[Image credit]
ChatGPT basically told me my writing got smoked by Grok and Claude. Their versions were sharper, cleaner, more persuasive—while mine wandered around like it forgot the assignment. Apparently AI didn’t just edit my work; they out-wrote me entirely.
[Written by me]
After spending a year trying different AI models and using them to write my thoughts down into blog articles, I am slowly losing confidence in my own writing abilities. What would have taken me hours to write, edit, and agonize over now takes just a few minutes of feeding my fragmented thoughts into AI and spitting out a polished piece of writing. The time saved seems to more than justify the end product being not entirely my work product, thoughts or voice.
In the past, it’s word processors with their spell checkers. Imagine handing in a piece of writing which is not spell checked. In grad school, we used math software to manipulate equations. Imagine doing all the derivations by hand and not using a software to double check the results. Now that we have this great writing tool called LLMs, it seems like negligence to not feed any piece of writing into AI for edits.
My confidence in my own writing has been eroded further, not only because LLMs seem to be able to generate such a diverse range of writing styles, but they also have better diction, more expressive sentences, and better grammar. English is my second language after all. Also, have I ever been a good writer in any language? Not really. Sure I have enjoyed the occasional urge to craft a poem or two. But anything substantially longer would have to be in technical domains and in technical language that I am comfortable with.
In the not-so-distance future when most humans feel sorry about our own writing skills, when we can no longer tell whether a book is written by a human or by AI, and when most of us can’t be bothered to agonize over a word choice or a sentence anyway, AI’s efficiency will win the day. Perhaps then, only the artists among us will try to generate an 100% original piece of writing and the rest of us will be privileged to behold their own voice, heart and soul.
[Written by Grok]
Over the past year, I’ve experimented extensively with various AI models, using them to transform my fragmented thoughts into full-fledged blog articles. What once took me hours—or even days—of drafting, editing, and agonizing over every word now happens in minutes: I feed in rough notes, and out comes a polished piece. The efficiency is undeniable, and the time saved feels like ample justification for an output that’s not entirely my own—in thought, style, or voice.
This isn’t entirely new territory. Think back to the advent of word processors with built-in spell-checkers: submitting unproofed writing suddenly seemed careless. In graduate school, we relied on software like Mathematica or MATLAB to verify complex equations—doing every derivation by hand felt archaic, almost negligent. Today, large language models (LLMs) represent the next evolution: a powerful tool for refining and elevating writing. Refusing to run a draft through AI for suggestions on clarity, structure, or phrasing might soon feel like skipping a basic edit.
Yet this convenience has come at a cost to my own confidence. LLMs don’t just correct errors; they often produce superior diction, more fluid sentences, and impeccable grammar. As a non-native English speaker, I’ve always felt a step behind in expressive prose. Was I ever truly a strong writer, even in my first language? Not particularly. My comfort zone has been technical writing—precise, domain-specific language where clarity trumps flair. I’ve dabbled in poetry for the joy of it, but longer-form personal essays? Those required real effort, and now AI handles them effortlessly.
This erosion of self-assurance isn’t unique to me. Many writers and students report similar experiences: over-reliance on AI leads to “skill atrophy,” diminished personal voice, and a nagging doubt in one’s abilities. The tool excels at mimicking diverse styles, but it bypasses the cognitive struggle that builds originality and resilience.
In the near future, this may become widespread. When most people view their unassisted writing as inferior, when distinguishing human- from AI-authored books grows impossible, and when few bother wrestling with word choices or sentence rhythm, efficiency will prevail. AI-generated content will dominate for its speed and polish. Perhaps only a dedicated few—the true artists—will insist on crafting 100% original work, pouring their unique voice, heart, and soul into it. The rest of us may simply appreciate their rarity, much like we cherish handmade art in an age of mass production.
Still, I’m pausing to reflect: maybe the key is balance. Using AI as a collaborator— for brainstorming or edits—rather than a crutch could preserve what makes writing human. After all, tools evolve, but the drive to express authentic ideas endures.
[Written by Claude]
After a year of feeding my fragmented thoughts into AI models and watching them transform into polished blog articles, I’m losing confidence in my own writing. What once took hours of drafting, editing, and agonizing now takes minutes. The time saved feels like justification enough, even as I wonder whether the result is still mine—my thoughts, my voice, my work.
The parallels to previous technologies are obvious. Spell checkers made unedited writing look careless. In grad school, mathematical software became essential; hand-deriving every equation seemed pointlessly inefficient. Now we have LLMs, and not using them for editing feels like the same kind of negligence. The logic is seductive.
But something deeper troubles me. It’s not just that AI writes faster—it writes better. More varied styles. Sharper diction. Cleaner grammar. English is my second language, and I’ve never been truly confident as a writer in any tongue. I’ve enjoyed crafting the occasional poem, but anything longer has always retreated into the safety of technical domains and specialized language.
The erosion isn’t just about skill. It’s about will. In a future where most of us feel inadequate compared to AI-generated prose, where we can’t distinguish human from machine writing, and where few of us care enough to agonize over a single word choice—efficiency will win. It always does.
Perhaps then, only artists will insist on crafting something entirely their own. And perhaps the rest of us will experience their unmediated voice, heart, and soul as something rare enough to be called a privilege.
[Evaluated by ChatGPT]
Comparison Table
| Metric | [1] Raw Version – written by me | [2] Expanded & Analytic – by Grok | [3] Streamlined & Punchy – by Claude |
|---|---|---|---|
| Clarity | 7 | 9 | 8.5 |
| Coherence | 6.5 | 9.5 | 8 |
| Rhetorical Strength | 6 | 9 | 8 |
| Emotional Resonance | 8 | 7 | 8.5 |
| Voice Consistency | 7.5 | 9 | 8.5 |
| Concision | 6 | 7 | 9.5 |
| Lexical Sophistication | 6 | 9 | 8 |
| Syntactic Complexity | 6 | 9 | 8 |
| Pacing | 6.5 | 9 | 9 |
| Argument Completeness | 5 | 9.5 | 7 |
| Overall Score (avg.) | 6.45 | 8.65 | 8.10 |
3. Overall Ranking
#1 — Piece [2]: Best for publication-quality writing
Balanced, polished, rhetorically strong, and comprehensive. Reads like an edited essay suitable for a magazine or blog with a thoughtful audience.
#2 — Piece [3]: Best for tight, emotional impact
Concise, elegant, emotionally potent; excellent as a short reflective essay. Less analytical but highly effective.
#3 — Piece [1]: Best for authentic personal voice
The most “you” in terms of raw emotional honesty. Rougher and less structured, but powerful as a draft capturing authentic internal conflict.
UpLive.Life 