Inference is now accepting submissions submit your article


Dispatches

The European Space Agency has released Euclid's first substantial survey dataset, containing observations of more than a billion celestial objects and approximately 15,000 strong-lensing candidates. The release supplies material for testing cosmological models; it does not yet establish new conclusions about dark matter or dark energy ...Read more

Google DeepMind has introduced AlphaGenome 2, trained on more than 100 million cellular profiles across 68 tissues. Its predictions connect DNA variation with gene regulation in specific cell types, potentially improving disease research. Performance claims await broader independent validation ...Read more

The European Space Agency has released Euclid's first major survey dataset, containing observations of more than 900 million galaxies and approximately 15,000 strong gravitational-lensing candidates. Covering 1,900 square degrees, the release provides an extensive new observational foundation for investigating dark matter and dark energy ...Read more

Recent Articles


Twin Universes and Bimetric Gravity [long_description] Jean-Pierre Luminet traces the idea of twin universes from Andrei Sakharov’s cosmological symmetries to quantum cosmology, bimetric gravity, and theories involving negative mass. Distinguishing these traditions, Luminet examines the mathematical and observational demands placed on two-metric theories and subjects the Janus cosmological model and its proposed alternatives to black holes to scrutiny, arguing that imaginative cosmological narratives must ultimately be supported by consistent equations, stable dynamics, and testable predictions.
Imagine an Organism [long_description] David J. Lobina takes up Blaise Agüera y Arcas’s What is Intelligence?, which argues that advances in AI have resolved longstanding questions about intelligence, language, and cognition. Drawing on research into infant development, linguistics, and cognitive science, Lobina challenges the book’s identification of intelligence with prediction and its parallels between neural networks and the human brain, arguing instead that modern AI and human cognition operate in fundamentally different ways.


Letters to the Editors



From the Archives


Endmark

Copyright © Inference 2026

ISSN #2576–4403