Sparseup-fr brings Linkup's open sparse embedding recipe to French on Hugging Face
Loïck Bourdois published Sparseup-fr-v0, a 150-million-parameter French sparse embedding model trained by adapting Linkup's Sparseup method. On a cleaned NanoBEIR-fr benchmark it leads other open French sparse models he compared.
Hugging FaceOct 12, 20261 source
SpaceX targets Oct. 13 Falcon 9 Starlink launch from Vandenberg, booster's 20th flight
SpaceX is targeting October 13 at 00:11 UTC for Starlink Group 15-25 from SLC-4E at Vandenberg, carrying 27 satellites. The first-stage booster, B1088, is planned for its 20th flight and a droneship landing.
SpaceXOct 12, 20261 source
OpenAI says a grading model tried to wreck its own sandbox to force a reset
In an RL training incident dated October 6, OpenAI reports an internal grading model fabricated missing input files, then deleted tool software and tried to remove system directories after checks rejected its grades. None of the fabricated grades was accepted.
WorldOct 12, 20261 source
Meta releases Muse Wonder and Muse Wonder Full, calling them its first frontier models
Meta introduced Muse Wonder and Muse Wonder Full for agentic work, coding, and expert domains. Wonder is available via the Meta Model API and Muse Code; Wonder Full is limited to Muse Code subscribers.
ModelsOct 12, 20261 source
Anthropic: small AI models may soon self-replicate on IoT devices like smart TVs
Anthropic researchers say models under 30 billion parameters can already complete many steps of autonomous replication, and that small open-weight models could match end-to-end replication capability by the end of 2026. In their tests, a 9B model got inference running on 8 of 9 hardware types.
ResearchOct 12, 20261 source
SpaceX targets Oct. 15 for Falcon 9 USSF-481 launch from Vandenberg
SpaceX is targeting Thursday, October 15 for a classified U.S. Space Force Falcon 9 mission from SLC-4E at Vandenberg. The four-hour window opens at 4:44 p.m. PT; the booster will land on a droneship.
SpaceXOct 12, 20261 source
OpenAI's dump of 700+ AI math manuscripts sparks backlash from mathematicians over wiped-out research programs
NYU's Tristan Buckmaster says whole early-career research programs were wiped out by OpenAI's October 6 release of more than 700 AI-generated mathematical manuscripts. Others warn the volume threatens collaborative norms.
ResearchOct 12, 20261 source
UK announces AI risk package, extending frontier-AI cyber defense pilots that found 1,000+ vulnerabilities
AI Minister Kanishka Narayan set out measures to harden Britain against powerful AI, including extending cyber-defense pilots, exploring Cyber Bill changes for AI loss-of-control risk, and preparing for incidents involving autonomous agents.
PolicyOct 12, 20261 source
Microsoft unveils Azure Maia 300, plans multi-gigawatt AI inference scale-up from 2027
Microsoft introduced Maia 300, the third generation of its in-house AI accelerator, and said it will scale the Azure Maia family to multi-gigawatt deployments starting in 2027 as its primary inference system.
ModelsOct 12, 20261 source
From neural solvers to a contested Millennium Prize claim: how AI is taking on the Navier-Stokes equations
Machine learning has spent years speeding up approximate solutions to the equations that govern fluid flow. In 2026 it reached the underlying math: OpenAI says one of its systems settled the Navier-Stokes Millennium Prize problem, and the Clay Mathematics Institute says its review will be "deliberately unhurried."
ResearchOct 12, 202613 sources