Daily Brief Archive
Explore our complete collection of AI insights. Every edition, every breakthrough, all in one place.
Google launches native Gemini Mac app and releases Flash TTS speech model supporting 70+ languages; Baidu open-sources 8B parameter ERNIE-Image text-to-image model, topping open-source rankings on GenEval; Snap lays off 1,000 employees, explicitly citing AI-driven efficiency gains to save $500 million annually
NVIDIA releases open-source quantum AI model Ising, boosting decoding speed by 2.5x and accuracy by 3x; AWS launches Bio Discovery drug development platform, reducing antibody design time from one year to weeks; Microsoft launches self-developed image model MAI-Image-2-Efficient, cutting costs by 41% and accelerating shift away from OpenAI dependence
Stanford's 2026 AI Index shows global AI adoption surpassing 50% for the first time, with U.S.-China model performance gaps nearly closed; Harvey raises $200 million at an $11 billion valuation, hosting over 25,000 legal AI agents; global AI compute shortage intensifies, with NVIDIA's latest chip spot prices rising nearly 50% in weeks
MiniMax open-sources M2.7 model with SWE-Pro score of 56.22% and compatibility with Huawei Ascend chips; SoftBank partners with NEC and seven other Japanese firms to establish a national AI foundation model company backed by 1 trillion yen in government funding; Mistral secures $830 million to build a Paris data center, targeting 200MW of compute capacity by 2027
Vidu AI secures nearly 2 billion RMB Series B funding led by Alibaba Cloud, valued at 12 billion RMB; Morgan Stanley report shows global AI token usage surged 250% in Q1, compute demand systematically outpacing supply; Cisco acquires Galileo and Astrix Security to build a three-layer AI agent security architecture
Five departments release AI human-like interaction management measures effective July 15, banning virtual intimate relationship services for minors; US Treasury and Federal Reserve convene Wall Street CEOs to address Anthropic Mythos cybersecurity threat, where model's vulnerability exploitation success rate jumps from 0.9% to 72.4%; Qwen 3.5-Omni full-modal model sets 215 SOTA records, supporting 256K token context and voice cloning in 113 languages
Meta signs $21 billion AI compute deal with CoreWeave, cumulative contract value reaches $35 billion; open-source video model HappyHorse-1.0 tops global blind evaluation leaderboard, 15B parameters outperform ByteDance's Seedance 2.0; Musk overhauls xAI engineering team, appoints SpaceX Starlink executive as president
Meta releases closed-source model Muse Spark, marking a strategic shift from its Llama open-source approach; Anthropic's Claude Mythos discovers thousands of zero-day vulnerabilities, leading to restricted access and the launch of Project Glasswing with over 40 partner companies for defensive collaboration; Safetensors joins the PyTorch Foundation, becoming the industry standard for model weight storage
Anthropic's annualized revenue surges from $9 billion to $30 billion, surpassing OpenAI, and secures 3.5GW of Google TPU compute; Zhipu's GLM-5.1, a 754B-parameter MIT-licensed open-source model, outperforms GPT-5.4 and Opus 4.6 on SWE-bench Pro; Anthropic refrains from public release of Claude Mythos due to its potent cyberattack capabilities, launching Project Glasswing with 40+ partners for defensive collaboration