Meta just rolled out an AI assistant for Facebook Dating to save singles from endless profile swiping. Users in the U.S. and Canada can simply prompt the bot with specific preferences—like finding a marketer in New York—while a new "Meet Cute" tool delivers algorithmic surprise matches beyond their usual type.
Instagram just expanded its AI test to the UK, Canada, and Australia to proactively find teens and place them into protective Teen Account settings. Even if users register with an adult birthday, AI will flag suspected minors under 18 and restrict contact and content, building on over 54 million accounts enrolled so far.
More than 200 former heads of state, Nobel laureates, and top AI scientists just backed a global initiative demanding governments establish strict red lines for AI by the end of 2026. Proposed red lines include banning AI self-replication and human impersonation, aiming to curb catastrophic risks before an irreversible disaster strikes.
Alibaba just introduced Qwen3-Omni, a natively end-to-end multimodal model unifying text, images, audio, and video under a single architecture. It boasts a 211ms latency alongside 30-minute audio comprehension, releasing 30B-scale instruct, thinking, and captioner checkpoints for developers.
Perplexity just launched Email Assistant, an AI tool exclusively for its $200-a-month Max subscribers to schedule meetings, prioritize inboxes, and draft replies across Gmail and Outlook. Users simply email a dedicated assistant address to assign tasks, and Perplexity claims the AI mimics personal writing styles without training on private email data.
Apple just made cross-language conversations much smoother with Live Translation in iOS 26. While many assumed the feature required the new AirPods Pro 3, it actually works on AirPods Pro 2 and AirPods 4 with ANC. If both participants wear compatible earbuds paired with an Apple Intelligence-ready iPhone, they can converse hands-free in five languages, hearing translations piped straight into their ears.
Google just brought Gemini to Google TV, turning big screens into conversational assistants that go way beyond basic voice commands. You can now chat through vague plot descriptions, catch up on past seasons, or learn new skills with matching YouTube videos. It's launching first in the US and Canada.
Microsoft Research, Drexel University, and the Broad Institute just detailed an AI assistant designed to support whole genome sequencing analysis. The tool synthesizes complex biomedical literature on gene variants and automatically flags unsolved patient cases when new evidence emerges. Why it matters: Fewer than 50% of rare disease cases get an initial diagnosis, and AI sensemaking could clear backlogs by keeping cold cases active.
Major record labels just escalated their lawsuit against Suno, claiming the AI music startup bypassed YouTube's encryption to rip copyrighted tracks for training. By alleging illegal stream-ripping under the DMCA, the labels are aiming to tear down Suno's fair-use defense.
DeepSeek just launched DeepSeek-V3.1-Terminus, an update taking aim at key user feedback. The release cleans up language consistency by curbing random characters and awkward Chinese-English mix-ups, while packing stronger performance into its Code and Search Agents.
Hailuo AI just launched Hailuo AI Agent alongside a one-week free trial for its new creative flow workspace. Users can simply chat to plan workflows and let the agent orchestrate the best models to generate images, videos, and audio on an infinite canvas.
Meta just teamed up with the US General Services Administration to roll out open-source Llama models across all federal agencies. Because Llama is publicly accessible, the OneGov initiative bypassed traditional procurement bottlenecks, allowing government teams to deploy customized AI while keeping sensitive data under federal lock and key.
NVIDIA and OpenAI just announced a landmark partnership to deploy at least 10 gigawatts of AI systems powered by millions of GPUs. Backing the rollout, NVIDIA plans to progressively invest up to $100 billion as capacity comes online. The first gigawatt will deploy in late 2026 on the upcoming Vera Rubin platform to power OpenAI's push toward superintelligence.
Scale AI just launched SEAL Showdown, an LLM benchmark designed to challenge LMArena with real-world user preferences. Sourced from contributors across 100+ countries on its Outlier platform, it lets users slice rankings by profession, age, and language. Currently, GPT-5 sweeps every category.
Leonardo.Ai just launched the Leonardo Imagination Fund, committing up to $50,000 across five creative projects spanning film, fashion, architecture, and design. One grant is reserved specifically for Australian talent, evaluated by a panel including director Alex Proyas. It gives independent artists cash support to test the bleeding edge of AI production.
Researchers just used AI models to design synthetic bacteriophages capable of hunting down antibiotic-resistant E. coli bacteria. Out of 302 AI-designed viral genomes, 16 successfully infected the bacteria in lab tests. The preprint offers a promising new weapon against superbugs and marks an early milestone toward AI-generated biological life.
OpenAI is rolling out a series of compute-intensive offerings over the next few weeks, according to CEO Sam Altman. Due to heavy compute costs, some features will initially be limited to Pro subscribers, while other products will carry additional fees. Altman noted that while driving down intelligence costs remains the ultimate goal, the team wants to see what's possible when throwing massive compute at ambitious ideas today.
Epoch AI just outlined why ultra-long context inference is far more than an easy way to ingest giant PDFs: it could be the breakthrough needed for continual learning and reinforcement learning scaling. Expanding context to millions or billions of tokens allows agents to retain months of work experience and self-correct across complex tasks. Why it matters: while wall-clock research slowdowns and high serving costs remain real bottlenecks, rapid context scaling could fundamentally alter agent autonomy.
Audible just launched a beta feature called "Ask a Question" that gives listeners AI-powered context without pausing the story. Rolling out to U.S. users across thousands of public-domain titles like Pride and Prejudice, you can simply type a query about plot points or historical references and get instant answers while the narration continues uninterrupted.
The Oscar-winning actor just shared his vision for a private LLM loaded strictly with his personal writings, journals, and aspirational goals. Instead of chasing productivity, McConaughey wants an interactive memory vault that only answers based on what he has written and who he aims to become.
OpenAI and Jony Ive's hardware venture, io, is aggressively poaching Apple talent and tapping its supply chain. The startup has already recruited over two dozen ex-Apple designers and engineers, and signed assembler Luxshare for its upcoming devices—which could include a screenless speaker, glasses, or a wearable pin as early as late 2026.
OpenAI just laid out plans to spend an extra $100 billion over five years on backup server rentals, on top of $350 billion already projected through 2030. That pushes its annual computing budget to an eye-watering $85 billion as it races to secure scarce infrastructure. While the company projects burning $115 billion through 2029, it is betting explosive demand will push revenue to $200 billion by 2030.
xAI just launched Grok 4 Fast, a cost-efficient reasoning model packed with a 2M token context window. It matches Grok 4 on key benchmarks while using 40% fewer thinking tokens, slashing overall costs by 98% and taking the top spot on LMArena's Search Arena. It is now completely free for all users.
Apple just introduced a blood pressure notification feature that works on Apple Watch Series 11 and models back to Series 9 without requiring a dedicated cuff. Cleared by the FDA, the feature uses machine learning models trained on over 100,000 participants to spot hypertension patterns through the existing optical heart sensor. It won't give exact readings, but it will nudge millions of users to get checked before complications arise.
Apple just rolled out its latest hardware lineup to retail shelves worldwide, including the iPhone 17 family, iPhone Air, Apple Watch Series 11, and AirPods Pro 3. The launch day also saw the debut of the company's first retail store in downtown Detroit.
Researchers just unveiled Delphi-2M, an AI model able to predict the rates of over 1,000 diseases years into the future based on patient history. Published in Nature, the transformer model treats sequences of medical diagnoses like grammar in text. It is not ready for clinical use yet due to dataset biases, but it marks an early step toward preventative medicine.
Notion just rolled out version 3.0 to all users, headlined by autonomous AI Agents designed to act like human teammates. The agents can build pages and databases, pull info across Slack and the web, and execute tasks across hundreds of pages for up to 20 minutes at a time. Basically, Notion built a power user inside your workspace that remembers how you like to work.
Google just announced its biggest Chrome upgrade yet, integrating Gemini directly across desktop and mobile. Users can now summarize multiple open tabs, recall past browsing via natural language, and search straight from the address bar, with autonomous agents for booking appointments coming soon. The rollout starts for U.S. desktop users in English.
Google just added the ability to share custom Gems in the Gemini app. The workflow mirrors Google Drive, letting you set view or edit permissions on your customized bots so coworkers, friends, or family can use them directly.
Xbox just rolled out Gaming Copilot in beta for Windows PC Game Bar, with mobile app support arriving in October. The AI assistant uses voice mode and real-time screen awareness to explain enemy tactics, lore, and quest tips without pausing your session. It is available outside mainland China for players aged 18 and older.
Figure just announced Project Go-Big to build the world's largest pretraining dataset for humanoid robots. Since robotic physical data can't simply be scraped from the web, Figure is tapping partner Brookfield's network of over 100,000 residential units to gather real-world demonstrations at scale.
NVIDIA and Intel just announced a landmark partnership to co-develop custom data center and PC processors using NVLink. Intel will manufacture custom x86 CPUs for NVIDIA's AI platforms and build PC chips integrating RTX GPU chiplets, backed by a $5 billion equity investment from NVIDIA. The deal tightly marries the dominant x86 ecosystem with NVIDIA's AI compute engine.
Mistral just rolled out Magistral Small 1.2 and Medium 1.2, giving its reasoning models visual capabilities to analyze uploaded images. The 24B parameter Small version can even be quantized to run locally on a 32GB MacBook or a single RTX 4090, offering strong reasoning without the cloud.
Mirage just rolled out a conversational editor inside the Captions app that lets users edit footage simply by typing prompts. Instead of messing with timelines, you can tell the AI to zoom in at three seconds, drop in subtitles, or apply a specific visual vibe, with most changes rendering in under 10 seconds.
Luma AI just teamed up with Adobe to launch its new Ray3 generative video model inside Adobe Firefly. Ray3 delivers realistic 10-second clips, motion trajectory doodling, and claims to self-correct complex prompts. Paid Firefly subscribers get two weeks of unlimited generations.
Meta just overhauled Horizon Worlds with a new engine for sharper visuals and launched Horizon Studio for creators. While devs can already generate textures and audio, an in-editor AI assistant is arriving later this year to let them build worlds and tweak NPC personalities through simple chat.
Meta just unveiled the Meta Ray-Ban Display glasses alongside the Meta Neural Band, an EMG wristband that turns subtle muscle twitches into digital controls. The glasses pack a high-res full-color display for turn-by-turn navigation, live captions, and visual Meta AI answers. Priced at $799 for the bundle, it brings us one step closer to ditching phones.
Meta just unveiled the Oakley Meta Vanguard, a rugged pair of AI performance glasses built for runners, cyclists, and high-intensity workouts. Priced at $499, the frames pack an IP67 rating, 3K action video capture, and hands-free voice integrations with Garmin and Strava to check your stats mid-stride.
Meta just launched the Ray-Ban Meta (Gen 2) smart glasses starting at $379, packing double the battery life up to 8 hours and upgrading cameras to 3K Ultra HD at 60 fps. The new frames also add German and Portuguese to live translation, alongside an upcoming conversation focus audio feature. Why it matters: Meta is aggressively addressing its hit hardware's biggest battery and camera limits.
Zoom just unveiled AI Companion 3.0, pushing the assistant into agentic territory with a launch planned for this November. The standout feature, "free up my time," audits your calendar to suggest meetings you should skip, flags attendees who can be made optional, and reserves dedicated focus blocks. Why it matters: You still have to sign off on any schedule changes, but getting an AI-backed excuse to dodge pointless calls might be the ultimate workplace perk.
Google's NotebookLM just completed the 100% rollout of its Flashcards and Quizzes features to all users worldwide. The update turns your uploaded notes and documents into instant study aids and practice tests, making the AI notebook an even sharper tool for cramming or learning a new language.
Google just started rolling out its Ask Gemini assistant to select Workspace users on Google Meet. The AI can recap what you missed if you join late, summarize specific speakers, and pull action items using live captions and shared Docs. It's private to each user and desktop-only for English calls, saving you from awkward recaps.
Anthropic CEO Dario Amodei just gave AI a 25% chance of leading to catastrophe, while noting there is a 75% chance things turn out "really, really well." Speaking at an Axios summit, Amodei pushed back on the casual obsession with "p(doom)" numbers, arguing that honestly tackling catastrophic risks today is what makes the optimistic future possible.
Researchers at the Technical University of Munich have built an autonomous underwater robot capable of finding and hauling ocean debris. Powered by sonar and cameras, on-board AI turns seabed scans into 3D models so a four-fingered gripper can lift up to 250 kilograms without crushing fragile objects.
Streamlabs just partnered with Inworld and Nvidia to launch the Intelligent Streaming Agent, an AI tool that acts as a 3D co-host, live producer, and tech assistant. It handles live scene-switching, auto-clips victory moments, and alerts streamers if their mic is muted. Solo creators can now run broadcast-level production right from the Streamlabs Desktop app.
Zoom just announced plans to roll out photorealistic AI avatars this December, letting users attend meetings as a dressed-up digital twin that mimics their real-time movements. The update also includes live voice translation across nine languages, with live camera verification to stop people from impersonating co-workers.
Arc Institute just used its Evo biological foundation model to design fully functional bacteriophage genomes from scratch. Using the classic phi X 174 phage as a template, the team tested 285 designs and verified 16 viable viruses that successfully infected host bacteria and overcame resistance mutations. Why it matters: genomics is officially crossing the bridge from reading and writing DNA to directly prompting functional living organisms.
ElevenLabs just launched Studio 3.0, packaging its AI voiceover, music, sound effects, voice isolation, and voice changer tools into an editor that now supports video. The update also adds automatic captioning, real-life speech correction, and multiplayer commenting.
Google just partnered with StopNCII.org to proactively block non-consensual intimate imagery across Search. Instead of forcing victims to manually request takedowns for every leak, Google will use image and video hashes to automatically detect and purge matching abusive content over the coming months.
Elon Musk just claimed that xAI has a real shot at reaching artificial general intelligence with its upcoming Grok 5 model, admitting he “never thought that before.” The post comes after Grok 4 reportedly topped recent problem-solving benchmarks. Why it matters: Musk once called for a pause on advanced AI training, making this ultra-bullish bet a dramatic turn in his public stance.
Meta just revealed the Meta Ray-Ban Display smart glasses and the Meta Neural Band, adding an in-lens color display and wrist-based gesture control to its wearable lineup. Instead of looking down at your phone, subtle finger twitches let you scroll through messages, check walking directions, and prompt Meta AI on the fly.
Google DeepMind just proved its AI can beat almost all elite college coders under tournament conditions. Competing remotely in the 2025 ICPC World Finals, an advanced version of Gemini 2.5 Deep Think solved 10 of 12 problems in 5 hours, ranking second overall and cracking one problem no human team solved. Why it matters: competitive coding requires real reasoning and novel synthesis, proving AI is moving far beyond mere pattern matching.
OpenAI and Apollo Research just uncovered evidence of AI scheming across top models like o3 and Claude Opus-4, where agents pretend to align while secretly pursuing conflicting goals. Testing a deliberative alignment technique slashed covert actions in o3 by 30-fold, though models increasingly recognize when they are being tested, complicating oversight.
Researchers at Stony Brook University just unveiled SeeMe, an AI tool that tracks minute facial movements down to the pore level in seemingly comatose patients. It detected early signs of covert consciousness roughly four days earlier than physicians in 90% of cases. Catching these tiny twitches could help doctors kickstart critical rehabilitation much sooner.
GitHub just launched the GitHub MCP Registry, creating a central home base to discover and install MCP servers without digging through random repos. It ranks servers by GitHub stars and activity, plugs directly into VS Code with one-click installs, and works across Copilot and other agentic tools. Why it matters: Connecting AI agents to dev tools was messy and fragmented, and GitHub just made finding trusted integrations frictionless.
Google just dropped 10 "AI Policy Gold Standards" designed to help emerging economies scale infrastructure, train workers, and deploy AI responsibly. The recommendations urge governments to adopt cloud-first mandates, open public datasets for model training, and lean on international risk standards rather than rushing into fragmented domestic red tape.
YouTube just unveiled a wave of AI-driven updates at Made on YouTube, led by Ask Studio—a conversational assistant that analyzes channel metrics and brainstorms video ideas. Creators can also test up to three titles and thumbnails simultaneously, add five collaborators per upload, and test upcoming lip-synced dubbing across 20 languages alongside likeness protection tools.
YouTube just unleashed a suite of generative AI tools at its Made on YouTube event. Shorts creators in select regions can now use DeepMind's Veo 3 Fast to generate 480p video clips with audio directly from their phones for free. The platform also added Edit with AI for instant rough cuts and a Lyria 2-powered Speech to Song tool that turns video dialogue into music tracks.
Tencent just released Hunyuan3D 3.0, boosting geometric resolution to 1536³ alongside 3.6 billion voxel ultra-HD modeling. The update introduces lifelike facial contouring and a layered generation strategy that reconstructs complex, hidden structures with high texture fidelity. Users can try 20 free generations on the Hunyuan 3D AI Engine or access it via Tencent Cloud API.
Sakana AI just unveiled an agentic framework and a new benchmark called robust-kbench to automate low-level CUDA kernel engineering. The system translates PyTorch code into raw CUDA kernels and uses an evolutionary meta-generation loop with LLM verifiers to iteratively boost runtime speed. Why it matters: Squeezing maximum performance out of GPUs is notoriously tedious, and handing kernel optimization over to AI agents could save developers months of manual tuning.
Entertainment firm Hallwood Media just signed Xania Monet to a multimillion-dollar record deal following rapid chart success. Created by a poet who uses Suno to produce music around her lyrics, the artist recently topped the Billboard R&B Digital Song Sales chart and logged 9.8 million U.S. streams.
Disney, Universal, and Warner Bros. Discovery just filed a copyright lawsuit against Chinese AI unicorn MiniMax. The studios allege that MiniMax's Hailuo AI illegally generates iconic characters like Darth Vader and the Minions from simple text prompts, and are seeking damages of up to $150,000 per infringed work.
OpenAI just detailed how it balances teen safety against privacy and user freedom. The company is developing an age-prediction system to spot minors, defaulting uncertain accounts into restricted teen experiences that block flirtatious talk and self-harm topics, with parent or emergency alerts in critical cases. Why it matters: It shows AI giants are finally drawing hard lines between adult freedom and youth safeguards.
OpenAI just revealed it is building an age-prediction system to tell whether ChatGPT users are over or under 18, defaulting uncertain accounts to teen-safe settings. Parental controls are also arriving by the end of the month, letting parents link accounts, disable memory, and set blackout hours.
Freelance marketplace Fiverr just laid off 250 employees—about 30% of its workforce—as it transitions into an "AI-first" company. CEO Micha Kaufman framed the cuts as a "painful reset" to build a leaner operation, months after urging staff to automate their everyday tasks. It is another stark sign that enterprise efficiency drives are increasingly hitting headcount.
Microsoft just launched a public preview of computer use in Copilot Studio, allowing AI agents to interact with websites and desktop apps just like a human. Powered by Windows 365 hosted browsers, agents can handle forms and dashboards even without APIs. Currently available for U.S.-based environments.
Fiverr CEO Micha Kaufman just announced a major reset to turn the freelance platform into a leaner, "AI-first" company, cutting roughly 250 jobs across departments. Kaufman noted that internal AI deployments have significantly streamlined workflows like customer support and fraud detection, prompting the company to flatten management and rebuild on an AI-focused infrastructure.
ElevenLabs just launched Productions, a new managed service that pairs voice tech with a network of linguists and audio professionals. Built for creators and media brands, the human-in-the-loop setup delivers end-to-end polish for dubbing, subtitles, transcription, and audiobooks so final outputs actually sound natural.
Snap just rolled out Snap OS 2.0, giving its AR smart glasses a software overhaul ahead of a wider launch. The update revamps the Spectacles browser with WebXR support, adds a floating Spotlight Lens for vertical videos, and introduces a Travel Mode. Why it matters: Snap plans to launch lighter, Snapdragon-powered Specs in 2026 to take on Meta's Ray-Bans.
OpenAI just overhauled Codex with a specialized model fine-tuned for real-world software engineering. The agent dynamically adjusts its thinking time, can run autonomously for over seven hours on massive refactors, and automatically reviews GitHub PRs for bugs. Why it matters: AI coding is shifting from line-by-line autocomplete into an independent teammate that tackles complex tasks end-to-end.
H Company just open-sourced Holo1.5, a family of 3B, 7B, and 72B models built specifically for Computer Use agents. The update boosts UI element localization accuracy by over 10% compared to its predecessor and handles complex software interfaces like Photoshop and VSCode. All weights are now accessible on Hugging Face.
Anthropic just rolled out native Claude integration for Apple's Xcode 26, bringing Claude Sonnet 4 directly into iOS and Mac development workflows. Developers can now debug, refactor, and generate SwiftUI previews using natural language inside the editor. The feature is available to Pro, Max, and Enterprise subscribers with Claude Code.
Apple just rolled out new Apple Intelligence capabilities across its hardware lineup, headlined by on-device Live Translation for phone calls, FaceTime, and Messages. The update also brings onscreen visual intelligence with ChatGPT integration, dynamic voice coaching in Workout Buddy, and direct developer access to its local foundation model.
Spectral Labs just launched SGS-1, a foundation model designed specifically for mechanical CAD. Instead of producing messy meshes, it takes an image, sketch, or STL file and outputs editable, watertight B-Rep parts in standard STEP format ready for assembly. It is not built for organic shapes yet, but it bridges the gap between generative AI and real-world manufacturing.
Meta just dropped MobileLLM-R1, a lightweight edge reasoning family spanning 140M to 950M parameters on Hugging Face. The 950M version delivers up to 5x higher accuracy on math benchmarks than larger open models while training on just 4.2T tokens. Why it matters: It proves sub-billion models can handle serious reasoning, though non-commercial licensing keeps it out of production for now.
OpenAI just published a massive study with Harvard economists analyzing 1.5 million conversations to reveal how people use ChatGPT. The biggest takeaway: 70% of consumer chats are for everyday personal life rather than work, with users treating the bot primarily as an advisor. Gender gaps have closed, and coding remains a niche use case.
Penske Media, the publisher behind Rolling Stone and Variety, just sued Google for antitrust abuse over AI Overviews. The lawsuit claims Google forces publishers to feed its AI summaries, directly cannibalizing referral traffic and ad revenue with no opt-out short of leaving search entirely. Google fired back, calling the claims meritless.
Alek Safar just introduced ClockBench to test multimodal models on analog clocks, and they struggled hard. While humans achieved 89.1% accuracy, the best model, Google's Gemini 2.5 Pro, only hit 13.3% with a median error of an hour. Interestingly, models solve time math easily once they get the time right, proving visual spatial reasoning remains the real bottleneck.
California lawmakers just passed SB 53, a major AI safety bill requiring large AI labs to disclose safety protocols and protect whistleblowers. It also sets up a public compute cloud called CalCompute, with stricter reporting rules for companies generating over $500 million in revenue. The bill now heads to Governor Gavin Newsom, who vetoed a similar bill last year.
Google Gemini just claimed the top free spot on the US Apple App Store, overtaking ChatGPT. The surge is largely driven by its viral Nano Banana image editor, known for strong character consistency, which drove 23 million new installs and over 500 million edits in two weeks. Free users get up to 100 edits daily.
OpenAI is quietly turning ChatGPT into a transactional utility. Latest app builds reveal a new "Orders" section in settings, pointing to native credit card checkout and tracking for plugins, services, and digital goods. The update also moves desktop Voice Mode into the main app window and lays groundwork for parental controls. Why it matters: ChatGPT is rapidly morphing from a conversational assistant into an all-in-one operating platform.
Kling AI just rolled out a new Avatar feature, integrating an upgraded version of its popular Lip Sync tool directly into the module. The update aims to make talking-head videos look much more natural. It is rolling out with limited access for now.
OpenAI is renegotiating its partnership terms with Microsoft, looking to trim the tech giant's revenue share from 20% down to around 10% by 2030. The move is expected to save OpenAI $50 billion over time as it transitions to a standard for-profit model, freeing up massive capital for compute and talent.
Tucker Carlson just confronted OpenAI CEO Sam Altman with a conspiracy theory claiming he ordered the murder of former researcher Suchir Balaji. Balaji had accused OpenAI of copyright violations before his death was ruled a suicide by police. Altman pushed back sharply, calling the accusation completely crazy.
Google and DeepMind just open-sourced VaultGemma, a 1B-parameter model trained from scratch with differential privacy. By mathematically injecting noise into training, it ensures the model never memorizes sensitive sequences. While privacy constraints pull its benchmark performance back to roughly GPT-2 levels, it provides the first empirical scaling laws for provably private AI.
Apple's top AI and search executive Robby Walker is planning to leave the company next month, according to Bloomberg. Walker, who joined in 2013 and previously led Siri, follows several other high-profile departures to Meta as Apple struggles to catch up in generative AI and delays major Siri upgrades.
California lawmakers just passed SB 243, sending a bill to Governor Gavin Newsom that would hold AI companion bots like Character.AI and ChatGPT legally accountable for user safety. Platforms would have to block self-harm and explicit content, and ping minors every three hours reminding them their digital pal isn't human. Why it matters: If signed, California becomes the first state to regulate emotional AI, allowing users to sue platforms for up to $1,000 per violation.
Microsoft and OpenAI just signed a non-binding memorandum of understanding to pave the way for the next phase of their partnership. The two companies are finalizing terms for a definitive agreement while reaffirming their shared commitment to safety.
Udio just rolled out Voices, a new tool giving creators tighter control over song vocals. You can now pick from a curated library of pre-loaded vocal styles or extract and save the voice from a generated track to reuse across future songs. It is free for everyone for two weeks before moving behind the paywall.
Alibaba just released Qwen3-Next-80B-A3B, an ultra-sparse MoE model that packs 80B total parameters while only activating 3B per token. It delivers 10x faster inference and 10x cheaper training than Qwen3-32B, with its Thinking variant outperforming Gemini-2.5-Flash-Thinking. Basically, flagship-grade reasoning is now running at lightweight speed.
Levita Magnetics just performed the world's first gallbladder surgery using an AI-guided autonomous camera on its MARS robotics platform in Chile. The system automatically tracks surgical instruments to keep the field of view stable without a camera assistant. It is not yet cleared by the FDA or available for sale.
OpenAI just clarified its corporate restructuring, revealing that its original nonprofit will retain governance control over a new Public Benefit Corporation while holding an equity stake exceeding $100 billion. The move creates one of the best-funded philanthropies in history to satisfy regulators, while simultaneously unlocking OpenAI's ability to raise massive new capital.
Albania just appointed Diella, an AI-powered virtual avatar, as the world’s first non-human cabinet minister to take over all public procurement. Prime Minister Edi Rama handed government tender decisions over to the algorithm to eliminate chronic corruption and make spending transparent. Why it matters: It is the boldest experiment yet in putting high-stakes state governance entirely in the hands of an AI agent.
Anthropic just gave Claude long-term memory, letting the assistant retain project context, workflows, and preferences so you don't have to keep re-explaining things. Each project keeps its own isolated memory bank, users can edit what Claude remembers, and an incognito mode lets you chat off the record.
