Weekly briefing October 26–29, 2026
A concise selection of this week’s science and technology, with original sources and context attached to every story.
The weekly selection
Coverage through Thursday, October 29. Summaries distinguish published findings from company claims and early candidates.
Google announced Gemini 4 Carbon, a model it says improves coding and cyber-defense tasks, including vulnerability discovery, exploit testing, and patching. The first rollout is limited to selected cyber defenders while safeguards are tested; the benchmark scores are company-reported.[1]
Read Google’s announcement
Mistral released the open weights for its 1-trillion-parameter Large 4 model, so organizations can run it on their own infrastructure. That offers more deployment control, while Reuters notes that public weights are also harder to contain.[2][3]
Read Mistral’s releaseOpenAI says its models produced results in mathematics, computer science, and physics, including a claimed refutation of the Exponential Time Hypothesis. The computer-science proof is public for review; the physics group is holding its materials while it prepares a write-up. Independent researchers are still assessing the claims.[4][5]
Read OpenAI’s announcement
After more than 2.2 billion miles, NASA says Lucy has entered the region around Jupiter’s L4 Trojan asteroids. Its first close asteroid encounter is planned for 2027; studying these primitive bodies may help refine theories of solar-system formation.[6][7]
Read NASA’s update
The Vera C. Rubin Observatory identified 2026 ED138, a trans-Neptunian object beyond Pluto, as a dwarf-planet candidate. Early estimates put it about 46 AU from the Sun and 600–1,500 km across, but follow-up observations are still needed to confirm its size and status.[8][11]
Read Rubin Observatory’s announcementA Nature paper reports that MCT1 can transport lactate into mitochondria in heart cells, providing a parallel route for energy production. In mouse studies, loss of MCT1 impaired the heart’s ability to adapt to stress; the finding points to a research target, not an established treatment.[9][10]
Read the Nature paperPrimary research and official announcements are preferred; company claims and candidate classifications are identified in the summaries.