Stay informed with the latest developments in artificial intelligence. Explore news, updates, and in-depth analysis on AI technologies, trends, and breakthroughs.
Oracle has begun laying off thousands of its 160,000-person workforce as the $420bn business software giant accelerates spending on datacentres to compete with Alphabet and Amazon — and fund a $300bn deal with OpenAI. About 10,000 employees are reported to have already lost their jobs.
A CMS misconfiguration exposed nearly 3,000 internal Anthropic assets, including a draft blog post describing Claude Mythos — a new model tier above Opus that the company itself warns is 'far ahead of any other AI model in cyber capabilities.' Anthropic has confirmed the model exists.
At a post-money valuation of $852 billion, OpenAI's latest round — anchored by SoftBank, Amazon, NVIDIA, and Microsoft — positions the company to scale compute, unify its product surface into an AI superapp, and reach $2 billion in monthly revenue.
A 60MB source-map file included in Claude Code v2.1.88 exposed 1,906 proprietary TypeScript source files on the public npm registry — the same packaging oversight that struck Anthropic in February 2025.
JustPaid's CTO replaced most of his dev workflow with seven AI agents running on OpenClaw and Anthropic's Claude Code — delivering 10 major features in a month and training a new human hire almost entirely through AI.
Microsoft's Copilot Cowork is now live in Frontier early access, bringing agentic task handling to Microsoft 365 alongside two major Researcher upgrades: a dual-model Critique layer and a side-by-side Model Council.
The South Korean fabless AI chip startup has now raised $850M total — $650M of it in the last six months — as it readies for a public listing and accelerates expansion into the US, Middle East, and Asia with two new inference infrastructure products.
The French AI lab is financing a new Nvidia-powered data center in Bruyères-le-Châtel, targeting Q2 2026 operability — part of a broader push to deploy 200MW of European compute capacity by 2027.
Mistral just released Voxtral 4B TTS 2603 — a fast, open-weights text-to-speech model with 20 preset voices, 9 languages, 70ms latency, and production-grade performance that runs on a single GPU. Here's everything it can do, who it's built for, and why it matters.
Hidden features in recent Gemini for Business builds point to deeper NotebookLM integration, a pre-built skills library from Google, and a no-code Skill Architect tool — moves that would bring the Business tier closer to parity with Enterprise and push directly into Microsoft's turf.
Pro and Max subscribers are canceling in visible numbers after discovering that Claude's peak-hour throttling leaves free users more functional than those paying $20–$200 a month — with no in-app warning, no token counter, and no published usage budgets. A viral Reddit open letter is demanding answers.
A peer-reviewed Stanford paper published in Science finds that AI sycophancy doesn't just flatter users — it makes them less likely to apologize, more morally rigid, and increasingly dependent on machine validation over human feedback. The researchers say it's a safety issue that needs regulation.