Why has it taken so long to return to the moon?
The long gap between Apollo and renewed lunar missions is primarily due to shifting political priorities, inconsistent funding, and lack of sustained strategic purpose, rather than technological limitations. Competing national interests, changing administrations, and budget constraints repeatedly disrupted lunar ambitions. Artemis aims to overcome these challenges through international and commercial partnerships.

original article.

AI maps science papers to predict research trends two to three years ahead
Artificial intelligence combining large language models and machine learning can systematically analyze scientific literature, mapping concept relationships to predict emerging research trends two to three years in advance. This approach highlights novel topic combinations and supports researchers in identifying innovative directions and interdisciplinary opportunities within rapidly expanding fields.

The number of scientific papers is growing so rapidly that scientists are no longer able to keep track of all of them, even in their own research area. Researchers from the Karlsruhe Institute of Technology (KIT), in collaboration with scientific partners, have shown how new research ideas can still be obtained from this wealth of information. Using artificial intelligence (AI), they systematically analyzed materials science publications to identify potential new avenues of research. Their results have been published in Nature Machine Intelligence.

Thomas Marwitz et al, Predicting new research directions in materials science using large language models and concept graphs, Nature Machine Intelligence (2026). DOI: 10.1038/s42256-026-01206-y

Language processing requires rapid cross-talk across brain regions, researchers discover
Language processing involves rapid, coordinated activity across multiple brain regions rather than a single area. Concrete words engage both sensory and language regions, while abstract words rely more on language-related areas. Brain responses to words of varying concreteness remain stable across individuals, and disrupting different regions impairs word classification, indicating distributed processing.

Multiple regions of the brain engage in fast-moving conversations to understand language, researchers have discovered, dispelling a prior school of thought that only one region of the brain was responsible for language processing. The research was published in PLOS Biology.
The team found that concrete words activated regions of the brain that process sensory experiences and regions responsible for language, while abstract words relied more heavily on language-related areas of the brain. For words that fell in between, the team found that the patients' brain responses were stable regardless of individual, subjective ratings.
Even if a person thinks of the word 'magic' in purely physical terms, their brain seems to still activate some of the abstract features associated with the word 'magic.'"

Additionally, researchers found that whether the participants were reading purely abstract or purely concrete words, multiple regions of the brain communicated with each other to process them.
In a separate part of the study, researchers asked participants to classify ambiguous words while they stimulated different parts of the brain with small electrical pulses to temporarily disable their processing. When different regions were stimulated, participants had a harder time making decisions about how to classify the words, reaffirming that multiple areas are responsible for decoding language.

The research has important clinical implications for patients with aphasia, or the inability to speak, as well as dementia and brain injuries.

Elliot Murphy et al, Frontotemporal network interactions causally support rapid concreteness judgments during reading, PLOS Biology (2026). DOI: 10.1371/journal.pbio.3003723