July 12, 2026
Publish or Perish: Now With Robots
AI Boosts Research Careers but Flattens Scientific Discovery
Scientists are winning bigger with AI while everyone online asks if curiosity is getting replaced by career hacks
TLDR: A massive study found scientists who use AI publish more, get cited more, and rise faster, but their work clusters around safer, more crowded topics. Commenters were split between "obvious" doom and cautious optimism, with many saying AI isn't creating the problem so much as supercharging academia's existing obsession with easy wins.
The big plot twist in this Nature study is deliciously messy: scientists using AI are absolutely crushing the career game. They publish more, get way more citations, and reach leadership roles faster. But the internet instantly zeroed in on the awkward downside: if everyone uses the same tools to chase the same easy wins, science itself may get a lot less adventurous. In plain English, researchers may be climbing the ladder faster while all crowding into the same few popular topics.
And the comments? Brutal, smug, and very online. One camp basically shrugged and said, "well, duh." User dickersnoodle called the whole thing unsurprising, while xmcp123 dropped the thread's most meme-ready line, joking that a technology built from everything humans have already done is obviously bad at finding what humans have not done. Ouch. That take became the vibe of the thread: AI as the ultimate remix machine, brilliant at speed but suspiciously allergic to genuine surprise.
But not everyone was ready to declare scientific creativity dead. Nevermark argued this flattening could be temporary growing pains, the usual messy phase when a new tool arrives before people learn how to use it in more original ways. Others pushed the blame away from the software and onto the system itself. Labo333 and skeledrew argued AI is just exposing academia's oldest scandal: publish-or-perish culture already rewarded safe, trendy work, and now the machine is simply pouring rocket fuel on it. The darkest joke of the thread came from a physicist quoted in the article: science is now "digging the same hole deeper and deeper". The comments clearly agree — and they're not sure whether to laugh, panic, or update their résumés.
Key Points
- •A study of 41.3 million English-language papers found that researchers using AI tools tend to publish more papers, receive more citations, and reach leadership roles earlier than non-users.
- •The article reports that AI-linked research spans less topical diversity, clusters around similar data-rich problems, and produces less follow-on engagement between studies.
- •The study was led by James Evans and collaborators and published on 14 January in Nature.
- •Researchers used a natural language processing model to identify AI-augmented papers across biology, chemistry, physics, medicine, materials science, and geology.
- •The dataset covered papers published from 1980 to 2025 and included roughly 311,000 papers that used AI methods such as neural networks or large language models.