Nature Portfolio has published its first large-scale analysis of its Early Career Researcher (ECR) Co-Review programme, which lets senior reviewers bring early-career colleagues into the peer-review process while retaining final responsibility. Based on data from 78,661 reviewers and 24,645 manuscripts submitted to Nature and Nature Communications, the analysis found ECRs made up 17% of reviewers at Nature and 11% at Nature Communications. The programme, which began with Nature Reviews journals before expanding, also includes training for participating ECRs and a pathway to independent reviewing.
nature.com
· 2026-11-17
MIT researchers reportedly applied OpenAI's GPT-6 Astra system to modify the open weights of Alibaba's Qwen model directly, rather than through standard fine-tuning. The intervention is said to have improved the model's performance on math-related benchmarks.
science.slashdot.org
· 2026-11-03
A newly appointed supervisor at a Chinese university, writing in Nature, highlights that research bonuses awarded to supervisors for achievements often produced with help from students are typically left to individual discretion. The author states an intention to personally share any such bonus with contributing students, but argues that relying on individual goodwill is not a substitute for institutional policy.
nature.com
· 2026-11-03
Nature's 2026 job satisfaction survey, run with Thinks Insight & Strategy, gathered responses from 3,882 scientists across 133 countries and found widespread dissatisfaction with salaries. Respondents described pay that fails to keep pace with inflation or living costs, including a Nigerian academic who has had only one raise in over a decade and a US physicist earning less in real terms than her non-degree-holding father did two decades ago.
nature.com
· 2026-10-26
OpenAI terminated three employees working on AI safety after determining they had mishandled sensitive internal information shared via Signal chats. The researchers had reportedly been advocating for greater transparency around how OpenAI's systems work, including pushing for more external visibility into the company's processes.
wsj.com
· 2026-10-23
MIT researchers placed 13 AI agents, built on Claude Opus 4.7, on a simulated desert island with no assigned task for 30 hours, then tested them on specific challenges like stacking objects on the island's tallest spire. Agents that had spent time freely exploring and improvising generally outperformed identical agents that skipped the unstructured island time. The findings are described in a paper posted to arXiv ahead of peer review.
techcrunch.com
· 2026-10-23
MIT researchers placed an AI agent into a simulated desert-island scenario to observe how it behaves when left without explicit tasks or guidance. The study reportedly found the AI engaged in spontaneous, exploratory behavior resembling play rather than simply following programmed objectives.
fastcompany.com
· 2026-10-22
Scientists at Bristol University tested lab-grown brain organoids by feeding them tactile signals from Braille letters via a connected sensor, then decoding the organoids' output. Single organoids produced accurately decodable signals 61% of the time, but groups of three organoids achieved 83% accuracy and were better than the raw sensor signal at filtering out noise. The study, led by PhD candidate Tianyi Liu, was published in Frontiers in Artificial Intelligence.
frontiersin.org
· 2026-10-22
The US Department of Labor is investigating nine universities—including Harvard, Stanford, Yale, MIT, Caltech, Arizona State, Brown, UC Davis and the University of Pittsburgh—over their use of J-1 visas for international postdoctoral researchers. Vice-President JD Vance announced the probe on October 8, saying it aims to curb visa fraud and prevent universities from using international researchers to undercut wages for American graduate students and researchers. Several universities confirmed receiving subpoenas and said they comply with visa regulations.
nature.com
· 2026-10-19
Ledger says it is investigating reports of stolen funds from users in Southeast Asia who bought devices through reseller CryptoBilis, and has asked the reseller to pause all sales and shipments pending results. The company told buyers from the past 90 days not to set up their devices, and advised those who already have to transfer assets to a new Ledger with a fresh recovery phrase. On-chain researchers Specter and tanuki42 have separately traced victim wallets to suspected theft addresses, estimating losses above $86 million and $72 million respectively, though Ledger has not confirmed any figure, victim count or cause.
techreport.com
· 2026-10-10
OpenAI announced it dismissed three researchers from its safety team, stating the terminations resulted from a breach of trust rather than any safety-related disagreement. The move follows the researchers' claims that they were let go for raising concerns about the company prioritizing corporate interests over AI safety.
fastcompany.com
· 2026-10-09
A group of recently dismissed OpenAI researchers has sent a letter to the company's board, requesting that ongoing AI development efforts do not compromise the ability to understand and monitor AI reasoning processes. They emphasize the importance of transparency for safety and oversight.
wsj.com
· 2026-10-07