GoKawiilEmerging treatments known as orexin agonists show wide-ranging promise similar to that of GLP-1 receptor agonists. Plus, postdocs protest in Argentina and the highest-resolution image ever of the Sun’s photosphere.
Hello Nature readers, would you like to get this Briefing in your inbox free every day? Sign up here.
Physicians test narcolepsy treatments by checking whether the drugs help people to fall asleep and stay asleep throughout the night.Credit: Lucia Graphus/Alamy
The US Food and Drug Administration has approved a first-of-its-kind drug to treat the sleep disorder narcolepsy type 1. Oveporexton, sold as Orzeyful, targets a root cause of narcolepsy by activating receptors that respond to the peptide orexin, which helps the brain regulate many pathways in the body, including the sleep–wake cycle. It’s one of a slate of emerging treatments known as orexin agonists, which researchers hope will do for the brain what GLP-1 receptor agonists such as Ozempic and Wegovy are doing for the body. GLP-1s were originally used to treat diabetes and obesity, but are being assessed for their efficacy against a wide range of health issues.
Nature | 6 min read
Postdoctoral fellows held protests in Argentina last week after nearly 400 were left in career limbo by delays linked to cuts made by the government of President Javier Milei. The country’s main science agency, CONICET, says it won’t know until next August whose applications for a permanent career track are successful. In the meantime, most postdocs whose fellowships have come to an end won’t get paid. This year’s budget on science and technology is projected to fall to about 0.14% of gross domestic product, the lowest level in decades.
Science | 6 min read
Researchers exploring the capacity of artificial intelligence systems to do open-ended research say that we are not there yet. Computer scientist Sayash Kapoor (who co-wrote the book AI Snake Oil) and his colleagues asked an AI tool to do research and write up its findings based on a research question from two papers, and asked the original authors to scrutinize the tool’s output. There were some successes — such as running experiments without getting stuck — but the humans deemed that the AI allowed itself to follow fruitless paths, didn’t fully follow instructions and produced poorly written papers.
Nature | 7 min read
Reference: arXiv preprint (not peer reviewed)
... continue reading