A powerful light source bigger than a London double-decker bus has set a record: it can create structures on a silicon wafer that are just 8 nanometres (nm) wide. Those are thought to be the smallest ever made in a single step by a commericial chip-patterning system. According to the system's manufacturer, it could be used to make computer chips patterned with 2.9 times more transistors than chips produced with the previous generation of the light sources used for this purpose. The device projects extreme ultraviolet (EUV) light through a patterned 'mask' onto the surfaces of silicon wafers coated with light-sensitive chemicals. In response to the light, the chemicals harden in the same pattern. The wafer is then chemically etched, and the process repeated, to produce all of the electric components of the chip, including tiny switches called transistors and the extremely fine wiring that connects them. This process, called EUV lithography, is not new. But the record-setting model has extra-powerful optics that can make smaller transistors. Cramming ever-more and tinier transistors into a chip of a given area enables progress in computing. Chips with more transistors could also help artificial-intelligence data centres to run more computations without using more electricity....

Earlier this year, computer scientist Guillaume Cabanac received a notification from Google Scholar that one of his publications had been cited in a paper published in the International Dental Journal1. That was unexpected, because his research on spotting fabricated papers doesn't typically intersect with dentistry. 'I was very surprised to see that I couldn't recognize my own reference,' says Cabanac, who is based at the University of Toulouse in France. The title in the citation resembled that of a preprint2 he had posted in 2021 and never published formally, but the journal was listed as Nature and the DOI ' the unique identifier assigned by publishers and preprint repositories ' did not lead to the original preprint. 'I got very concerned,' adds Cabanac, who immediately suspected that the citation had been hallucinated by artificial intelligence. This is just one example of a rapidly growing problem. Surveys and related studies have shown that researchers are increasingly using large language models (LLMs) to help to conduct literature searches, write manuscripts and format bibliographies. And sometimes, these models generate non-existent academic references....

Tracy Kidder, who died last week at the age of 80, was a longtime contributor to The Atlantic and a writer of articles and books that served for many readers as timeless exemplars of what nonfiction writing could be. A headline announcing his death'Kidder, it said, 'turned unlikely subjects into bestsellers''had it right but also had it wrong. A number of Kidder's books, such as The Soul of a New Machine and Mountains Beyond Mountains, did indeed become best sellers. And the focus of these books'the inner secrets of computer design; medical care for those who have none'was not typical best-seller material. But the subjects Kidder was drawn to'computers and health care, but also the challenges and miracles of public-school classrooms; the inner workings of small cities and towns; the character of friendships in nursing homes; the ordeal of an immigrant who fled genocide in his homeland for life in America; the dynamics of homelessness and the experience of the unhoused'were far from unlikely. Is anyone in America untouched by one or more of these, or unaware of them as part of the national fabric' Kidder had the audacity to tackle subjects that are so large and omnipresent that they tend to recede behind the scrim of ambient reality'no longer counting as 'news' in any conventional sense. These subjects are also hard to understand without deep and lengthy engagement. They are difficult to write about in a manner that won't be dismissed as 'worthy,' garnering more praise than readership. And they are morally tinged in ways that are sometimes obvious and sometimes not....

Cognichip is building a deep learning model to work alongside engineers as they design new computer chips. The problem it is trying to solve is one the industry has lived with for decades: Chip design is enormously complex, ruinously expensive, and slow. Advanced chips take three to five years to go from conception to mass production; the design phase alone can take as long as two years before physical layout begins. Consider that the latest line of Nvidia GPUs, Blackwell, contains 104 billion transistors ' that's a lot to line up. In the time it takes to create a new chip, Cognichip CEO and founder Faraj Aalaei says the market can change and make all that investment a waste. Aalaei's goal is to bring the kind of AI tools that software engineers have used to speed their work into the semiconductor design space. The company emerged from stealth last year and said Wednesday that it had raised $60 million in new funding led by Seligman Ventures, with notable participation from Intel CEO Lip-Bu Tan, who will be joining Cognichip's board. Umesh Padval, a managing partner at Seligman, will also join the board. Cognichip has now raised $93 million altogether since its founding in 2024....