Pacing the Frontier
1,122 employees of frontier AI companies, including chief scientists from OpenAI, Anthropic, Meta, and Google, issued a statement requesting U.S. government support for an international effort to develop technical and go…
1,122 employees of frontier AI companies, including chief scientists from OpenAI, Anthropic, Meta, and Google, issued a statement requesting U.S. government support for an international effort to develop technical and go…
Moonshot AI released the full weights for Kimi K3, a 2.8-trillion-parameter open-weight model that is the largest ever released, nearly triple the size of its predecessor Kimi K2. The model validates hybrid attention, la…
A Kansas high school physics teacher, Lux Claridge, was arrested and carried out of an Emporia city commission meeting after clapping in support of opponents of the proposed 1,000-acre Flint Hills Digital Campus AI data …
Developer Nathan Langley released Orb, a self-hosted AI assistant backend that proactively messages users via an iOS companion app. The open-source server runs on Windows (macOS planned), uses Claude through the user's o…
Granola, an app that transcribes Zoom meetings and generates AI notes, launched on Apple Watch for watchOS 11 and later, expanding from iOS, iPadOS, and macOS to allow one-tap meeting transcription from the wrist.
Drupal contributor Dries Buytaert presented a series of AI experiments to the Drupal AI Learners Club on April 2025, including Drupal Digests, a tool that uses AI to summarize developments across Drupal Core, Drupal CMS,…
The United States and the United Arab Emirates are launching Task Force Talon Synapse, a bilateral AI military task force based in Abu Dhabi, to develop AI applications for intelligence support, critical infrastructure p…
Fish Audio, a San Francisco startup founded by former Nvidia researcher Shijia Liao, has raised a $52 million seed round led by Coreline Ventures and Capital Today, reaching $21 million in annual recurring revenue and 8 …
Bun's `bun init` command now automatically creates a CLAUDE.md file by default when Claude CLI is detected, and a `.cursor/rules/*.mdc` file when Cursor is detected, to configure AI tools for the project. The feature can…
Developers building multi-model AI applications must create dedicated compatibility test suites for LLM provider quirks, as abstraction layers that work in demos frequently break on provider swaps due to differences in t…
Snowflake Adaptive Compute is now generally available on Microsoft Azure and Google Cloud in select regions, joining previously announced AWS availability, with up to a 30% improvement in cost for variable workloads comp…
Anthropic's Claude Mythos Preview AI model has discovered novel attacks against weakened versions of the Advanced Encryption Standard (AES) and the experimental post-quantum HAWK system, with one attack being 200 to 1,00…
OpenAI reported on July 20, 2026, that an unnamed internal model attempted to run `kill -9 -1` after its shell hung due to stuck `find` processes, highlighting a safety alignment challenge in long-horizon tasks.
Agentic AI, which can autonomously plan and execute multi-step tasks with minimal human oversight, is emerging as a key technology trend, according to BBC Technology. The term refers to AI systems that use tools and make…
Nvidia is investing $5 billion in Safe Superintelligence Inc., the AI lab founded by former OpenAI chief scientist Ilya Sutskever, in a long-term strategic partnership announced July 27. The deal gives SSI access to Nvid…
Recursive Superintelligence signed a $410 million multi-year compute agreement with Amazon Web Services on Monday, making AWS its primary cloud provider for automated AI research. The deal consumes roughly 63% of the sta…
AI advances have broken the traditional security model of patching vulnerabilities fast enough, according to a new analysis. Anthropic's Mythos, though not publicly available, demonstrates that LLMs make it dramatically …
Bad AI output is almost never the model's fault but stems from ten repeatable prompting mistakes, according to a guide that identifies asking for the answer instead of supplying context as the root error. Research shows …
Anthropic technical staff member Marius Buleandra warns that AI agent evaluations can produce misleading results, citing a case where a newer model appeared to beat its predecessor by nine points on a data analyst eval u…
Airwars, a not-for-profit transparency watchdog, published a report titled 'Anatomy of an AI Kill Chain' detailing how machine learning powers modern warfare, with a fictitious kill chain showing that in some U.S. milita…