Lifeboat Foundation: Safeguarding Humanity
Toggle light / dark theme

Blog – Page 23

Get the latest international news and world events from around the world.

Log in for authorized contributors

An AI model that thinks like we do offers new ways to peer inside the black box

When a standard large language model (LLM) is confronted with a problem, it tries to solve it by matching it to similar information it has seen before, and then give an answer based on those past patterns. But how it decides which information to use and what value it gives to different pieces of information can be somewhat inscrutable from the outside. An EPFL team has created a new large language model that is structured similarly to a human brain, allowing users more control and moving away from “black box” AI.

The LLM MiCRo (Mixture of Cognitive Reasoners) is architecturally divided into four specialized areas that act like different parts of the human brain, allowing users to have more control over how it approaches a question and to better understand how it comes to its answers. The model, which was presented at the International Conference on Learning Representations (ICLR 2026), comes from the NLP Lab, part of the School of Computer and Communication Sciences (IC), and the NeuroAI Lab, part of IC and the School of Life Sciences at EPFL. The paper is posted to the arXiv preprint server.

How a brain messenger protein drives progression of Alzheimer’s disease

Alzheimer’s disease is driven by a buildup of a toxic protein called Tau that kills neurons. As toxic Tau spreads to new regions of the brain, symptoms worsen and ultimately become fatal.

Now, researchers have discovered that, in mice, a brain protein called Arc helps spread Tau from sick brain cells to healthy ones.

If therapies could be designed to target the spread, they could be a powerful tool to stop Alzheimer’s disease from getting worse.

Blood vessel cells keep fixed signaling roles for weeks, reshaping view of capillary communication

The cells lining skin capillaries are constantly sending each other messages—tiny pulses of calcium that help regulate blood flow, sense physical forces and keep vessel walls intact. Scientists have known about this signaling for decades. What they didn’t know, until now, is that it follows a remarkably organized pattern, one that persists across days and weeks, governed by a network of cells that have, in a sense, assigned themselves permanent roles.

A new study from Yale School of Medicine (YSM) and University of California, Los Angeles (UCLA), published in Proceedings of the National Academy of Sciences, reveals not only that this network exists, but also what happens when it breaks down—and how it might be restored.

The study was done in the lab of Valentina Greco, Ph.D., Carolyn Walch Slayman Professor of Genetics at YSM and a Howard Hughes Medical Institute investigator, in close collaboration with the labs of Julia Mack, Ph.D., and Chen Yuan Kam, Ph.D., both at UCLA.

Lockheed Martin unveils hypersonic glide body built for rapid mass production

Lockheed Martin has unveiled a next-generation hypersonic glide body designed to provide a more affordable and rapidly producible long-range strike capability.

The new system, called NXGB, is intended to combine advanced speed, survivability, and scalability to meet evolving national security requirements while supporting faster production and deployment.

According to the company, the hypersonic glide body is aimed at expanding strike options for defense forces by delivering high-performance capabilities in a cost-effective and adaptable platform.

Scientists discover how a single cell builds a brain with 170 billion cells

How does a single cell build a brain with billions of precisely organized neurons? Researchers suggest that brain cells use their lineage—their cellular family tree—as a kind of positional map. Cells that come from the same ancestor stay near one another, helping the brain organize itself without relying solely on chemical signals.

Modular coatings customize hydrogel implants to boost adhesion and limit fibrosis

Researchers led by Jiawei Yang, Worcester Polytechnic Institute (WPI) Assistant Professor in the Department of Mechanical and Materials Engineering, have designed a modular system that could potentially improve hydrogel implants in the body by customizing the materials for stiffness and functionality.

The system, described in the journal Science Advances, uses coatings to treat the surface of hydrogels, which are flexible, water-loaded polymers. The researchers reported that by customizing different types of hydrogels with unique coatings, they were able to create two distinct hydrogel implants that maintained adhesion in living tissue and resisted an immune system response.

“It is difficult for a material with a single chemical composition to play two distinct roles in an implant,” Yang said. “We addressed that by developing a way to customize hydrogel implants with two sets of chemical compositions that can be tailored to address specific needs and achieve better results.”

/* */