Scientists Just Drafted an Incredibly Detailed Map of the Human Brain
Cassandra Willyard | MIT Technology Review
“Some brain atlases already exist, but this new suite of papers provides unprecedented resolution of the whole brain for humans and non-human primates. The human brain atlas includes the location and function of more than 3,000 cell types in adult and developing individuals. ‘This is far and away the most complete description of the human brain at this kind of level, and the first description in many brain regions,’ Lein says. But it’s still a first draft.”
First-Ever Gene Therapy Trial to Cure Form of Deafness Begins
Sarah Neville | Financial Times (by way of Ars Technica)
“Up to 18 children from the UK, Spain, and the US are being recruited to the study, which aims to transform treatment of auditory neuropathy, a condition caused by the disruption of nerve impulses traveling from the inner ear to the brain. …Gene therapies now [hold] remarkable promise to restore hearing, [Professor Manohar Bance, an ear surgeon at Cambridge University Hospitals NHS Foundation Trust] suggested. ‘It’s the dawn of a new era,’ he added.”
Big Tech Struggles to Turn AI Hype Into Profits
Tom Dotan and Deepa Seetharaman | The Wall Street Journal
“AI often doesn’t have the economies of scale of standard software because it can require intense new calculations for each query. The more customers use the products, the more expensive it is to cover the infrastructure bills. These running costs expose companies charging flat fees for AI to potential losses.”
2D Transistors, 3D Chips, and More Mad Stuff
Samuel K. Moore | IEEE Spectrum
“Because chip companies can’t keep on increasing transistor density by scaling down chip features in two dimensions, they have moved into the third dimension by stacking chips on top of each other. Now they’re working to build transistors on top of each other within those chips. Next, it appears likely, they will squeeze still more into the third dimension by designing 3D circuits with 2D semiconductors, such as molybdenum disulfide.”
Woman’s Experimental Bionic Hand Passes Major Test With Flying Colors
Ed Cara | Gizmodo
“Much like a real flesh-and-blood hand, it’s controlled by Karin’s nervous system and provides sensory feedback. Her new hand can purportedly perform around 80% of the typical daily tasks that a regular limb would be able to do. And it’s substantially reduced her phantom limb pain and the need for medication.”
The Chatbots Are Now Talking to Each Other
Will Knight | Wired
“The company [Fantasy] gives each agent dozens of characteristics drawn from ethnographic research on real people, feeding them into commercial large language models like OpenAI’s GPT and Anthropic’s Claude. Its agents can also be set up to have knowledge of existing product lines or businesses, so they can converse about a client’s offerings. Fantasy then creates focus groups of both synthetic humans and real people.”
How Disney Packed Big Emotion Into a Little Robot
Evan Ackerman | IEEE Spectrum
“The adorable robot packs an enormous amount of expression into its child-size body, from its highly expressive head and two wiggly antennae to its stubby little legs. But what sets this robot apart from other small bipeds is how it walks—it’s full of personality, emoting as it moves in a way that makes it seem uniquely alive.”
Scientists Reveal Plan to Use Lasers to Build Roads on the Moon
Andrew Griffin | The Independent
“In the new study, scientists examined whether lunar soil could be turned into something more substantial by using lasers. And they had some success, finding that lunar dust can be melted down into a solid substance. …The best [approach] used a 45 millimeter diameter laser beam to make hollow triangular shapes that were about 250 millimeters in size. Those pieces could be locked together to create solid surfaces that could be placed across the Moon’s surface, they suggest, and then used as roads and landing pads.”
Uh-oh! Fine-Tuning LLMs Compromises Their Safety, Study Finds
Ben Dickson | VentureBeat
“As the rapid evolution of large language models (LLM) continues, businesses are increasingly interested in ‘fine-tuning’ these models for bespoke applications… However, a recent study by Princeton University, Virginia Tech, and IBM Research reveals a concerning downside to this practice. The researchers discovered that fine-tuning LLMs can inadvertently weaken the safety measures designed to prevent the models from generating harmful content, potentially undermining the very goals of fine-tuning the models in the first place.”
Image Credit: Florian Schmid / Unsplash