Google just released MedGemma 27B Multimodal and a 400M-parameter image encoder called MedSigLIP for health AI development. Both run on a single GPU—with lightweight variants even fitting on mobile devices—to help researchers build private, tailored medical tools without relying on closed APIs.
California lawmakers just revived their push to regulate frontier AI with a toned-down bill called SB 53. Introduced by State Senator Scott Wiener, it would mandate leading labs like OpenAI and Google to publish safety protocols and report major incidents, while dropping developer liability rules that killed the previous SB 1047.
Google just updated Veo 3 inside its Flow filmmaking tool, allowing creators to generate full video and spoken dialogue directly from a single starting frame. The update dramatically improves character consistency by letting users prompt dialogue and actions from a single image. Why it matters: AI video is closing the gap toward full cinematic production, though audio generation remains experimental.
OpenAI is quietly testing a new 'Study Together' feature with select ChatGPT subscribers. Instead of just handing over answers, the mode asks users follow-up questions to prompt critical thinking, serving as an apparent counter to Google's LearnLM. It could mark a shift to discourage AI cheating and turn the bot into a real tutor.
Google just launched its Veo 3 video generation model globally for all Gemini Pro subscribers, expanding access across Europe, India, Indonesia, and more. Members get 3 video generations per day that replenish daily. It marks a big step toward putting frontier video creation tools directly into everyday consumers' hands.
ChatGPT referrals to news publishers have surged past 25 million in 2025—up over 25x from under 1 million a year earlier, according to Similarweb. But with organic search traffic plunging from over 2.3 billion visits to under 1.7 billion as Google's AI answers take over, the bump is hardly enough to offset publisher losses.
Google Arts & Culture just launched Moving Archives, teaming up with the Harley-Davidson Museum to bring century-old archival photographs to life using its Veo video generation model. Visitors can toggle between the original still photos and subtle AI animations, accompanied by Gemini-generated text and audio commentary. It offers a fresh look at how generative video can make dusty historical archives feel remarkably tangible.
OpenAI just confirmed it is testing Google's TPUs, but shut down rumors of any large-scale rollout. While the company continues to rely heavily on Nvidia GPUs alongside AMD and CoreWeave, it is also pushing ahead with an in-house chip expected to tape out later this year.
Google just rolled out Google Vids to all Workspace for Education users alongside 30+ free Gemini AI tools embedded directly in Google Classroom. Teachers can now whip up lesson plans and differentiated materials on the fly, with interactive NotebookLM study guides and student-facing Gems coming soon. It essentially turns schools' standard daily software into a built-in AI tutor without extra costs.
OpenAI has reportedly started leasing Google's TPU chips to handle inference for ChatGPT and other services. The move marks its first major compute shift beyond Nvidia hardware, aiming to rein in spiraling inference costs and diversify a single-vendor supply chain.
Google confirmed it is tapping YouTube's vault of 20 billion videos to train flagship models including Gemini and video generator Veo 3. While creators can opt out of third-party AI scraping, they have no way to stop Google from using their uploads. It underscores a growing flashpoint: creators unwittingly training commercial tools that could one day replace them.
Google just launched Search Live in its mobile app, letting users have free-flowing voice conversations with Search powered by a custom Gemini model. It streams spoken answers, shows web links on-screen, and keeps running in the background. It basically turns search into an active audio call, though it's currently limited to U.S. Labs users.
Pope Leo XIV just made the threat of AI to humanity a signature issue of his papacy, pushing back against tech leaders who spent years courting Rome. While giants like Google and Microsoft pitch the promises of AI, the Vatican is pushing for a binding international treaty to protect human dignity, justice, and labor.
YouTube just announced it will integrate Google's flagship Veo 3 AI video generator into YouTube Shorts later this summer. The platform previously relied on an older Veo model for background generation, but Veo 3 brings significant upgrades directly to mobile creators. With Shorts averaging over 200 billion daily views, YouTube is betting big on generative video to beat TikTok and Reels.
Canva just introduced Create a Video Clip, bringing Google's Veo 3 model straight into its design workspace. Users can now generate 8-second cinematic video clips with synchronized audio from a single text prompt, then immediately tweak them in the Canva editor. Paid subscribers get an initial quota of five video generations per month.
California experts just released a 52-page blueprint to replace the vetoed SB 1047 bill, led by Stanford's Fei-Fei Li and other researchers. The report warns that AI companies can't objectively evaluate their own risks, urging mandatory third-party audits, whistleblower protections, and legal safe harbors for independent safety researchers.
Google just released Gemini 2.5 Pro and Flash as stable models while introducing a preview of 2.5 Flash-Lite. Flash-Lite is pitched as the fastest and most cost-effective 2.5 model yet, packing a 1M-token context window, adjustable thinking budgets, and lower latency for high-volume tasks. Why it matters: Google is aggressively undercutting API costs for high-throughput production workloads.
Adobe just launched its Firefly generative AI app on iPhone and Android, bringing text-to-image, text-to-video, and Generative Fill to mobile devices. The app even lets users tap into third-party models like Google's Imagen and Veo, with all projects automatically syncing back to Creative Cloud desktop and web workflows.
Cursor just launched a new $200/month Ultra plan packing 20x more usage than Pro, tailored for heavy coders who want predictable bills instead of pay-as-you-go spikes. The company is also upgrading Pro by shifting from request caps to compute limits, unlocking unlimited tool calls and unlimited access to its Auto model.
AI pioneer Geoffrey Hinton just warned that mundane intellectual labor is on the verge of being entirely replaced, calling mass job displacement the biggest immediate threat. While one worker with AI could soon replace ten office staff in fields like legal and call centers, physical trades like plumbing remain safe for now.
