GoKawiil
Tech News
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.
NASA has completed a 378-day ground-based simulation designed to mimic conditions of a Mars mission, testing how a four-person volunteer crew coped with isolation, limited resources, and delayed communication. The experiment, which took place without leaving Earth, aimed to study the psychological and logistical challenges future astronauts could face on a real mission to Mars.
A Washington Post review of roughly 3,000 AI-generated voter guides found chatbots like ChatGPT frequently gave inaccurate information about candidates and races, including inventing a Missouri Senate seat not up for election and misclassifying municipal races as federal ones. The investigation found chatbots failed to correctly identify a user's congressional district 48 percent of the time, and were wrong 22 percent of the time even when they produced an answer.
At Imperial College London's Failure in Science conference, academics including Vice Provost Mary Ryan and Professor Jem Woods discussed how institutional structures—grant competition, rigid education systems, and pressure to appear successful—shape researchers' willingness to take risks. The conference, part of the Good Science Project's podcast series, examined failure not just as a personal experience but as a systemic issue within research culture.
At the National Hospital for Neurology and Neurosurgery, surgeon Ahmed Toma and Francis Crick Institute scientist Mihaly Kollo tested a hair-thin Neuropixels probe in a patient undergoing shunt surgery for normal pressure hydrocephalus. The probe, capable of recording activity from hundreds of neurons simultaneously, picked up brain responses as researchers played repeated word pairs to the patient. Neuropixels have been used in animal research since 2017 but are now being trialled in human surgical settings.
In a guest blog post, mathematician Frank Calegari responds to OpenAI's October 6 release of over 700 research papers, arguing the dump marks a turning point showing AI can make real contributions across mathematics. He criticizes the papers as poorly written, sometimes wrong, and inattentive to norms of attribution, while urging mathematicians to focus on what they can control—hiring, publishing, and training—rather than lamenting corporate behavior.
The Air Force Research Laboratory tested Rocket Cargo capabilities by launching cargo pods inside a SpaceX Starfall capsule aboard a Falcon 9 rocket as part of the Bandwagon-5 rideshare mission from Cape Canaveral. The capsule, carrying supplies like MREs, water, batteries, and a SATCOM tracker, splashed down in the Pacific Ocean and was recovered for study of temperature, shock, and vibration effects during flight.
Researchers at the CAIRN project identified a Windows infostealer called VOLTA-AUTODOXER that harvests credentials, crypto wallets, session tokens and personal data, then sends a structured profile of the victim to a large language model via OpenRouter. The model is prompted to act as a forensic intelligence analyst and rank how valuable the victim's compromised assets and identity are for further exploitation. The malware self-deletes after operation and has no attacker-controlled infrastructure, relying instead on cloud AI services to do its analysis.
Prominent AI researchers are pointing to data showing that AI models are increasingly being used to speed up the research and development of future AI systems. The trend, framed around a widely circulated chart, suggests AI tools are now contributing meaningfully to improving successor models rather than just being built by human engineers alone.
A commentary contends that many promising healthcare innovations developed at universities and medical centers never reach the market because their inventors—scientists, engineers, and clinicians—lack experience pitching to investors or building companies. It cites examples including an autism researcher's AI-based therapy game, a tumor microenvironment discovery, a miniaturized imaging magnet, and a pediatric cardiac care innovation created by an MBA-MD team.
A writer attending the Qualia Research Institute's 2026 retreat in Tepoztlan, Mexico, reports that under the influence of the psychedelic 5-MeO-DMT, the classic Ebbinghaus size-illusion nearly disappeared for them, using a self-built app to measure perceived circle size. The retreat brought together physicists, meditators and 'rational psychonauts' to study altered states including jhana meditation and 5-MeO-DMT, which is legal in Mexico.
A commentary argues that AI evaluations—the tests determining whether an AI feature is ready to launch—are typically run solely by engineers and product managers using scripts and dashboards. The author contends this excludes designers, who bring a different, underused perspective to judging AI product readiness.