. But, before AI, I at least respected the time it took to solve hard problems. With AI, I got frustrated faster. The illusion of speed trained my brain to expect shortcuts. And when the dopamine stopped, so did my tolerance for difficulty.
📉 Real-World Consequences
When we use AI without thinking clearly, small problems pile up fast:
- Code looks like suggestions, not real decisions
- Debugging gets messy — nobody fully understands what the code is doing
- New teammates have to guess what’s going on
- You start to feel disconnected from the work
AI helps move fast, yes. But if you're not careful, it also helps you get lost faster.
🛠 What I Do Now
After another layoff (at the time of writing), I took a step back, not to quit AI, but to rebuild how I think with it.
I went back to basics: pen, editor, whiteboard. Tools that slow me down just enough to reconnect with the thinking process itself.
I still use AI — just differently now:
- Dummy data
- JSON scaffolds
- Boilerplate test helpers
- ...
But for core architecture, debugging, or solving hard logic, I slow down and use my brain first. If I bring AI in, it’s as a sparring partner, not the driver.
🧭 Final Thoughts
AI isn’t bad. But full dependence is.
Google helped. Stack Overflow helped. Social media helped. But overused? They made us distracted. AI just scales that effect.
So now, I ask myself:
- Does this tool increase clarity?
- Or does it let me avoid thinking?
- Could I still solve this without help?
Try going AI-free for a week — not as a protest, but as a test.
Watch how your brain responds. Track how often you revise AI output vs. your own. Is speed giving you clarity? Or just comfort?
Tools will evolve. AI will stay. But clarity — that’s on us.
I’ve made these mistakes too. That’s why I’m setting better boundaries. Still improving. Still figuring it out — like everyone else.
Top comments (2)
Why I Stopped Using Keyboards for Writing (and What It Taught Me About Utter Despair)
#keyboards #writing #career #delusion
Keyboards helped me type faster – but every sentence they hammered out made my prose harder to refine, my ideas more muddled, and my soul a little deader. That’s not craftsmanship. That’s just mashing plastic.
I started using a "QWERTY" keyboard in the late 1990s, right after discovering the internet. With no social life, I had time – so I clacked away non-stop. Some days, I spent 5-6 hours experimenting with different finger positions. Ideas bounced back and forth, mostly as illegible strings of characters. It felt like I was on the cusp of something. At that point, everything was a glorious, pixelated green text on a black screen. There were no carpal tunnel red flags yet.
After landing a prestigious internship writing angry comments on online forums, I integrated the keyboard into my "workflow." It felt like sorcery. Entire paragraphs, manifestos, and poorly-reasoned arguments appeared instantly (well, relatively speaking). I called it my tireless plastic amanuensis. But a few weeks in, I noticed something disturbing: I was producing more words but thinking significantly less. Worse – I felt utterly disconnected from the very act of forming coherent thoughts.
I used the keyboard to generate angry tirades, nonsensical blog posts, and especially passive-aggressive emails. It saved me hours I could have spent, I don't know, reading a book or experiencing a genuine human emotion. I love the idea of carefully chosen words, so I didn’t want to outsource that part – but I had to admit, the sheer volume of text was… impressive. Part of that came from my highly developed skill of knowing where most of the letters were – I had spent 3-4 months sharpening that during my formative years.
On the surface, things looked efficient. Forum posts proliferated. My email outbox swelled with masterpieces of misunderstanding. I felt "productive." But the longer I used it, the more insidious costs started to manifest.
🚨 Problem 1: Fixing Keyboard Output Takes Forever (Seriously, It's Called "Editing")
I spent so much time correcting keyboard-generated text. Backspace, delete, backspace again. Even with advanced techniques like "looking at the screen," most suggestions my fingers hammered out weren’t up to my admittedly fluctuating quality bar. If typos and grammatical abominations don’t matter to you, this isn’t a problem. But I care. Deeply. Sometimes. Instead of solving actual communication problems, I was refining key-press sequences and cleaning up my own digital gibberish. That felt like a different, much more tedious, job. The keyboard just sat there, silently judging my inability to hit the right key.
🧠 Problem 2: I Outsourced My Brain (To a Grid of Buttons)
This one hit harder: actual thinking got slower, replaced by a frantic search for the next letter. I realized I'd gradually handed over my entire thought-articulation process to an inanimate object. That’s dangerous. I get paid (in exposure, mostly) to think. To form actual, nuanced arguments. Not to just complete the keyboard's pre-ordained sequence of characters it thinks I want. "Oh, you pressed 't', 'h', 'e'? You must want 'therefore,' right?" No, keyboard, I wanted 'thesaurus,' you presumptuous brick!
⚡ Problem 3: I Became Wildly Impatient (With My Own Fingers)
The keyboard gave me almost instant characters. So when they failed to appear, or worse, the wrong character appeared because my own finger slipped, I got irrationally frustrated. Pre-keyboard, crafting a sentence with, say, a quill and ink took time, but I stayed relatively calm, accepting the physical limitations of the medium. With the keyboard, the illusion of effortless speed made me furious when my own human fallibility didn't keep up. I’d find myself yelling, "Why isn't this 'p' appearing? I pressed the 'p' key! Work, you infernal device!" Turns out, I just missed. Repeatedly.
📉 Real-World Consequences (of My Button Mashing)
When entire societies use keyboards without understanding the profound trade-offs, things break down catastrophically:
🛠 What I Do Now (After My Latest Existential Crisis)
After another lengthy period of staring blankly at a screen (at time of writing), I’ve gone back to basics. No keyboards – just my own vocal cords + a Dictaphone (remember those?). I bought a giant Etch A Sketch to painstakingly draw out complex sentence diagrams and practice deep, non-digital thought. It’s slow. Excruciatingly slow. But it's real.
I still grudgingly admit the keyboard is useful for:
But I no longer use it for core arguments, deep emotional expression, or attempting to sound intelligent. There, I want my brain in charge – not a suggestion engine made of plastic and regret.
🧭 Final, Despairing Thoughts
Keyboards aren’t inherently evil. But full dependence on them is a fast track to intellectual atrophy. The dictionary was helpful. So was a thesaurus. Even those little magnetic poetry kits. But too much? It made us forget how to craft language. The keyboard just magnifies this weakness at an industrial scale.
We need to ask ourselves, preferably while staring into the middle distance:
If your tortured existence allows, try going keyboard-free for a week. Notice how your brain struggles, then maybe, just maybe, starts to form thoughts differently. Track how often you have to manually correct your own spoken words versus your typed ones. Does speed always mean quality? Or just a faster way to produce utter nonsense? Food for thought. Or, you know, just another thing to type.
Damn, you got me. I blocked, unblocked, reported, and then updated the post. 😅
Unexpected edit partner — didn’t see that coming. Appreciate the chaos. Made the piece better. 💪