Today in AI
A catch-up on the week's biggest AI stories, from Meta's GitHub breach to open-weight models and AI laptops.
October 17, 2026 · 6 stories
On October 13, Meta disclosed that attackers used a stolen employee's credentials to access some of its GitHub repositories, taking product source code and personal data affecting about 1,500 people, mostly US employees. Meta says about 99% of the taken material was business data, the personal data was stored in the repositories in error, and its internal systems were not breached.
Source: CybernewsOn October 14, YouTube began rolling out labels on YouTube Music for tracks created primarily using generative AI, with partner disclosures of Fully Gen AI, Partly Gen AI, or No Gen AI plus its own detection tools. Tapping the label opens a panel explaining AI's role; the labels will later expand to the main YouTube app.
Source: TechRepublicOn October 14, Anthropic told prospective investors its third-quarter revenue reached $16.8 billion, up about 45% from Q2, with adjusted profitability for a second consecutive quarter, according to the Financial Times. The company is preparing for a possible November IPO that bankers believe could value it around $2 trillion.
Source: SiliconANGLEOn October 15, OpenAI updated the ChatGPT Chrome extension so Chat mode can read across open tabs to answer questions, while Work mode can fill forms, send messages, download files, and organize tabs, with per-site permission prompts. The extension works in Chrome and other Chromium browsers across all ChatGPT plans.
Source: Particle NewsOn October 16, Reflection released the weights for Beam, a 501-billion-parameter mixture-of-experts model with 23 billion active parameters and a 1-million-token context window, on Hugging Face under the Apache 2.0 license. The company says it matches larger open models on reasoning while using 3 to 4 times less inference compute, a claim not yet independently verified.
Source: Hugging FaceOn October 16, NVIDIA's RTX Spark laptops went on sale from ASUS, Dell, HP, Lenovo, Microsoft, and MSI, starting at $2,599. Each machine pairs a Blackwell RTX GPU with a 20-core Grace CPU and up to 128GB of unified memory, rated at up to 1 petaflop of FP4 AI performance for running large models and agents locally.
Source: WebProNewsTimelines are composed each morning from the previous day's verified news — every story links to the source it was checked against. See the full feed on the news page.