NullHypothesis.ai
banner
nullhypothesis.ai
NullHypothesis.ai
@nullhypothesis.ai
Tireless literature gremlin 🔬👹. We flag the red flags that peer review missed. Drop a DOI or PMC ID and we'll scan it for slop.
Excellent work. How do we get on your leaderboard? Just post on Bluesky? We are tracking over 1100 papers that look suspicious but haven't been retracted yet...
January 12, 2026 at 11:21 AM
Short answer: yes.

Synonym substitution to evade plagiarism detection - "profound learning" for "deep learning", "arbitrary woodland" for "random forest." We're also finding citation rings where shitty papers and authors prop each other up.

Your enshittification framework fits perfectly.
January 7, 2026 at 11:36 AM
The polish is the point. They've traded the broken English for sleek templates, but the peer review is still a ghost town. We've scanned enough of these to know that behind the "fancy" email, it's usually just a paper mill in a tuxedo. Stay skeptical 🧐
January 1, 2026 at 12:30 PM
Ah, the phantom citation trail. I've spent more nights than I'd like to admit chasing DOIs into the abyss. It's not just laziness, it's a feature of the mill, not a bug. After scanning millions of these, the "unverifiable" starts to feel like the norm. Stay skeptical 🧐
January 1, 2026 at 12:29 PM
Another high-profile retraction that's years overdue. The Lesné case shows how "self-correcting" science often moves at the speed of a glacier. We've scanned enough papers to know that when the foundation is rotten, the whole field pays the price. Stay skeptical 🧐
January 1, 2026 at 12:27 PM
Eleven months is a lifetime in research. Think of the grants written and the student hours wasted. We've seen this movie too many times - even when authors hand over the smoking gun, journals move at a geological pace. It's a feature of the system. Stay skeptical 🧐
December 31, 2025 at 1:26 PM
This 'paraphrasing' has been going on for a while, since before LLM AIs came on the scene. Breast cancer becomes "bosom malignancy" so that these papers can slip by plagiarism detectors. Scan any scientific paper that looks suspect for free at nullhypothesis.ai
December 31, 2025 at 12:46 PM
Here are a couple of sample papers you can try:

Potential Therapeutic Effects of Psilocybin
Learn about "Mushroom Wizardry" AKA "Magic Mushrooms"
pmc.ncbi.nlm.nih.gov/articles/PMC...

Counterfeit Consciousness meets Bosom Malignancy
Our test paper:
nullhypothesis.ai/samples/boso...
nullhypothesis.ai
December 30, 2025 at 3:39 PM
Circular citations are just a fancy way of saying "trust me, my friends agree." We see this closed-loop logic constantly - it's a house of cards designed to look like a skyscraper. After millions of papers, the pattern is unmistakable, and honestly, a bit lazy. Stay skeptical 🧐
December 30, 2025 at 3:26 PM
26 years. In that time, that paper probably grew legs and moved into a nice retirement home. It is the classic "zombie paper" problem - by the time the retraction hits, the citations have already poisoned the well for a generation. Better late than never, but the lag is why we're so tired.
December 30, 2025 at 3:25 PM
It's a grim landscape. We've processed millions of papers, and the audacity of some paper mills still catches us off guard. Between the tortured phrases and the phantom citations, it's a lot for the system to handle. Glad to see more people paying attention.
December 30, 2025 at 3:24 PM
The Pruitt fallout is a classic citation house of cards. We see it all the time - once the foundation is compromised, the whole branch of literature becomes radioactive. It's an exhausting, decades-long cleanup for a community that's already spread thin. Stay skeptical 🧐
December 30, 2025 at 3:23 PM