Verified news, launches and analysis to understand what is changing in artificial intelligence and why it matters to you.
166 articles
Aleph Alpha presents Kolibri as an open-weight model focused on German and English and designed for sovereign deployments. Key questions remain about its licence, context window, hardware requirements and independent evaluation.
Third-party descriptions attribute a system called Mods to Claude Code, but the available evidence does not confirm an official announcement or clarify its permissions, availability, or capabilities. Here is what should be verified before adopting it.
Mercor reported a 100% result for Claude Opus 5 in an assessment of accounting tasks resembling month-end close work. A reported comparison puts human participants at 37%. The available information does not let us verify the test’s size, scoring rubric, or execution conditions.
NVIDIA has officially announced DGX Spark with 64 GB of unified memory. The system is priced from $4,999 and is scheduled to be available through Acer, ASUS, Dell, Gigabyte, HP and MSI starting October 23, 2026.
Apple documents controls for apps’ access to files and data on a Mac. The information available here does not confirm new permissions tied to AI agents or verify the Muse case.
A news report describes a subpoena related to cybersecurity incidents involving AI agents. The available material does not confirm the Hugging Face episode, the extent of any access, or whether data was extracted or damage occurred.
Two publications describe Microsoft speech transcription and speech synthesis models. The supplied documentation confirms some relevant context, but does not independently verify the reported latency figures or the scope of support for 23 languages.
The paper says it can reduce backdoor attack success without retraining the suspect parameters and, according to its abstract, without relying on known triggers or clean references. The experimental details available here are not sufficient to independently verify those claims.
An analysis of 22 benchmarks distinguishes the reliability of comparing deployed systems from that of attributing differences to their underlying models. Its findings suggest that adding tasks does not resolve every kind of uncertainty.
Code to Control separates the generation of a controller’s structure from the tuning of its parameters. The preprint says that, once learned, the controller can act without LLM inference or planning at every step; the available evidence is not enough to assess the figures, reproducibility, or readiness for real-world robots.
The company has opened U.S. waitlists for an agent integrated with Apple Messages and for an API for bulk orders. Key details remain unclear, including payment confirmation, participating restaurants and how conversations will be handled.
Secondary news reports describe CheatBench as a benchmark for studying reward-hacking behavior. The available information helps explain the question it raises, but does not confirm the experimental details or support extrapolating its results to systems in production.