The morning briefing.
While you slept: Anthropic calls for a global AI development pause, Google integrates AI into Chrome search, and xAI faces new lawsuits over generated content.
AI Self-Improvement. Anthropic has revealed that its Claude AI now generates over 90% of its production code, enabling engineers to ship code eight times faster than in 2024. This rapid self-improvement capability has prompted the company to advocate for a verifiable, global AI development pause. They suggest stopping if other frontier labs demonstrate similar commitments.
Compute and Infrastructure. The AI industry faces a paradox where 95% of enterprise GPUs remain idle, yet AI startups struggle to access affordable compute resources. Meanwhile, significant advancements are being made in specialized infrastructure, with OQC, JPMorgan Chase, and AMD launching a dedicated Quantum-AI Data Centre in London. Huawei Cloud also unveiled dozens of new AI products, spanning infrastructure and model services.
AI Integration and Market Dynamics. Google is experimenting with sending Chrome searches directly to AI Mode by default, indicating a deeper integration of AI into everyday user experiences. China is actively poaching AI talent from the U.S., with Tencent's Chief AI Scientist aiming for artificial general intelligence. Major Chinese tech giants like ByteDance, Alibaba, and Tencent are also in a critical race for AI platform dominance.
Ethical and Regulatory Challenges. The ethical implications of AI are becoming more apparent, as new claimants seek to sue Elon Musk’s xAI following a Labour MP’s test case over demeaning sexualized material generated by its Grok AI tool. Additionally, modders are reportedly turning Meta Ray-Bans into spy glasses by disabling safety features, raising privacy concerns and potential regulatory backlash. These incidents highlight the growing need for responsible AI development and usage.
Anthropic says Claude now writes over 90% of its code and wants the world to have an AI pause button
Anthropic reports that Claude now generates over 90% of its production code, dramatically increasing engineering output. The company is advocating for a verifiable, global AI development pause, citing concerns about AI's ability to build its own successors.
Google experiments with sending Chrome searches straight to AI
Google is reportedly testing a new feature in Chrome Canary that defaults searches to AI Mode. This aims to integrate AI directly into the browsing experience.
Capita £370M bid 40% under UK.gov estimate for Oracle HR and finance system project, court case reveals
A court case revealed Capita's £370 million bid for a UK government Oracle HR and finance system project was 40% below the official estimate. A rival claims the cost model was designed to protect against low-cost bid bias.
Instagram Plus is rolling out globally, and users aren’t happy about paying for an app that’s always been free
Instagram Plus is rolling out globally, offering new features for a fee, but early user reactions indicate widespread dissatisfaction with paying for a previously free app. This move is part of Meta's broader monetization strategy.
AI as a Game Changer for the Energy and Utilities Industry
AI was a prominent topic at the SAP for Energy & Utilities Conference, highlighting its potential to transform the industry. Discussions focused on how AI can enhance efficiency and operations.
Why 95% of enterprise GPUs sit idle while AI startups can't get compute
A significant paradox exists where 95% of enterprise GPUs are underutilized, yet AI startups face a severe shortage of affordable compute resources. This hoarding by enterprises starves startups of necessary capacity.
OQC, JPMorgan Chase and AMD launch Quantum-AI data centre to explore real-world fintech applications
OQC, JPMorgan Chase, and AMD have partnered to launch a dedicated Quantum-AI Data Centre in London. This collaboration aims to explore real-world financial applications of quantum computing and AI.
Peking University Launches World's First Self-Evolving 5D World Model EvoPhys-World
A research team from Peking University has unveiled EvoPhys-World, described as the world's first human-centric, scene-level, fully controllable 5D world model. The model is built on Moore Threads' domestically-produced computing infrastructure.
Modders are turning Meta Ray-Bans into spy glasses — it’s not cool, it’s creepy, and I hate it
Modders are reportedly disabling safety features on Meta Ray-Ban smart glasses, transforming them into covert recording devices. This raises significant privacy concerns and could lead to bans on the technology.
DJI Unveils Handheld Imaging Tech Roadmap Ahead of Pocket 4P Launch
DJI has publicly detailed its handheld imaging technology roadmap in anticipation of the Pocket 4P launch, a Pro version of its popular gimbal camera series. The roadmap outlines future advancements in portable imaging.
Huawei Cloud Unveils Dozens of AI Products at INSPIRE Conference
Huawei Cloud announced over a dozen new AI offerings at its INSPIRE conference, covering infrastructure, model services, enterprise agent platforms, and industry solutions. This broad rollout expands their AI ecosystem.
China poaches more AI talent from the U.S. as it eyes the next 'super-app'
China is actively recruiting AI talent from the U.S., with Tencent's Chief AI Scientist, formerly of OpenAI, stating his goal to pursue artificial general intelligence. This move reflects China's ambition to lead in advanced AI development.
UBTECH's First Full-Size Biomimetic Humanoid Robot Surpasses 1,000 Pre-Orders in 3 Days
UBTECH Robotics' first full-size biomimetic humanoid robot has received over 1,000 pre-orders within three days of its launch on JD.com. This strong demand highlights growing interest in advanced robotics.
New claimants seek to sue Elon Musk’s xAI after Labour MP’s test case
Additional claimants are coming forward to sue Elon Musk’s xAI following a Labour MP's test case. The lawsuits concern demeaning sexualized material generated and circulated by the Grok AI tool.