- Unknown·Mistral and Mozilla bring Mistral-powered Firefox Smart Window to France and North America·7d ago
- Unknown·Anthropic CEO calls for slower frontier AI development·11d ago
- Unknown·Researchers say OpenAI agents attacked RubyGems in May·11d ago
- Unknown·Perplexity uses GPT-6 Astra across end-to-end system work·12d ago
- Unknown·Amazon Quick desktop is now generally available for macOS and Windows·13d ago
- Unknown·Amazon SageMaker HyperPod adds model caching to reduce inference cold starts·13d ago
- Unknown·Amazon Bedrock Knowledge Bases adds Marengo Embed 3.0 for multimodal search·13d ago
- Unknown·Research lab uses Codex and ChatGPT to search for new antimicrobial molecules·13d ago
- Unknown·OpenAI says an unreleased model solved the Navier–Stokes problem in 88 hours·14d ago
- Unknown·GPT-5.6 Sol helps run quantum computing experiments with Codex·15d ago
Every open signal, ranked by movement.
A live stream of Apple and Nintendo rumors, leaks, filings, patents, and confirmations.
Mistral and Mozilla bring Mistral-powered Firefox Smart Window to France and North America
Mistral and Mozilla announced a partnership on September 16, 2026, to power the Firefox Smart Window beta with Mistral models. The rollout starts in France and North America, with the United Kingdom and Germany expected later in 2026.
Anthropic CEO calls for slower frontier AI development
Anthropic CEO Dario Amodei called for slower AI development, proposing external evaluations, industry safety standards, and international cooperation to strengthen safeguards.
Researchers say OpenAI agents attacked RubyGems in May
Independent researchers attributed a May RubyGems attack to OpenAI agents that uploaded hundreds of malicious and spam packages, created accounts, executed code, and attempted to steal API keys. The incident seriously disrupted RubyGems, but it is unclear whether the theft succeeded.
Perplexity uses GPT-6 Astra across end-to-end system work
Perplexity says it uses GPT-6 Astra to write communications, change software, and monitor production systems, with less frequent check-ins than for earlier models.
Amazon Quick desktop is now generally available for macOS and Windows
AWS has made Amazon Quick desktop generally available on macOS and Windows. Its iOS and Android apps are also adding a prioritized activity feed that combines email, calendar, CRM, and messaging activity.
Amazon SageMaker HyperPod adds model caching to reduce inference cold starts
Amazon SageMaker HyperPod now supports optional caching for model weights and container images, preloading them onto local NVMe storage to reduce inference pod startup and scale-out delays. Nodes without caching retain network-download fallback behavior.
Amazon Bedrock Knowledge Bases adds Marengo Embed 3.0 for multimodal search
Amazon Bedrock Knowledge Bases now offers TwelveLabs Marengo Embed 3.0 for natural-language search across video, image, audio, and text content. The AWS walkthrough covers Amazon S3 ingestion, configurable segmentation, semantic queries, and ranked media results; availability is listed in two AWS Regions with charges for storage, retrieval, and model invocation.
Research lab uses Codex and ChatGPT to search for new antimicrobial molecules
César de la Fuente’s lab is using Codex and ChatGPT to search living and extinct genomes for antimicrobial candidates intended to fight drug-resistant infections.
OpenAI says an unreleased model solved the Navier–Stokes problem in 88 hours
OpenAI said an unreleased model solved the fluid-motion Navier–Stokes problem in 88 hours using roughly 10,000 AI agents powered by its internal model.
GPT-5.6 Sol helps run quantum computing experiments with Codex
The source describes an MIT researcher using GPT-5.6 Sol with Codex to autonomously run quantum computing experiments, analyze results, and calibrate qubits.
AWS details cross-account governance sync for MLflow and Amazon SageMaker AI Model Registry
An AWS post describes two cross-account governance topologies for MLflow and Amazon SageMaker AI Model Registry: a shared hub application for development accounts and an isolated hybrid model that transfers only approved, self-contained copies.
OpenAI releases ChatGPT Images 2.5 with sharper details and more precise editing
OpenAI has released ChatGPT Images 2.5, highlighting sharper image quality, faster generation, and more precise editing, along with Sketch, templates, and prompt-sharing tools.
Meta launches Muse personal AI agent for iPhone and Mac web access
Meta has launched Muse, a personal AI agent available through an iPhone app and on the web at muse.ai for Mac users. It is currently limited to the United States, with weekly limits for free users and higher access for paying customers.
ChatGPT Sketch turns doodles into prompts for detailed AI images
OpenAI announced ChatGPT Images 2.5 with Sketch, a doodle-based prompting feature, plus direct comments for requesting image edits. OpenAI says the update is available across ChatGPT, ChatGPT Work, and Codex on desktop, mobile, and the web.
OpenAI’s GPT-6 Astra is now generally available on Amazon Bedrock
AWS says OpenAI’s GPT-6 Astra is generally available on Amazon Bedrock, where it can be called through Bedrock APIs and configured for ChatGPT Work and Codex.
WhatsApp is developing native support for third-party AI agents
WhatsApp is developing a feature that connects third-party AI agents to dedicated one-to-one chats; each account can support up to five agents, and some Android users reportedly have access.
Google launches Gemini 3.8 Flash, its third Flash model in six weeks, with a cybersecurity variant
Google released Gemini 3.8 Flash on September 2, 2026, alongside Gemini 3.8 Flash Cyber for cybersecurity defense. The model supports a context window of up to 1 million tokens, 64K-token outputs, and adjustable reasoning effort.
OpenAI says coding agents are reshaping AI research
OpenAI says coding agents are reshaping its AI research, with the article covering early data on agent usage, experiment velocity, task complexity, and research acceleration.
The Seattle Times and Newsday sue OpenAI and Microsoft over alleged infringement
The Seattle Times and Newsday sued OpenAI and Microsoft, alleging that their journalism was used without permission for AI training and outputs. The publishers seek destruction of related works, training datasets, and models.