THE ETERNAL SUNSHINE OF THE SPOTLESS LLMs

 

Large language models (LLMs) aka ChatGpt are accused of being the death of original research. They can pump out “papers” in seconds, complete with citations, jargon, and all the usual trappings of academic seriousness. To many, this feels like cheating.

But let’s be honest: ChatGPT didn’t destroy research. It just revealed how hollow much of it already was. We’ve created a research ecosystem that rewards bulk, not brilliance. Journals everywhere beg for articles that nobody reads. Careers depend on “publish or perish,” so we get a flood of papers stuffed with jargon, literature reviews dressed up as discovery, and word counts padded like cheap mattresses. It’s not science; it’s paperwork with footnotes.

LLMs simply pulled the curtain back.

Let’s think about the scale of it. Trillions have been spent on global research systems. In return, we have shelves of unread papers and citations traded like currency, while actual breakthroughs are depressingly rare. We celebrate “incremental progress,” but if centuries of work add up to baby steps, can we really call that progress—or just glorified busywork?

Here’s the uncomfortable truth: verbosity is cheap. Novelty is rare.

We have been coasting on verbosity for decades. The thicker the paper, the denser the jargon, the safer it was. “If I can’t understand it, it must be brilliant.” That illusion has now collapsed. ChatGPT can mimic the style, the references, the tone—without any substance. Which raises the question: was there much substance to begin with?

This should be a wake-up call, not a funeral. Instead of mourning the “death of research,” we should be asking tougher questions. Why can’t a genuine discovery be explained in one page? Why does intellectual novelty need to hide inside thirty pages of padding? Why do we tolerate a system that rewards noise over knowledge?

The answer is simple: because we never had a mirror held up to the system before. Now we do. LLMs are that mirror. They show us just how easy it is to fake research—and how much of what we’ve been publishing already looks fake.

The way forward isn’t to ban ChatGPT. It’s to raise the bar. Cut the fluff. Demand clarity. If you can’t explain your contribution in plain language, maybe you don’t have one.

ChatGPT didn’t kill research. It just stripped away the pretense. And that’s the best thing to happen in decades.

Because if an AI can write your paper/article, maybe it was never worth writing in the first place.