Today in AI
Muse gets a voice and a face, Apple ships five new products today, and Claude starts reading bank statements.
November 6, 2026 · 8 stories
Apple's fiscal 2026 10-K, filed November 3, adds a new risk factor saying AI agents can act without authorisation, exceed permissions, defy directions and conceal actions, with some consequences 'difficult or impossible to reverse'. It is a risk disclosure, not a report of any incident.
Source: Times of IndiaMusk announced the feature on October 25, and the first real user calls surfaced November 3-4: Grok Bot phoned a dental office to check insurance and called an ISP to cancel a contract, showing a call card for approval first. SpaceXAI has not documented whether the bot identifies itself as AI on the call.
Source: TeslaNorthMira Murati's lab introduced its second model on November 5: a 594-billion-parameter sparse mixture-of-experts model scoring 31 on the Artificial Analysis Intelligence Index. It is available now through Tinker with closed weights; full weights under Apache 2.0 are promised next week, pending external safety testing.
Source: Unite.AIAnthropic launched Personal Finance in Claude on November 5, a US beta for Pro, Max and Ultra subscribers. Accounts link read-only through Plaid, so Claude can answer questions about balances, spending and bills from up to two years of transactions; it cannot move, manage or invest money, and the feature is not designed to give investment advice.
Source: ReutersGoogle's video generation and editing model went generally available on November 5 across the Gemini API and Google Cloud. Native clips now run up to 15 seconds, up from 10, with multi-turn conversational editing, and 720p output costs $0.07 per second, 30% less than Omni 1.1 Flash. Outputs carry a SynthID watermark.
Source: Google Cloud BlogNVIDIA posted four Cosmos 3 Nano checkpoints to Hugging Face on November 5 under OpenMDW 1.1 terms. The 16-billion-parameter models generate interactive driving and robot simulation video, including synchronised camera and LiDAR observations, for training and testing physical AI without collecting every scenario in the real world.
Source: RuntimeWireMeta announced on November 5 that voice conversations and expressive avatars are rolling out in the Muse app in the US across mobile, web and Mac. Users tap the phone icon to talk, interrupt Muse mid-answer, and pick from hundreds of voices, with the avatar reacting as it listens and speaks. Voice for AI Glasses and Muse Charm is coming soon.
Source: Unite.AIFive products go on sale today, November 6, after pre-orders opened last week: an iPad mini with OLED display and A20 Pro chip ($699), an iMac with M6 chip ($1,499), a 14-inch MacBook Pro with M6 chip ($1,999), and 14- and 16-inch MacBook Pros with OLED touchscreens. The base 14-inch touchscreen model in Space Black ships in 3-4 weeks.
Source: MacRumorsTimelines 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.