X just rolled out Stories, an AI feature powered by Grok that serves up short summaries of trending topics in the Explore tab. Instead of reading the original news articles, Grok builds its overviews entirely from the posts and conversations happening across the platform. The feature is currently exclusive to paying X Premium subscribers.
Apple is reportedly preparing a suite of AI features for iOS 18 powered by its Ajax LLM, including text summarization in Messages and Safari. Simple tasks will run on-device with strict privacy warnings, while complex queries could tap server-side power from Google or OpenAI. Details are expected at WWDC on June 10th.
OpenAI just sparked speculation about a direct challenge to Google after registering the search.chatgpt.com domain. Reports indicate an official announcement could arrive as early as May 9. Rather than relying on clunky Bing add-ons, the new tool aims to deeply integrate real-time web retrieval directly into ChatGPT to rethink the traditional search experience.
Microsoft just updated Azure OpenAI Service terms to explicitly bar U.S. police departments from using its AI models for facial recognition. The new rules also ban law enforcement globally from running real-time facial recognition on body cameras and dashcams, drawing a firm line on biometric surveillance.
OpenAI just landed its first major music video project with artist Washed Out's new single "The Hardest Part." Director Paul Trillo spent six weeks cherry-picking 55 clips from roughly 700 Sora-generated prompts to pull off sweeping visuals on a budget, though audience reactions have been sharply mixed.
Microsoft initially poured $1 billion into OpenAI in 2019 largely because it was worried about falling behind Google's machine learning scale, according to unsealed internal emails. CTO Kevin Scott warned Satya Nadella and Bill Gates that Microsoft took six months just to replicate Google's BERT model due to lagging infrastructure, prompting Nadella to push for the deal.
Udio just bumped its track limit to 15 minutes and stretched extension context from 30 seconds to 2 minutes. That extra memory means the model can remember parts from earlier in your song, finally making consistent choruses possible in long mixes.
Ideogram just rolled out a new Pro subscription tier for power creators at $48 a month billed annually. Subscribers get 3,000 fast prompts to churn out up to 12,000 images monthly in under 15 seconds, plus an upcoming CSV bulk upload feature. It supports commercial use, though enterprise copyright indemnification is still missing.
Anthropic just launched a new Claude Team plan alongside its first official iOS app. The team tier costs $30 per seat per month (with a 5-seat minimum), unlocking higher usage limits, full Claude 3 model access, and a 200K context window. The free mobile app brings seamless chat sync and direct photo analysis to your phone.
Meshy just launched Meshy-3, rolling out high-poly sculpture generation and complete PBR texture maps straight from text prompts. The release also overhauls the Image-to-3D pipeline and adds a smart healing tool to clean up messy AI artifacts, making generated assets far more practical for game engines.
Google is reportedly shelling out $5 million to $6 million a year to News Corp to develop AI-related content and products. While tech giants scramble to secure media partnerships to feed their models, News Corp quickly pushed back, stating it has absolutely no AI licensing deal in place with Google.
OpenAI just rolled out its "Memory" feature to all ChatGPT Plus subscribers outside of Europe and Korea. The chatbot can now remember key details, preferences, and context across different conversations. You'll get notified when a memory updates, and you can easily view, delete, or turn off memories in settings whenever you want.
GitHub just launched the technical preview of Copilot Workspace, a Copilot-native developer environment designed to build software from scratch in natural language. Starting from an issue or repo, agents brainstorm, plan, and test implementation while leaving developers in full control of every step. Why it matters: Coding is shifting rapidly from managing syntax to expressing intent.
OpenAI just struck a licensing partnership with the Financial Times to bring FT's journalism into ChatGPT. Users will see attributed summaries, quotes, and links to FT reporting in relevant answers. The FT has also deployed ChatGPT Enterprise to all staff, signaling that major publishers are opting to partner rather than fight.
Sam Altman’s OpenAI is reportedly in talks to partner with Worldcoin, his other high-profile venture focused on crypto and iris-scanning digital identity. Under the potential deal, OpenAI would provide AI solutions and services to Worldcoin, though the close ties between Altman’s two ventures are almost certain to draw fresh regulatory scrutiny.
Meta advertisers are watching budgets rapidly evaporate as the company's automated platform, Advantage Plus, misfires with skyrocketing costs and dismal returns. Multiple marketers reported CPMs spiking up to 10 times higher than normal while generating virtually zero revenue. Why it matters: The promise of 'set it and forget it' automation is colliding with costly bugs, driving frustrated businesses back to manual ad setups.
Beijing Humanoid Robot Innovation Center just unveiled Tiangong, a full-sized electric humanoid robot that runs steadily at 6 km/h with human-like motion. Weighing 43 kg, it navigates slopes blindly and will serve as an open-source base platform for developers.
Shengshu Technology and Tsinghua University just unveiled Vidu, a text-to-video AI capable of generating 16-second clips at 1080p resolution in one click. Built on their U-ViT architecture, it supports multi-camera transitions and simulates real-world lighting. While it doesn't quite match Sora's photorealism yet, its temporal consistency marks a serious step forward for Chinese video models.
Apple has reopened talks with OpenAI to power new generative AI features on iPhones later this year. The discussions focus on integrating a ChatGPT-like bot into iOS 18 alongside Apple's own models. Apple is still in active talks with Google over Gemini licensing as well, with no final decision made.
The U.S. Department of Homeland Security just formed a 22-member AI Safety and Security Board, tapping tech leaders like Sam Altman, Satya Nadella, and Sundar Pichai. The panel is tasked with shielding critical infrastructure like power grids and airports from AI-driven threats. Why it matters: As risks range from automated cyberattacks to bioweapons, Washington is asking the builders of AI to help safeguard the nation against it.
Drake just took down his Kendrick Lamar diss track "Taylor Made" from X after Tupac Shakur’s estate gave him a 24-hour ultimatum to pull it or get sued. The track used AI-generated vocals mimicking the late rap legend. Why it matters: It’s one of the most high-profile clashes yet over unauthorized AI voice clones in mainstream music.
Google just announced a $3 billion investment to build and expand data centers across Virginia and Indiana to power its AI workload. Alongside the infrastructure boost, Google.org is launching a $75 million fund and the Google AI Essentials course to train one million Americans in basic AI skills for free.
A Maryland high school athletic director was just arrested after allegedly using AI to fabricate a racist audio recording of his principal. The deepfaked clip went viral and got the principal suspended, before forensic analysts traced the audio back to AI tools used on the school network. Why it matters: It’s a chilling look at how accessible voice-cloning tech is already being weaponized for personal vendettas.
Profluent just released OpenCRISPR-1, the world’s first precision gene editor designed from scratch using protein language models. The AI-generated editor performs as well as or better than natural Cas9 in human cells despite carrying over 400 mutations. It signals a shift from hunting for enzymes in nature to designing custom biological tools on demand.
Sanctuary AI just unveiled the seventh-generation Phoenix humanoid, promising lower manufacturing costs and task learning times under 24 hours. The update focuses on upper-body dexterity, a lighter frame, and longer uptime as the company prepares to deploy systems into Magna auto facilities.
Google just released a new research report with fellow Andrew McAfee exploring how generative AI will shape the economy. The research argues that fears of mass technological unemployment are likely overblown, predicting AI will diffuse much faster than electricity or the steam engine and help boost entry-level workers' performance.
LA Metro is installing AI-powered cameras on city buses to automatically catch drivers illegally parked in dedicated bus lanes and stops. Violations won't be purely automated—footage gets sent to the transportation department for human review before tickets are issued. Testing kicks off this summer.
Synthesia just rolled out Expressive Avatars, an upgrade that brings real emotions and tighter lip-sync to its corporate video presenters. Instead of stitching together pre-recorded video clips, the avatars are generated entirely on the fly using multimodal deep learning. Why it matters: It takes the robotic edge off training and marketing videos, though getting hands right remains a work in progress.
Researchers at EPFL and Yale School of Medicine just released Llama-3[8B]-MeditronV1.0, fine-tuned on Meta's Llama 3 in 24 hours to help clinicians in low-resource and humanitarian settings. The open-source model beats all open models in its size on MedQA and MedMCQA, with weights and data openly available to ensure equitable access.
Tupac Shakur’s estate just threatened legal action against Drake over his Kendrick Lamar diss track "Taylor Made," which used unauthorized AI clones of the late rapper and Snoop Dogg. The estate gave Drake 24 hours to pull the track, calling it a blatant abuse of Tupac's legacy.
AI coding startup Augment just emerged from stealth with a massive $252M funding round at a near-unicorn $977M valuation. Backed by former Google CEO Eric Schmidt and top venture firms, the company is founded by ex-Microsoft and Google veterans aiming to fix code quality and developer burnout. Why it matters: Coding assistants are a brutal, cash-burning market, but Augment now has the runway to take a real shot at Microsoft's GitHub Copilot.
Rabbit just started handing out its $199 R1 AI gadgets to early buyers in New York, and initial hands-on tests reveal a charming yet clearly unfinished device. While basic queries and a rotating camera work fine, core promised features like email and spreadsheets are still missing.
Snowflake just launched Arctic LLM, an enterprise-focused open model licensed under Apache 2.0 for research and commercial use. Using a 480B-parameter MoE architecture that only activates 17B parameters at a time, it cost just $2 million to train and targets SQL generation and corporate coding workloads. Snowflake claims it outperforms rivals like Databricks’ DBRX on database tasks.
Moderna just rolled out ChatGPT Enterprise across its thousands of employees, with staff building 750 custom GPTs in just two months. Use cases range from Dose ID—which analyzes clinical trial data to optimize vaccine dosages—to legal review bots. Why it matters: Moderna is betting that generative AI can allow a few thousand people to operate like a 100,000-person pharma giant, aiming to launch 15 new treatments in five years.
SenseTime just launched SenseNova 5.0, a major model trained on over 10TB of tokens with a 200,000 context window. Alongside the cloud model, it rolled out an edge-side matrix that hits 78.3 words per second on flagship chips and cuts enterprise inference costs by 80%. Why it matters: SenseTime is pushing frontier multimodal AI directly into phones, cars, and edge devices.
Apple has cut its 2024 Vision Pro shipment forecast to 400,000–450,000 units, well below market expectations of 700,000 or more, according to analyst Ming-Chi Kuo. U.S. demand has dropped faster than expected, prompting order cuts ahead of international launches and likely delaying any new hardware iterations until after 2025.
Meta just expanded its Ray-Ban smart glasses line with new frames and rolled out multimodal AI to all users across the US and Canada. The glasses can now answer questions about your surroundings and share your first-person view during WhatsApp or Messenger video calls. Why it matters: Wearable AI is officially moving from basic audio prompts to real-time vision.
Adobe just launched its Firefly Image 3 foundation model, bringing a batch of new AI features to the Photoshop desktop beta. The highlight is Reference Image, allowing creators to upload a sample image to guide style and color instead of wrestling with complex prompts. Adobe confirmed uploaded reference images won’t be used to train its models and remain safe for commercial use.
Microsoft just launched Phi-3 Mini, a lightweight 3.8B-parameter AI model trained on synthetic bedtime stories. Microsoft claims it rivals models 10 times its size, like GPT-3.5, while running on phones and laptops. It is available now on Azure, Hugging Face, and Ollama, with 7B and 14B versions coming soon.
OpenAI just rolled out enterprise-grade controls for its API customers, introducing a Batch API that cuts prices by 50% for non-urgent tasks completed within 24 hours. The Assistants API also received a massive boost, expanding its file search limit by 500x up to 10,000 files alongside native streaming and project-level workspace controls.
OpenAI just joined forces with competitors including Google, Meta, and Microsoft to adopt new "Safety by Design" principles aimed at protecting children. Led by child protection group Thorn, the alliance pledged to scrub child abuse material across training, deployment, and ongoing platform maintenance, alongside releasing annual progress updates.
Hugging Face just launched the Open CoT Leaderboard to measure how much chain-of-thought prompting actually boosts LLM accuracy, rather than tracking absolute benchmark scores. Early results from 30 models reveal small 7B models often gain more from step-by-step reasoning than giants, though CoT can occasionally backfire and degrade accuracy. It gives developers a real look at prompt effectiveness beyond raw capability.
Microsoft just hired former Meta VP of infrastructure Jason Taylor as corporate VP and deputy CTO to build its next wave of AI supercomputing systems. Taylor previously spent over a decade managing Meta's AI data infrastructure and hardware budgets. Why it matters: As frontier models demand massive compute, Microsoft is securing veteran talent to scale its foundational data center architecture.
Meta just opened up its Quest operating system, dubbed Meta Horizon OS, to third-party hardware makers including Asus ROG and Lenovo. It basically mimics Google's Android playbook for mixed reality to counter Apple's closed Vision Pro ecosystem.
Microsoft AI CEO Mustafa Suleyman just delivered a TED Talk offering an eye-opening mental model for artificial intelligence, arguing that we are essentially creating a "new digital species." He pointed out that even leading AI creators struggle to define where the technology is heading, urging the world to look closely at this pivotal moment.
Apple reportedly acquired Datakalab, a Paris-based startup specializing in low-power computer vision and on-device AI compression algorithms. Finalized in December, the deal brings Datakalab's patents and engineering talent into Apple's ecosystem. Why it matters: As Apple prepares its AI push for iOS 18 and Vision Pro, running powerful models directly on your hardware without draining battery or risking privacy is the core battleground.
Adobe researchers just unveiled VideoGigaGAN, a generative AI model that can upscale blurry videos up to eight times their original resolution. It manages to deliver sharp, realistic details like skin texture without the flickering artifacts typical of video GANs. It's currently just a research preview, with no word yet on a Premiere Pro rollout.
Researchers just open-sourced JAT, a multi-purpose Transformer agent built to reproduce and improve upon Gato. Running on a single network, it tackles 157 tasks across Atari games, navigation, and robotic control, matching 65.8% of expert performance while supporting basic text and vision tasks. Why it matters: It opens up generalist embodied AI research with fully public models and datasets.
Drake just dropped a new diss track featuring AI-generated vocals of Tupac Shakur and Snoop Dogg to pressure Kendrick Lamar into responding. Listeners speculate he recorded the verses himself before applying AI voice-cloning filters. Why it matters: generative audio is no longer just a bedroom novelty, it is now an active weapon in mainstream pop-culture battles.
TED just partnered with artist Paul Trillo and OpenAI to envision what the conference might look like in 40 years, completely generated using the unreleased text-to-video model Sora. The concept video highlights how experimental creators are pushing the boundaries of AI video ahead of a public rollout.
Stanford’s HAI just released its seventh annual AI Index report, revealing that AI has surpassed or matched human baselines across almost all traditional benchmarks. In competition math alone, model accuracy jumped from 6.9% to 84.3% in two years, leaving researchers scrambling to design harder tests. Why it matters: Systems are mastering existing exams so fast that we now need new tests simply to see where humans still hold an edge.
The International Olympic Committee just unveiled its Olympic AI Agenda, laying out plans to use AI across judging, talent scouting, and broadcasting. Demos featured algorithms analyzing diving mechanics in under 0.1 seconds and Intel using phones to scout young athletes in Senegal, with initial rollouts planned for Paris 2024.
