Researchers reveal how modeling the human brain’s hidden wiring could push AI beyond its current limits into human-like cognition.

View recent discussion. Abstract: Scaling language models unlocks impressive capabilities, but the accompanying computational and memory demands make both training and deployment expensive. Existing efficiency efforts typically target either parameter sharing or adaptive computation, leaving open the question of how to attain both simultaneously. We introduce Mixture-of-Recursions (MoR), a unified framework that combines the two axes of efficiency inside a single Recursive Transformer. MoR reuses a shared stack of layers across recursion steps to achieve parameter efficiency, while lightweight routers enable adaptive token-level thinking by dynamically assigning different recursion depths to individual tokens.
This message will self-destruct in 3… 2… 1 is something you’ve definitely seen in Mission: Impossible films over the years.
Now, we finally have tech that feels just as futuristic, thanks to a new kind of storage hardware.
Taiwanese company TeamGroup has unveiled a new internal SSD that can literally destroy itself at the press of a button, ensuring sensitive data never falls into the wrong hands.
The device, called the P250Q-M80, is aimed at sectors where top-tier data security isn’t just a perk, it’s a necessity. We’re talking defense, industrial automation, AI development, and maybe even the crypto wallet you hold.
TeamGroup showcased the drive at Computex 2025, where it took home a Best Choice Award in the cybersecurity category.
Directly connected to bone, the leg allows wearers to climb stairs, walk at a normal speed, and kick balls.
The hand could lead to advanced prosthetic hands for amputees or robots with the dexterity and strength to perform household task.
Brain injuries can result in significant damage to the cerebral cortex, and restoring the cellular architecture of the tissue remains challenging. Here, the authors use a droplet printing technique to fabricate a simplified human cerebral cortical column and demonstrate its functionality and potential for future personalized therapy approaches.
Researchers at the University of Colorado Anschutz Medical Campus have discovered that amniotic fluid stem cells can be safely collected from vaginal fluid after childbirth rather than relying on more invasive methods that can pose some risk to the mother and fetus.
“We can then turn those cells into beating heart cells and use them later in treating congenital heart defects,” said the study’s senior author Jeffrey Jacot, Ph.D., associate professor of pediatrics and bioengineering at the University of Colorado Center for Bioengineering in the CU School of Medicine. “These results allow for an expanded and readily available source of amniotic stem cells beyond traditional collection through amniocentesis.”
The study was published today in the journal Stem Cells Translational Medicine.