Daily AI Briefing
The 120 stories that mattered in AI, curated and summarized from dozens of sources by AIBriefs.
Claude Opus 4.8 scores 69.2% on SWE Bench Pro (up from 64.3%) and introduces dynamic workflows in Claude Code. Fast mode runs at 2.5x speed and is now three times cheaper.
M3, a frontier open-weight model for coding and agentic workflows, is now available. Weights are open and accessible through platforms like Venice, Together Compute, and vLLM. The community is already shipping optimizations for faster inference.
Anthropic raised $65B in Series H at a $965B post-money valuation, with $47B revenue run-rate. The company also released Claude Opus 4.8, which fixes issues from Opus 4.7 and achieves SOTA on economically relevant benchmarks.
The 550B-parameter Mixture-of-Experts model with 55B active parameters achieves 5x higher throughput and up to 30% cost reduction for long-running agent tasks. It features a hybrid Mamba-2 MoE Transformer architecture with a 1M context window and is available on Hugging Face.
The US Commerce Department issued an export control directive on June 12, 2026, citing national security, forcing Anthropic to disable Fable 5 and Mythos 5 for all users. Anthropic stated the directive lacked specific details and that the alleged vulnerabilities are minor and also discoverable by other publicly-available models.
Trump's administration banned Anthropic's Mythos and Fable models, prompting Macron to seek workarounds. CEOs including Amodei and Hassabis called for a U.S.-led AI coalition at the G7 summit.
Anthropic confidentially submitted a draft S-1 to the SEC for a proposed IPO. The number of shares and price have not been set, and the offering depends on market conditions.
Gemini 3.5 Flash outperforms Gemini 3.1 Pro on benchmarks while running 4x faster. Other launches include Antigravity 2.0 desktop app, Managed Agents, and native Android vibe coding in AI Studio.
The U.S. Commerce Department invoked an export control directive to force Anthropic to take down Fable 5 and Mythos 5, citing unspecified national security concerns. Anthropic believes it may relate to a reported guardrail bypass but hasn't received details; the letter remains unreleased.
Anthropic raised $65 billion in Series H funding at a $965 billion post-money valuation, surpassing OpenAI as the most valuable private AI company. Its run-rate revenue has hit $47 billion, and the company expects its first profitable quarter soon.
Gemini Spark rolls out to Google's $99/month AI Ultra plan, integrating natively with Gmail, Calendar, Drive, and other Google apps. The agent runs on a dedicated VM powered by Gemini 3.5 Flash and Antigravity, with early hands-ons praising its effectiveness but noting privacy concerns over personal data access.
At HPE Discover Las Vegas, NVIDIA and HPE announced the HPE ProLiant Compute DL394 Gen12 with NVIDIA Vera CPU, available in 2027, and the NVIDIA Agent Toolkit for HPE Private Cloud AI. The Vera CPU is designed for agentic AI workloads, and Vera Rubin NVL72 rack-scale system is ramping to production.
Google is overhauling its search box for the first time in 25 years, introducing AI-powered suggestions, dynamic expansion, and "information agents" that autonomously gather data. AI Mode usage has doubled quarterly, with over 1 billion monthly users.
Antigravity 2.0 includes a desktop app for orchestrating multiple agents, a CLI tool, and an SDK for custom workflows. Google also introduced an AI Ultra plan at $100/month with 5x higher limits. The platform is powered by the new Gemini 3.5 Flash model.
Ideogram 4.0 is the company's first open-source text-to-image model, trained from scratch. It introduces structured JSON prompting, multilingual text rendering, layout controls, and native 2K resolution, and ranks #1 on Design Arena among open models.
Kuaishou's Kling AI seeks $2 billion in funding at an $18 billion valuation. General Atlantic is in talks to lead the round.
Hark, led by Brett Adcock, raised $700M at a $6B valuation from investors including Nvidia and Intel Capital. The secretive startup, which has revealed little about its 'universal' AI interface, plans to release multimodal models this summer and eventually hardware devices.
OpenAI has confidentially submitted a draft S-1 registration statement to the SEC for a potential IPO. The company has not yet decided on the timing for going public.
Moonshot AI's Kimi now valued at $30 billion pre-money, up from $4.3 billion in six months. Its annualized revenue exceeded $200 million in April following the K2.5 model launch.
Noam Shazeer, co-lead of Google's Gemini, is joining OpenAI, according to Reuters. The move marks a significant personnel shift in the AI research landscape.
Google announces Gemini 3.5 Flash, Neural Expressive design, Gemini Omni video model, Daily Brief agent, and Gemini Spark 24/7 personal AI agent. The app now serves over 900 million monthly users across 230 countries.
Anthropic sued the Defense Department in March after being declared a supply chain risk. Opening arguments in the DC court case are set for today.
OpenAI's method simulates deployment with recent user requests to predict rare failures before release. Validation using WildChat data showed accurate predictions of real-world failure rates. The approach extends to agentic coding by simulating tool calls.
The new $99.99 Google Home Speaker, first smart speaker in six years, goes on sale June 25. It replaces rigid Google Assistant commands with conversational Gemini interactions for smart home control and natural-language queries.
SpaceX's IPO filing reveals Anthropic pays $1.25 billion per month through May 2029. The deal amounts to roughly $15 billion annually, far larger than previously reported.
NVIDIA's Blackwell platform achieved the fastest training time at the largest reported scale in MLPerf Training 6.0. The results highlight full-stack innovation across silicon, systems, networking, and software.
Spotify and Universal Music Group announced a licensing deal enabling fans to create AI-powered covers and remixes of songs as a paid add-on for Premium users. The initiative is designed to generate additional revenue for artists and songwriters, with the companies emphasizing consent, credit, and compensation.
Midjourney will livestream its first hardware project announcement on Wednesday June 17 at 6pm PT from San Francisco. The company has been hiring for hardware and electrical engineering roles, and CEO David Holz has previously hinted at a physical AI device.
Xiaomi MiMo-V2.5-Pro UltraSpeed achieves over 1,000 tokens/s decode speed on a 1-trillion-parameter MoE model using a standard 8-GPU server. The API is available June 9–23 at 3x the standard price for 10x speed.
NVIDIA XR AI is now available in public beta, providing a developer library to build multimodal AI agents for AR glasses and XR devices. It integrates NVIDIA NeMo Agent Toolkit for tool use and multi-agent coordination, running on NVIDIA accelerated computing platforms.
Sensor Tower's State of AI 2026 report shows ChatGPT's audience share fell to 46% in May, down from over 50% in January. Google Gemini holds 28% and Claude 10%, while ChatGPT still leads with 1.1 billion monthly users.
Analysis of 400k+ Claude Code sessions shows domain experts succeed more, but the gap from intermediate users is small. Debugging time fell by nearly half over 7 months, and typical task value rose ~25%.
Anthropic's Mythos Preview has helped partners find over 10,000 vulnerabilities within a month of Project Glasswing's launch. Cloudflare found 2,000 bugs (400 critical/high); Mozilla fixed 271 in Firefox, 10x more than before. The model remains unreleased due to lack of safeguards.
The new 'gpt-bidi-1' model is designed to listen and speak simultaneously, handling interruptions naturally. OpenAI touts it as a 'major leap in intelligence', with users able to choose between Instant, Medium, and High modes.
Vicki Boykis reports that with Google's Gemma 4 family, local models achieve ~75% the accuracy/speed of frontier models for agentic coding workflows. She used gemma-4-26b-a4b via LM Studio to refactor Python code and write unit tests on a 2022 M2 Mac.
Anthropic is projected to hit $10.9 billion in revenue in Q2 2026, according to a source. This would be the company's first profitable quarter.
Nvidia and LG Group are building an AI factory for robotics, autonomous driving, and smart manufacturing. The collaboration integrates Nvidia's AI platforms with LG's consumer electronics and robotics expertise to develop home robots like CLOiD and modular robotics platforms.
Suno raised $400m in Series D funding at a $5.4bn valuation, more than doubling its valuation from $250m at $2.45bn in November 2025. The company teased its first music model developed in partnership with the music industry, while still facing lawsuits from Universal Music Group and Sony Music.
NVIDIA's GTC Live pregame show and Jensen Huang's keynote at COMPUTEX 2026 will take place on Monday, June 1 at 9 a.m. Taipei Time. The event will unveil the latest breakthroughs in AI and accelerated computing.
eve is an open-source agent framework now in public preview. It treats agents as directories of files with built-in production features like durable execution, sandboxed compute, and human-in-the-loop approval.
Hermes Agent now supports asynchronous subagents that no longer block the chat. Users can approve memory/skill updates via Write Gate and build custom agents with a GUI. Run `hermes update` to access.
NVIDIA's Nemotron-Labs introduces diffusion-based language models designed for near-instantaneous text generation. The method uses iterative denoising for parallel text synthesis, potentially surpassing autoregressive models in speed.
Developed with 173 scientists from biotech and pharma, the benchmark includes 750 expert-authored tasks across seven biological research areas. LifeSciBench aims to measure how well AI systems handle real-world life science research decisions.
Cerebras founder Andrew Feldman details the company's $63 billion IPO story and the design of its dinner-plate-sized AI chip in a No Priors podcast interview. He contrasts Cerebras' approach with competitors Nvidia and AMD, discussing the chip's unique architecture and market strategy.
Two papers propose process-reward optimization for training computer use agents (CUAs), addressing limitations of sparse reward and costly live environment interaction. Methods like filtered behavior cloning and multi-granularity reward models improve agent performance on complex digital workflows.
Manus is reportedly exploring a $1 billion fundraising round to buy back Meta's stake and undo Meta's takeover of the AI agent company. The move would reverse Meta's acquisition earlier this year. No financial terms have been disclosed yet.
DeepSeek's founder publicly states AGI as the company's goal while a $10 billion funding round progresses. The round values the Chinese AI firm at a significant multiple, though exact valuation remains undisclosed.
Suno raised capital at a $5.4 billion valuation, doubling from its $250 million round seven months ago. The AI music startup is one of the most valuable in the generative audio space.
Microsoft and Ernst & Young (EY) are committing $1 billion to help enterprise clients adopt AI technologies. The partnership will focus on deploying Microsoft's AI tools across EY's client base. No specific breakdown of the investment has been disclosed.
AI coding startup Cursor has reached a $3 billion annual sales rate. Bloomberg reports SpaceX has an agreement to acquire Cursor for $60 billion later this year, or pay $10 billion for their work together.
Google announced at I/O 2026 that 900 million people use its Gemini assistant and 50 billion images have been generated. The company is embedding AI agents directly into Search via an 'intelligent search box' rolling out today. New Android-powered smart glasses are coming this fall.
Anthropic is in negotiations to adopt Microsoft's AI chips, according to a report from The Information. The story was first reported by Bloomberg.
The Pentagon is testing rival AI models as it moves to replace Anthropic's technology, according to a Bloomberg report. No details on the models or timeline were disclosed.
AMD plans to invest $10 billion in Taiwan's AI sector, focusing on chip packaging and manufacturing for next-generation AI infrastructure. The investment aims to advance partnerships with Taiwanese companies.
The 1-trillion-parameter model achieves over 1000 tokens/s decode speed in collaboration with TileRT. The API launches at a limited-time price (3× MiMo-V2.5-Pro cost for ~10× speed) available from June 9–23, 2026 via application.
Investment firm Whale Rock is publicly highlighting Anthropic's user growth potential as the AI company prepares for its IPO. The comments come from Whale Rock's Alex Sacerdote, according to Bloomberg.
Mark Zuckerberg stated he does not believe in a centralized future where a few institutions advance AI. He advocates for building AI as a tool and putting it in individuals' hands to drive progress.
Anthropic CEO Dario Amodei said he left OpenAI because he didn't trust Sam Altman. 'Why argue with someone' you don't trust, he explained in a recent interview.
French companies have submitted bids to host a €10 billion AI gigafactory in Europe, according to a Bloomberg report. The proposed site would be one of the largest AI infrastructure projects on the continent.
Standard Chartered plans to cut thousands of jobs as part of an AI-driven efficiency initiative. The move reflects the growing trend of automation in the banking sector.
Poland is investing $11 million in AI voice company ElevenLabs to establish a technology hub in the country. The investment aims to boost Poland's AI ecosystem and create local jobs.
ASML CEO discusses strong AI chip demand, potential for data centers in space, and Elon Musk's Terafab supercomputing project. No financial details disclosed.
Elicit founders discuss their approach to building world models and process supervision for scientific reasoning. They explain how domain-specific reasoning primitives maintain trust as AI models become more capable.
Anthropic has deployed noted security researcher Nicholas Carlini to engage with government officials on AI safety, according to a WSJ report. Carlini's role focuses on addressing policymakers' concerns about advanced AI risks and building confidence in safety measures.
The acquisition creates the world's largest human experience dataset. Qualtrics plans to use its AI platform to predict patient experiences.
Researchers engineered a van der Waals crystal that mimics neuronal synapses using light-driven learning. The material could enable ultrafast, energy-efficient neuromorphic computing.
The update brings Screen reactions for selfie-in-screen recordings, Gemini Omni-powered video editing via natural language, and in-app music generation with lyrics. Pixel devices also get Android 17 and Wear OS 7.
Sanctuary AI demonstrated a wire-plugging task at an unnamed Tier 1 automotive supplier. The validation showcases the application of physical AI on existing robotic platforms. This marks progress toward commercial deployment in manufacturing.
ABC-130k contains 130,000 bimanual manipulation trajectories collected on two-arm YAM stations, distributed as MCAP files with subtask annotations. It is the largest open-source robotics teleoperation dataset, with train/val splits and accompanying code.
Updated i2L method turns one or more reference images into a style LoRA in a single forward pass, eliminating the need for explicit per-style training. The approach achieves efficient style transfer while preserving reference-specific appearance.
Epoch AI's Gradient Updates newsletter explores developing a standardized taxonomy for categorizing AI R&D tasks, similar to O*NET for occupations. It identifies key trends and challenges in mapping the diverse landscape of AI research.
Altman admitted he was "pretty wrong" about AI eliminating entry-level white-collar jobs, while Amodei now says automation may expand work. Both CEOs are reportedly preparing for their companies' IPOs.
On April 19, 2026, the Honor Lightning humanoid robot ran a half-marathon in 50 minutes and 26 seconds, beating the human world record by 7 minutes and the best robot time from 2025 by almost two hours. The article explores the engineering breakthroughs—including energy-efficient hardware and adaptive gait control—that enabled this leap in performance.
The startup's AI assistant aims to reach 50% of physicians in the region. Bloomberg's coverage explores the company's plans and market.
Kirkland & Ellis announced a $500 million investment in AI, the largest by a law firm. Ken Crutchfield's analysis argues the real bet is on data and knowledge management, not just AI technology.
Amazon's Nova2 model trails the latest releases from LLM leaders OpenAI and Anthropic. However, the company's AI chief believes Amazon can close the gap within the next year.
Autonomique's mobile manipulator robots and AI platform have been deployed at a Canadian Tier 1 automotive supplier. The system aims to address labor shortages and production complexity.
Claude Mythos Preview found over 10,000 high- or critical-severity vulnerabilities across critical software projects since Project Glasswing launched last month. Progress is now limited by how quickly the team can verify, disclose, and patch the vulnerabilities found by AI.
Prime Minister Kristen Michal backed a proposal to issue AI agents a personal identification code separate from their owners. The move aims to regulate AI autonomy and accountability.
Pramaana Labs raised a $27 million seed round led by Khosla Ventures to apply formal verification to AI, targeting high-stakes fields like law, drug discovery, and tax preparation. The approach aims to mathematically prove model correctness in safety-critical applications.
President Trump plans to sign an AI cybersecurity directive as early as Thursday, per Bloomberg. The directive aims to bolster cybersecurity for AI systems.
Microsoft has canceled most of its direct Claude Code licenses, moving engineers to GitHub Copilot CLI instead, according to The Verge. The decision comes as Uber's CTO revealed the firm exhausted its 2026 AI coding tools budget in four months, signaling AI adoption costs are proving a bottleneck.
GLAAD's report finds AI systems can amplify anti-LGBTQ bias, misinformation, discrimination, and privacy harms. The organization warns that the risk is growing as AI adoption increases.
The Linux Foundation is launching an initiative with Google, Microsoft, and OpenAI to create a trust layer for AI. The effort aims to establish open standards and governance frameworks for AI transparency. This collaboration brings together major competitors to address trust and accountability in AI development.
New feature lets agents autonomously search and discover resources on the Hugging Face Hub. Enables more dynamic and self-directed agent behavior.
Published in Nature, the research demonstrates AMIE's potential for managing chronic conditions through AI-guided dialogue. The study highlights the system's ability to conduct diagnostic conversations accurately.
The startup raised $6 million in seed funding to detect and stop dangerous AI agentic behavior in real time. It aims to address security risks from autonomous AI agents.
SoftBank founder Masayoshi Son's heavy investment in OpenAI has sparked concern among insiders, who view his devotion to CEO Sam Altman as risky. The bet highlights growing Japanese capital influence in AI.
Odyssey raised funding at a $1.45B valuation backed by Amazon and other investors. The round underscores growing interest in world models, considered the next frontier beyond large language models.
Microsoft Research proposes Next-Latent Prediction Transformers, where models learn to predict their own next latent state instead of the next token. This addresses the myopia of next-token prediction. The preprint is available on Microsoft Research.
ERA uses Gemini to write and optimize scientific code, tested on genomics, public health, and more benchmarks. It achieves expert-level performance and is part of Computational Discovery, now available through a trusted tester program in Google Labs.
The RTX Spark (N1X) chip features 20 Arm CPU cores, a Blackwell GPU with 6,144 CUDA cores, 128GB LPDDR5X RAM, and 300 GB/s bandwidth. It supports 120-billion-parameter models with million-token contexts and will debut in fall on Windows PCs from major manufacturers.
Google announced information agents, Gemini Spark, Android Halo, and Daily Brief at I/O 2026, with information agents launching this summer for Pro/Ultra subscribers (Ultra $100/month). The analysis questions whether consumers will adopt the overlapping ecosystem of products.
Exa Labs raised $250 million at a $2.2 billion valuation, according to Bloomberg. Parallel Web Systems, led by former Twitter CEO Parag Agrawal, raised $100 million at a $2 billion valuation from Sequoia Capital.
Coherent broke ground on an expanded manufacturing building in Sherman, Texas, producing lasers and optical components for AI systems. The facility scales production of compound semiconductors that wire AI infrastructure together.
Harrison.Rad 1.5 uses imaging, clinical context, and prior exams to draft reports for radiologist review. The model is available for research via chat.harrison.ai and API.
Anthropic paused a billing change for Claude Agent SDK on June 15, the day it was set to take effect. The change would have billed SDK usage at API rates, but users can continue using subscription rate limits for programmatic use.
Over 200 organizations have jointly called for a ban on the use of artificial intelligence in military kill chains. The appeal highlights concerns over autonomous weapons and the lack of human control in lethal decisions.
Google introduces three architectural patterns integrating Model Context Protocol (MCP) Apps and Agent-to-User Interface (A2UI) to balance custom iframe environments with native declarative rendering. The approach enables developers to serve rich agentic UIs without sacrificing flexibility.