A research team at MIT has developed a groundbreaking technique to reorganize 40,000 atoms within a solid crystal at room temperature, a feat previously requiring extreme cold and vacuum conditions. This method achieved picometer-precision atomic rearrangement in just 40 minutes.

Published in Nature, the research involves using an algorithm-controlled electron beam to displace entire columns of atoms within a 13-nanometer chromium bromide sulfide semiconductor crystal, rather than just moving atoms on a surface. This process creates stable three-dimensional atomic structures, generating vacancy-interstitial pairs that act as functional "defects" in quantum physics, potentially serving as qubits or ultra-precise magnetic sensors.

Unlike conventional materials where defects are problematic, in quantum physics, precisely placed defects can be highly valuable. The ability to create and stabilize these defects within the crystal, protected from the environment, opens the door to manufacturing quantum materials under ambient conditions, moving atomic engineering closer to a manufacturing system.

While currently limited to a single material and requiring advanced electron microscopes, and not yet resulting in a functional device, this advancement is significant for its ability to reorganize solid matter in three dimensions at room temperature without destroying the crystal. The team aims to explore the effects of millions of deliberately organized defects, potentially leading to the design of entirely new artificial matter with emergent physical properties not found in nature.

👉 Read the full article on Gizmodo en Español

Anthropic has unveiled Claude Fable 5, its fifth-generation AI model designed for the most ambitious knowledge work and coding problems. This model is engineered to tackle complex, asynchronous tasks that can span several days, which previous models struggled to sustain.

Claude Fable 5 can operate as an agent for days, capable of planning across stages, delegating to sub-agents, and self-checking its work. For coding, it handles ambitious projects like large migrations, complex implementations, and multi-day autonomous sessions, writing its own tests and using vision to verify outputs against goals.

In enterprise workflows, it manages complex, multi-stage knowledge work with minimal oversight, from deep research to review-ready deliverables. Its vision capabilities allow it to understand diagrams, charts, and tables embedded in files and PDFs, enhancing document-heavy tasks in finance, legal, analytics, and architecture.

For safety, Claude Fable 5 includes robust safeguards for cybersecurity and biology, automatically routing flagged queries in these domains to Opus 4.8. The model demonstrates state-of-the-art performance across various benchmarks, including CursorBench, FrontierBench, and ViBench, notably achieving the highest performance in "Vibe Coding."

👉 Read the full article on Anthropic

(Note: The original article was inaccessible. The following summary is based on the article title and general industry knowledge.)

GM is significantly expanding its business scope by making a substantial investment in the energy storage solutions market, targeting both AI data centers and the broader electrical grid. This strategic move is seen as an effort to leverage its expertise in electric vehicle (EV) battery technology and enter new growth sectors.

The rapid advancement of AI has led to a dramatic increase in power demand from data centers, making stable and sustainable energy sources crucial. GM's entry aims to address this escalating demand, enhancing the operational efficiency and reliability of data center infrastructure.

Furthermore, energy storage for the grid facilitates the integration of renewable energy and contributes to stabilizing power supply. GM is positioning itself to become a key player in energy management and infrastructure beyond the automotive industry, indicating an intensifying competitive landscape in this sector.

👉 Read the full article on TechCrunch