According to AI News... Anthropic ships Claude Agent View while Google upgrades Android with Gemini Intelligence, plus major robotics milestones.

According to TheRundownAI, Google pushes Gemini on Android, eyes SpaceX for orbital compute, as Anthropic and Amazon reshape AI work and tools.

According to @demishassabis, Google DeepMind demoed Gemini-powered pointer controls in Google AI Studio using motion, speech, and shorthand.

According to @emollick, OpenAI says Study Mode still works via /study and /learn, though the official page lacks menu access details.

According to TheRundownAI, Google made Android a system-wide intelligence layer with Gemini-native devices and agentic cursor features.

According to @emollick, OpenAI quietly pulled Study Mode, raising concerns that default assistant answers harm learning compared to tutor-style guidance.

According to @sundarpichai, Gemini Intelligence brings cross‑app automation, one‑tap form fill, and Rambler speech to polished text on Android.

According to TheRundownAI, Google DeepMind demoed Gemini in the mouse pointer, streamlining on-screen actions and context for faster AI assistance.

According to GoogleDeepMind, experimental demos show Gemini following motion, speech, and shorthand to control on-screen tasks.

According to @godofprompt, Google is set to replace Veo with Gemini Omni, a new video model that enables chat-based video editing inside Gemini.

According to @AINewsOfficial_ and Anthropic, Claude agents debut for finance with native Microsoft 365 integration, while Figure 03 adds 9 home tasks.

According to @godofprompt, Google adds step-by-step timelines and preps mid-flight steering and async tools; old Interactions API breaks June 6.

According to God of Prompt, OpenAI and Google embed resilient pixel-level watermarks in generated images, verifiable via C2PA and Gemini checks.

According to Google Gemini on X, Personal Intelligence and Nano Banana 2 now generate imagery in Gemini, with a live demo today at 12 PM PT.

According to @godofprompt, 53 LLMs miss the Car Wash Test; adding a goal-and-prerequisites step makes Claude Sonnet 4.7 answer correctly.

According to @godofprompt, 53 LLMs failed a car-wash reasoning test; adding goal-identification to the prompt fixed Claude Opus 4.7 instantly.

According to Google Gemini on X, creators can submit Gemini Canvas projects with a big 1–10 number by May 6 for a chance to be shown at I O 2026.

According to @emollick, ChatGPT, Claude, and Gemini often cite four poems about making LLMs, signaling shared cultural priors in AI outputs.

According to @godofprompt, ex-Google Maps PM Bilawal Sidhu used Gemini and Claude to ship a Palantir-like geospatial app in days, compressing execution time.

According to The Rundown AI, Google adds Gemini to cars, White House scrutiny on Anthropic rises, and OpenAI traces ChatGPT goblin replies.