Saturday, October 10, 2026
AI stories selected by our newsroom, explained in about a minute, with their sources.
Google's new image model costs about half as much per picture on the Gemini API, but input and text tokens now cost more.
· AI voice, from Notebookcheck, FourWeekMBA, XenoSpectrum
Archive images: Max Liebermann self-portrait (1925), Jules Beau (1902), NASA, Library of Congress, U.S. Army; newspaper pages from 1910 and 1920 (public domain, via Wikimedia Commons); Music: “Lofi HipHop Melody Loop 99 BPM.wav” by holizna (CC0, Freesound); Sound effects: Freesound (CC0)
Test edits, a tampered citation tool, a failed proxy attempt and millions of requests: what Wikimedia found, and what OpenAI says.
· AI voice, from Help Net Security, The Next Web
Archive photos: U.S. National Archives, Library of Congress (National Photo Co.), NASA; newspaper pages from 1910 and 1920 (public domain, via Wikimedia Commons); Music: “delicate melancholic cinematic underscore, unders...” by shqiptarbas (CC0, Freesound); Sound effects: Freesound (CC0)
OpenAI put 722 math manuscripts from an unreleased model on GitHub. Fewer than a quarter have formalized main results, and none are peer-reviewed yet.
· AI voice, from Gadget Review, Unite.AI, XenoSpectrum
Music: “Minimal Vintage Relaxing Meditative Ambient Piano...” by bassimat (CC0, Freesound); Sound effects: Freesound (CC0)
Our editorial selections come first, followed by our articles and stories reported by multiple outlets. Scripts are written with AI assistance from the cited sources. Automated checks compare names, numbers and quotations with those texts; they do not replace human verification. Voices are synthetic, and each video's credits identify the tools and media used. Archive images in the collage are decoration, not depictions of the people in the story.
