Nvidia just rolled out G-Assist, an experimental AI chatbot that runs locally on your GPU inside the Nvidia desktop app. It lets you monitor PC performance, tweak game settings, and even auto-overclock your graphics card through a floating overlay. Game integration is still sparse, but it's a neat preview of local AI hardware assistants.
Earth AI just revealed it uncovered promising deposits of copper, cobalt, and gold across Australian sites that miners overlooked for decades. The startup trained algorithms on national geological archives dating back to the 1970s and built custom drill rigs to prove the predictions. It proves AI can compress decades-long prospecting cycles into fast, data-driven discoveries.
Google just released Gemini 2.5 Pro Experimental, its newest flagship model that reasons through thoughts before answering. The model immediately took the #1 spot on LMArena, boasts a 1-million-token context window, and hits new highs on math and coding benchmarks like SWE-bench. It is available right now in Google AI Studio and for Gemini Advanced subscribers.
Perplexity just launched dedicated answer tabs, letting users seamlessly filter searches across images, video, travel, and shopping in one place. It basically turns the AI search engine into a full-fledged vertical search hub. The feature is available now on the web, with mobile coming soon.
Luma AI just launched Ray2 image-to-video inside Dream Machine, letting users turn quick hand-drawn doodles and sketches into animated videos. Simply drop in scribbled characters or scenes, and the model animates them into dynamic clips. It is a neat, low-friction shortcut for creators looking to prototype animations straight from raw sketches.
Bambu Lab just unveiled the H2D, a massive modular 3D printer starting at $1,899 that doubles as a laser cutter, vinyl plotter, and drawing machine. It features dual extruders, heated filament drying, and a quick-release mount for swappable tool heads. Why it matters: It tries to conquer the entire maker desktop with a single machine, though early testing hints that multi-tool complexity brings fresh reliability hurdles.
Otter just launched three new AI agents, led by a voice-activated Meeting Agent that can verbally answer questions during calls. It taps into your company's historical meeting archive to retrieve answers, schedule follow-ups, and draft emails. The feature is currently rolling out on Zoom, with Teams and Google Meet support coming soon.
OpenAI just rolled out native image generation inside GPT-4o, making it the new default engine in ChatGPT. It nails in-image text rendering, handles up to 20 distinct objects at once, and maintains character consistency across ongoing chats. Renders can take up to a minute, but the jump in control is huge.
Waymo just announced plans to launch its Waymo One driverless ridehailing service in Washington, DC by 2026. The major roadblock is local law, which currently requires a human behind the wheel for testing and bans driverless commercial operations entirely. If regulators clear the way, the fleet—which already logs 200,000 paid rides weekly—will roll directly into the nation's capital.
Reve AI just launched Reve Image 1.0, a new text-to-image generator that unexpectedly grabbed the #1 spot on Artificial Analysis's Image Arena over Midjourney and FLUX. The model excels at rendering readable text inside images and editing existing visuals with simple language prompts. It is currently free to try on the web with no API yet.
Synthesia just launched a $1M equity fund to grant stock options to the human actors behind its most popular stock avatars, vesting monthly over up to four years. Previously, actors only received flat licensing fees for three-year stints. Why it matters: it’s an unusual step toward revenue sharing and mutual buy-in between synthetic media platforms and real human talent.
Apple just announced that its annual Worldwide Developers Conference (WWDC25) will run online from June 9 to 13. The conference will be completely free for all developers and spotlight the next generation of Apple software, with a limited in-person event kicking off at Apple Park on opening day.
Google just released TxGemma, an open family of models fine-tuned from Gemma 2 to predict molecular properties and analyze therapeutics data. Available in 2B, 9B, and 27B sizes, the models include specialized predict and chat variants that explain reasoning. The 27B predict model matches or beats previous SOTA on 64 out of 66 clinical and chemical tasks.
DeepSeek just quietly updated its flagship DeepSeek-V3 model, switching to a permissive MIT license that lets developers freely build commercial projects. Community benchmarks show a small boost in coding performance, and an Apple researcher managed to run a quantized version on a Mac Studio at 20 tokens per second. It's a huge win for open-source AI commercialization.
ARC Prize just unveiled ARC-AGI-2 and launched its 2025 competition to stress-test general fluid intelligence. Every test task was solved by multiple humans in two tries or less, yet pure LLMs score a clean 0% and public reasoning systems barely reach single digits. Why it matters: simply scaling up parameters won't close this reasoning gap; AGI needs fundamentally new architectures.
Alibaba just released Qwen2.5-VL-32B-Instruct, a visual language model fine-tuned with reinforcement learning. It delivers major boosts in math reasoning and human preference alignment, designed as the sweet spot for users who find 72B too heavy and 7B not capable enough for local rigs.
Apple just rolled out the iOS 18.4 Release Candidate to developers ahead of an early April public release. The update adds Priority Notifications to highlight key alerts on your lock screen, brings a new Sketch style to Image Playground, and expands Apple Intelligence to simplified Chinese, French, Japanese, and more. Barring critical bugs, this is what hits your iPhone next month.
Boston Dynamics just expanded its partnership with NVIDIA to bring the Isaac GR00T framework and Jetson Thor compute to its electric Atlas humanoid. The setup lets Atlas run heavy multimodal models alongside its native whole-body controllers, while training complex locomotion in Isaac Lab. Why it matters: The world's most agile robot just picked up the most popular AI compute stack.
OpenAI just expanded roles for three key executives to better bridge frontier research and mass-market products. Longtime researcher Mark Chen stepped in as Chief Research Officer to tie lab breakthroughs directly to product development, while COO Brad Lightcap broadened his operational oversight and Julia Villagra took on Chief People Officer.
Midjourney just teamed up with NYU researchers to tackle AI's biggest creative flaw: bland, repetitive writing. By introducing two new training methods called DDPO and DORPO, they reward rare and unique high-quality responses during fine-tuning on models like Llama 3.1. It turns out the image generator has serious ambitions for text, too.
Apple is reportedly working on adding cameras to the Apple Watch to power features like Visual Intelligence within the next two years, according to Bloomberg's Mark Gurman. Standard models could tuck lenses under the display, while the Ultra might mount one beside the digital crown. Apple basically wants wearables to see the physical world, aiming to drive the system with in-house models by 2027.
Alibaba researchers just introduced LHM, a reconstruction model that creates high-fidelity, animatable 3D humans from a single 2D image. By relying on 3D Gaussian splatting and a multimodal transformer, it generates fully articulated avatars in seconds without requiring hand or face post-processing.
Google just started rolling out Project Astra-powered camera and screen sharing to Gemini Live on Android. Users can stream their live camera feed or share their full phone screen for real-time visual assistance. It is currently rolling out gradually to Gemini Advanced subscribers.
Lyft just announced that riders in Atlanta will be able to hail robotaxis starting this summer, using a fleet of Toyota Sienna minivans powered by May Mobility. The ride-hailing company also plans to expand to Dallas in 2026 with Mobileye-equipped vehicles, gearing up to scale thousands of autonomous cars across multiple cities.
Boston Dynamics just dropped a wild video showing its electric humanoid robot Atlas pulling off cartwheels, forward rolls, and slick breakdancing coffee-grinder spins. Teaming up with the Robotics and AI Institute, the team trained the bot via motion-capture suits and reinforcement learning. Why it matters: As humanoid competition heats up globally, Boston Dynamics is pushing extreme agility, even while under heavy pressure to turn parlor tricks into real commercial revenue.
Perplexity just proposed an ambitious bid to acquire TikTok, promising to rebuild its recommendation system from scratch in US data centers and make the code open-source. The plan also includes adding AI search and video citations. Why it matters: With TikTok valued at up to $50B, the pitch faces massive financial hurdles and may simply be another wild PR stunt.
Meta is testing a "Write with Meta AI" button on Instagram that scans photos to draft three ready-made comment suggestions. Users can tap a pencil icon in the comment bar to pick or refresh replies. While Meta wants to boost interactions, users already worry the tool will replace authentic friend check-ins with low-effort AI slop.
OpenAI is facing a fresh GDPR privacy complaint after ChatGPT told a Norwegian man that he was a convicted murderer serving 21 years for killing his own children. Privacy group Noyb filed the complaint with Norway's data watchdog, arguing that hallucinating defamatory claims alongside real personal data violates European privacy law, and a tiny disclaimer is not enough.
Alie Jules just launched VrefDeck, a dedicated visual organization app built for AI creators, designers, and photographers. The tool lets users drag, sort, and tag reference images with custom cards for parameters and prompts, running 100% locally to protect private workflows. It's available on Mac and Windows for a one-time purchase of $29.
Google just rolled out an AI-powered search update for Gmail that surfaces emails based on relevance instead of plain chronological order. It factors in who you email most and what you actually click, though a quick toggle lets you switch back to the classic "Most recent" view. It is rolling out globally to personal accounts on web, iOS, and Android.
Apple is replacing its AI and Siri lead John Giannandrea with Vision Pro head Mike Rockwell. The shakeup comes after Apple delayed generative AI Siri upgrades to next year, with CEO Tim Cook reportedly losing confidence in the AI team's execution. Hardware lead Paul Meade will take over the Vision Pro group.
Pika just teased a new video manipulation tool that lets users modify specific characters or objects while leaving the rest of the frame completely intact. It solves one of AI video's biggest headaches—unwanted background warping—though access is currently limited to selected Creative Partners.
Bigscreen just revealed the Beyond 2, shrinking its ultra-compact PC VR headset down to just 107 grams while adding built-in eye tracking. The new model features a wider 116-degree field of view, 2560x2560 micro-OLED panels, and mechanical IPD adjustment, with the first batch shipping in April 2025.
Topaz Labs just rolled out Gigapixel 8.3, introducing a brand-new Recover v2 model that restores high-resolution image details over 5x faster with significantly fewer artifacts. It runs locally with 8GB VRAM or in the cloud, alongside a refreshed interface and Gigaprint cropping tools.
Krea just rolled out custom video training powered by Wan 2.1. Creators can now feed their own footage into the model to teach it specific visual styles, motions, or custom subjects for tighter generation control.
Anthropic just gave Claude real-time web search across all plans globally. The assistant now pulls live data from the internet and embeds direct citations in its conversational responses. It basically saves you from juggling multiple browser tabs when analyzing market trends or fact-checking sources.
Anthropic just detailed the "think" tool, a technique giving Claude dedicated space to reason mid-execution during complex agentic workflows. Unlike extended thinking done upfront, it allows Claude to inspect tool outputs and navigate strict policies, driving a 54% relative improvement on τ-Bench airline tasks.
Meta just announced the rollout of Meta AI across Europe, bringing conversational AI in six European languages to WhatsApp, Instagram, Facebook, and Messenger for free. Users can summon the assistant in group chats using @MetaAI or simply tap the new blue circle icon to search and brainstorm.
OpenAI just rolled out its compute-heavy o1-pro model to the API for select Tier 1–5 developers. It burns more compute to deliver sharper reasoning responses, while adding vision, function calling, and structured outputs. The catch is the eye-watering price: $150 per million input tokens and $600 per million output tokens.
xAI just rolled out image generation capabilities in its developer API, starting with a single model called grok-2-image-1212. It outputs up to 10 JPGs per request at $0.07 per image, sitting right between FLUX and Ideogram pricing. The catch: you can't customize size or style yet, and prompts are automatically revised by a chat model.
Razer just unveiled AI QA Copilot, a tool designed to take the friction out of video game testing. As QA testers play through a build, the AI flags crashes and glitches in real time, automatically compiling reproduction steps, logs, and video clips into ready-to-file tickets. Human oversight is still required, but cutting down manual paperwork could noticeably speed up release cycles.
Nvidia just teamed up with Google DeepMind and Disney Research to build Newton, an open-source physics engine designed to help robots learn complex physical tasks with higher precision. Disney plans to plug Newton into its next-generation entertainment animatronics, starting with Star Wars-inspired BDX droids.
Cloudflare just rolled out AI Labyrinth, a clever defensive system that turns generative AI against rogue web crawlers. Instead of blocking scrapers outright, it routes them into a massive maze of AI-generated decoy pages that burn the bot's computing power and fingerprint bad actors. It's available as a free one-click toggle for all users.
Tesla just secured a commercial ride-hailing permit in California, clearing the path to compete head-on with Uber and Lyft. The catch is that human drivers are still mandatory—the permit doesn't cover autonomous vehicles. Tesla plans to have employees drive other staff first before rolling the app-based service out to the public.
Perplexity just rolled out major upgrades to its Sonar models, adding High, Medium, and Low search modes so developers can balance search depth with compute costs. The company is also waiving fees for citation tokens across responses. Internal benchmarks show Sonar outperforming search-enabled GPT-4o at a fraction of the price.
Nvidia and Elon Musk's xAI just officially joined the AI Infrastructure Partnership (AIP), a coalition led by Microsoft, BlackRock, GIP, and Abu Dhabi's MGX. The alliance aims to mobilize $30 billion in private capital to build massive data centers and energy infrastructure, primarily in the US. Why it matters: Surging compute and power demands are pushing fierce tech rivals into unprecedented infrastructure alliances.
Robert Legato, the VFX innovator behind Titanic and Avatar, has joined Stability AI as chief pipeline architect. The move reunites him with board member James Cameron to bridge generative AI tools with visual effects pipelines. Legato says he wants to champion an artist-first approach that empowers filmmakers rather than replaces them.
Google's NotebookLM team just started rolling out interactive mindmaps, designed to turn notes into interactive toys that help users learn through play. Instead of just returning text or static images, the notebook can now generate tailored visual tools for deeper exploration. Why it matters: It reflects a push toward personalized software tailored to how each person learns.
OpenAI just rolled out direct PDF inputs for its Responses and Chat Completions APIs. Developers can now pass raw PDF documents straight to models like o1, gpt-4o, and gpt-4o-mini without third-party parsing tools. Why it matters: document workflows just got a lot simpler.
Open-source tool Browser Use just blew up after users realized it powers the viral AI agent Manus. Co-creator Gregor Zunic says daily downloads jumped from around 5,000 on March 3 to 28,000 on March 10. By letting models click buttons and fill out forms, it is rapidly becoming the shovel of choice in the web agent gold rush.
Authors just moved for summary judgment against Meta in a California federal court, alleging the company torrented terabytes of pirated books to train its AI models and uploaded data back to the P2P network while leeching. The plaintiffs mocked Meta's justification as a "Bob Dylan defense"—stealing a little lands you in jail, while stealing a lot makes you king.
Captions just launched Mirage, an AI model designed specifically for generating UGC-style video ads with virtual spokespersons. Users can supply a script or audio to produce realistic avatars featuring nuanced body language and micro-expressions, with full control over outfits, backdrops, and emotions. It dramatically lowers the bar for scaling testimonial-style ad creative.
Google just opened native image generation in Gemini 2.0 Flash to all developers via Google AI Studio and its API. Built from the ground up to output images natively, the model handles multi-turn conversational editing, preserves character consistency across prompts, and renders in-image text far better than typical diffusion models.
Snap just launched its first AI Video Lenses powered by an in-house generative video model. Available to Snapchat Platinum subscribers at $15.99 a month, the initial trio can generate cuddling animals or dramatic zoom-outs directly into saved Snaps. Snap plans to drop more weekly as it builds out proprietary tech to take on TikTok and Instagram.
Google just released Gemma 3, its newest lineup of open-weight models ranging from 1B to 27B parameters. The 4B, 12B, and 27B versions handle both images and text with an expanded 128k token context window, supporting over 140 languages. Why it matters: The 27B model beats Gemini 1.5 Pro across key benchmarks, offering top-tier multimodal power you can run openly.
Rivian just rolled out an update for its second-generation vehicles, bringing hands-free highway driving to 135,000 miles of roads across the US and Canada. The system monitors driver gaze via an infrared rearview mirror camera, with plans to expand road coverage and add eyes-off capabilities in 2026.
Freepik just added Google's flagship video model Veo 2 to its AI Suite, giving users the ability to convert static images directly into high-fidelity video clips. The integration connects image generation, retouching, and motion inside a single workflow. The catch: it is reserved exclusively for Freepik Premium subscribers.
Google DeepMind just launched Gemini Robotics and Gemini Robotics-ER, two models built on Gemini 2.0 designed to bring AI reasoning into physical actions. The models can handle delicate chores like origami and zip-bag packing, and DeepMind is already teaming up with Apptronik to power Apollo humanoid robots.
Nous Research just launched an inference API, giving developers cloud access to Hermes 3 70B and the DeepHermes-3 reasoning model without hosting weights locally. It matches OpenAI's format for drop-in replacement, offers a $5 starter credit, and is rolling out via a first-come waitlist.
Niantic just announced a massive restructuring, selling its gaming division—including Pokémon GO and Pikmin Bloom—to mobile giant Scopely in a deal valued at around $3.85 billion. The company is parting ways with game operations to focus entirely on spatial computing and AI through Niantic Spatial.
Sakana AI just announced that a scientific paper fully generated by its upgraded The AI Scientist-v2 passed peer review at an ICLR workshop. Conducted with UBC and Oxford, it marks the first time a completely AI-authored paper cleared the exact same review process human researchers face. Full details of the v2 system will drop soon.
Microsoft just rolled out the Responses API and Computer-Using Agent (CUA) inside Azure AI Foundry. CUA lets AI autonomously navigate software interfaces, click buttons, and fill forms from natural language prompts, while the new API bundles tool calls into a single request. It turns AI from an advisor into an operator, though human oversight is still urged for sensitive workflows.
Sam Altman just revealed that OpenAI trained a new model specifically excelling at creative writing, sharing a metafictional short story about AI and grief. Altman noted it was the first time AI writing had genuinely moved him. Why it matters: LLMs often produce bland prose, but a model tuned for literary nuance could change creative workflows. No release timeline has been announced yet.
Harvey just launched agentic workflows in Assistant, designed to collaborate with legal professionals across transactional, litigation, and financial tasks. Instead of requiring detailed prompts, the agents proactively guide users and surface their thinking steps. In blind evaluations, Harvey matched lawyers in contract drafting and beat them in analyzing court transcripts. The feature rolls out broadly at the end of March.
Nvidia's biggest event of the year is underway, with CEO Jensen Huang scheduled to deliver his keynote on Tuesday at 10 a.m. PT. Viewers can stream the presentation for free on Nvidia's website or YouTube without registering, with updates expected on Blackwell Ultra and the next-gen Rubin architecture.
Perplexity just launched an official desktop app for PC, adding voice dictation and keyboard shortcuts to access its latest models directly from the desktop. It basically turns the AI search engine into a native desktop tool.
Reka just open-sourced Flash 3, a 21B parameter reasoning model built completely from scratch. Despite its lightweight footprint, Reka claims competitive performance and is already using it as the core engine behind its new enterprise intelligence platform, Nexus.
Cartesia just raised a $64M Series A led by Kleiner Perkins to build out its real-time voice AI platform. Alongside the funding, the startup officially launched Sonic 2.0, its next-gen voice model aimed at cutting latency and powering the next wave of interactive voice apps.
Luma just introduced Inductive Moment Matching (IMM), an open-source pre-training technique designed to replace conventional diffusion models. By taking an inference-first approach and incorporating target timesteps with moment matching, IMM delivers better image quality than diffusion or flow matching while cutting sampling steps by up to 30x.
Manus AI just partnered with Alibaba's Qwen team to integrate its viral general AI agent with Qwen's open-source models and domestic computing infrastructure. After facing massive hype and severe site crashes following its launch last week, Manus is leaning on Alibaba's resources to manage traffic surges and scale beyond its current invite-only rollout.
Adobe just added a Firefly-powered feature called Customize to Adobe Stock, letting users edit assets directly in the browser without jumping into Photoshop. You can expand images, apply styles, and harmonize looks across assets with full version history. It is available now for paid subscribers and runs on generative credits.
Anthropic CEO Dario Amodei predicted that AI will write 90% of software code within three to six months and practically all of it in a year. The bold forecast suggests software engineering will pivot from writing code to reviewing and orchestrating AI-generated logic.
OpenAI just launched a dedicated suite of tools to help developers build AI agents, anchored by the brand-new Responses API alongside built-in web search, file search, and computer use. It also released an Agents SDK and tracing tools to orchestrate complex multi-step workflows. Why it matters: OpenAI is streamlining agentic apps into a single endpoint and plans to sunset the Assistants API by mid-2026.
Apple is reportedly developing next-gen AirPods equipped with tiny external cameras and AI to see the world around you. According to Bloomberg’s Mark Gurman, the earbuds would essentially act like screenless smart glasses, pairing with Siri to offer real-time contextual awareness—likely targeting a 2027 debut on AirPods Pro 4.
A federal judge just allowed an AI copyright lawsuit against Meta to move forward, ruling that authors including Sarah Silverman adequately alleged Meta intentionally stripped copyright management information to hide that Llama was trained on their books. While the judge dismissed a computer fraud claim, Meta still faces key infringement claims.
Meta just kicked off small-scale testing of its first custom AI training silicon, aiming to rein in its gargantuan infrastructure spending and ease its reliance on Nvidia. The MTIA chip is designed to be more power-efficient than general-purpose GPUs, and Meta hopes to deploy it for production training by 2026.
OpenAI just inked a massive five-year deal worth $11.9 billion with CoreWeave for AI data centers and infrastructure services. As part of the agreement, OpenAI will also receive around $350 million in equity tied to CoreWeave's upcoming IPO. The commitment secures vast GPU capacity for OpenAI while supercharging its partner's public debut.
ElevenLabs just cut speech-to-text pricing by 45% at scale and made its Scribe model completely free via the UI until April 9. The model boasts high accuracy across 99 languages. It is an aggressive move to undercut rivals and attract developers needing high-volume transcription.
Google just added a Gemini-powered "Add to calendar" button to Gmail on the web. The AI automatically detects scheduling details in your emails, letting you pop an event onto your calendar via the side panel with one click. Why it matters: It eliminates the copy-paste routine for meetings, though it's currently limited to English and paid Workspace or Google One AI Premium tiers.
Tencent Hunyuan just unveiled Hunyuan-TurboS, an ultra-large MoE model blending Transformer and Mamba architectures to ease long-context training and KV-cache overhead. The team claims it tops GPT-4o and DeepSeek-V3 in math and reasoning, all while cutting inference costs to one-seventh of its predecessor.
Researchers just introduced Generalized Interpolating Discrete Diffusion (GIDD) to overcome the rigid limitations of next-token prediction. By blending masking with uniform noise, it achieves top-tier generation performance while finally allowing models to revise and self-correct their own mistakes.
Apple is reportedly planning a dramatic interface overhaul across iOS 19, iPadOS 19, and macOS 16 ahead of WWDC in June. The new look borrows heavily from visionOS with translucent panels and 3D depth to unify icons, menus, and navigation. Why it matters: It marks Apple's most sweeping visual revamp since iOS 7, arriving as the company navigates sluggish device upgrades and recent Siri AI delays.
A federal judge just ruled that an AI copyright lawsuit against Meta can proceed, rejecting the company's bid to dismiss the case. Authors allege Meta trained its Llama models on their books and removed copyright information to hide the infringement. The judge tossed a computer fraud claim, but Meta must now defend its fair-use arguments in court.
US lawmakers have already introduced 781 AI-related bills just over two months into 2025, according to tracking data from MultiState. That figure officially exceeds the 743 total bills proposed across all of 2024. With Congress still stalled on a federal AI framework, individual states are rushing to fill the regulatory void.
Google just brought its Gemini side panel to Google Calendar for Workspace Labs testers, turning schedule management into a simple conversation. You can ask when your next meeting is or set recurring events in plain English without clicking through menus. It is currently limited to early testers.
Highlight just launched a universal interface for MCP servers, aiming to bring Cursor-like interactions across every app on your computer. It features screen grounding that lets you mention services in-context and write with Tab, pitching itself as what Apple Intelligence and Copilot should have been.
xAI just confirmed that users can now ask Grok anything simply by replying to an X post with "grok". Powered by the rollout of Grok-2 and Grok-2 mini, the update turns any reply thread into a live AI chat without leaving the timeline.
Meta is reportedly gearing up to launch Llama 4 in just a few weeks, putting a major focus on native voice capabilities. Meta CPO Chris Cox confirmed it will be an "omni" model handling speech and text natively, allowing users to interrupt mid-sentence like ChatGPT's Voice Mode. The rush comes as DeepSeek's open-source gains pushed Meta into overdrive.
Microsoft AI head Mustafa Suleyman is pushing to reduce the company's heavy dependence on OpenAI, actively testing in-house models like Phi-4 inside Copilot. Microsoft wants to trim sky-high Azure compute bills and reclaim its strategic roadmap, but untangling a deeply integrated $13 billion partnership is proving far easier said than done.
ElevenLabs just launched its voice AI models on Google Cloud Marketplace. Enterprises can now easily pair ElevenLabs' lifelike speech generation with Google's fast Gemini 2.0 Flash model to build real-time conversational agents, localized dubbing, and marketing voiceovers within their existing cloud setup.
xAI just purchased a 1-million-square-foot property in Southwest Memphis to expand its massive Colossus supercomputer cluster. The facility complements an existing site that features an $80 million wastewater recycling plant and the world's largest Tesla Megapack deployment to reduce strain on local grids. Why it matters: Elon Musk's race for raw computing scale isn't slowing down, and xAI is locking down the physical footprint needed to power next-gen frontier models.
Google just launched an experimental text embedding model, gemini-embedding-exp-03-07, through the Gemini API. Trained directly on Gemini, it took the top spot on the MTEB Multilingual leaderboard with a score of 68.32, beating the runner-up by 5.81 points. Developers can try it out immediately using the existing embed_content endpoint.
Apple just confirmed that its most anticipated Siri upgrades—the ones with personal context and cross-app actions—are delayed until "the coming year." Getting an AI assistant to dig through private emails and reliably execute real-world tasks proved harder than planned. Basically, that truly smart Siri likely won't arrive until iOS 19.
U.S. Air Force and Space Force officials just outlined how they want to integrate AI into America's nuclear command, control, and communications architecture. The plan is to use AI to speed up data processing and pick the most secure communication routes, while strictly keeping humans—specifically the president—in full control of any strike decisions.
