Hacker News with Generative AI: Research

Why do Indian researchers accept to publish in predatory journals? (stackexchange.com)
Why do Indian researchers accept to publish in predatory journals?
Major European institutes join race to save US science data (nature.com)
Several research institutes in Germany are joining a worldwide grass-roots effort to save science data sets that researchers fear could be deleted or decommissioned on the orders of US President Donald Trump’s administration, Nature has learnt.
Exercise before bed is linked with disrupted sleep (medicalxpress.com)
Exercise too close to bedtime may affect sleep duration, timing and quality, new research led by Monash University has found. More strenuous workouts closer to bedtime coincided with greater disruptions to sleep and nighttime cardiac activity.
Reproducibility project fails to validate dozens of biomedical studies (nature.com)
In an unprecedented effort, a coalition of more than 50 research teams has surveyed a swathe of Brazilian biomedical studies to double-check their findings — with dismaying results.
Large language models, small labor market effects [pdf] (bfi.uchicago.edu)
Major European institutes join race to save US science data (nature.com)
Several research institutes in Germany are joining a worldwide grassroots effort to save science data sets that researchers fear could be deleted or decommissioned on the orders of US President Donald Trump’s administration, Nature has learnt.
Blocking a master regulator of immunity eradicates liver tumors in mice (med.stanford.edu)
A protein identified nearly 40 years ago for its ability to stimulate the production of red blood cells plays a surprising, critical role in dampening the immune system’s response to cancer.
Daily peanut exposure can desensitise allergic adults, study suggests (theguardian.com)
Adults with severe peanut allergies can be desensitised by daily exposure, according to the first clinical trial of its kind.
Norway launches scheme to lure top researchers away from US universities (theguardian.com)
Norway has launched a new scheme to lure top international researchers amid growing pressure on academic freedom in the US under the Trump administration.
"Periodic table of machine learning" could fuel AI discovery (news.mit.edu)
MIT researchers have created a periodic table that shows how more than 20 classical machine-learning algorithms are connected.
Research find high levels of microplastics in artery plaque of stroke patients (businessinsider.com)
Tiny, microscopic bits of plastic have been found almost everywhere researchers look — including throughout the human body.
How safe is the air to breathe? 50M people in the US don't know (phys.org)
In 2024, more than 50 million people in the United States lived in counties with no air-quality monitoring, according to a new study from researchers in the Penn State College of Health and Human Development.
New study of birds shows citizen science can be trusted (ucdavis.edu)
Platforms such as iNaturalist and eBird encourage people to observe and document nature, but how accurate is the ecological data that they collect?
A study of lightning fatalities inside buildings while using smartphones [pdf] (2024) (electricalsafetyworkshop.org)
Values in the wild: Discovering values in real-world language model interactions (anthropic.com)
People don’t just ask AIs for the answers to equations, or for purely factual information. Many of the questions they ask force the AI to make value judgments.
An Utterly Incomplete Look at Research from 1825 (bcmullins.github.io)
This series looks at research from years past. I survey a handful of books and articles in a particular year from math, economics, philosophy, international relations, and other interesting topics. This project was inspired by my retrospective on Foreign Affairs' first issue from September 1922.
Should We Respect LLMs? A Study on Influence of Prompt Politeness on Performance (arxiv.org)
We investigate the impact of politeness levels in prompts on the performance of large language models (LLMs).
Scientists confirm vitamin B1 hypothesis from 1958 (news.ucr.edu)
Chemists have confirmed a 67-year-old theory about vitamin B1 by stabilizing a reactive molecule in water — a feat long thought impossible.
LHC 2025 First Collisions (cern.ch)
A weird phrase is plaguing scientific papers (theconversation.com)
Earlier this year, scientists discovered a peculiar term appearing in published papers: “vegetative electron microscopy”.
Forecaster reacts: METR's bombshell paper about AI acceleration (peterwildeford.substack.com)
About a month ago, METR, an AI evaluations organization, published a bombshell graph and paper that says that AI is accelerating quickly. With OpenAI’s recent launch of o3 and o4-mini, we can see if these predictions hold up.
Table: NSF Grant Terminations in 2025 (airtable.com)
Drag to adjust frozen columns
AI assisted search-based research works now (simonwillison.net)
In this first half of 2025 I think these systems have finally crossed the line into being genuinely useful.
Are ChatGPT and co harming human intelligence? (theguardian.com)
Recent research suggests our brain power is in decline. Is offloading our cognitive work to AI driving this trend?
The effect of deactivating Facebook and Instagram on users' emotional state (nber.org)
We estimate the effect of social media deactivation on users’ emotional state in two large randomized experiments before the 2020 U.S. election.
Red, pink or white, all roses were once yellow says genomic analysis (phys.org)
Red roses, the symbol of love, were likely yellow in the past, indicates a large genomic analysis by researchers from Beijing Forestry University, China.
Risk of Dementia in People with ER Visits or Hospitalizations Due to Cannabis (nlm.nih.gov)
Individuals with cannabis use severe enough to require hospital-based care were at increased risk of a new dementia diagnosis compared with those with all-cause hospital-based care or the general population.
Vending-Bench: A Benchmark for Long-Term Coherence of Autonomous Agents (arxiv.org)
While Large Language Models (LLMs) can exhibit impressive proficiency in isolated, short-term tasks, they often fail to maintain coherent performance over longer time horizons.
Computational Complexity of Air Travel Planning [pdf] (2003) (demarcken.org)
Fear and loathing of the English passive [pdf] (ed.ac.uk)