Jeff Dean and Google's AI Brain Trust Depart to Found Discovery Loop
In what may be the most consequential talent move in AI this year, Jeff Dean — Google's chief scientist and a figure synonymous with the company's technical DNA — has left after nearly 27 years to co-found Discovery Loop, a startup aiming to automate the scientific method itself.
Dean isn't going alone. Joining him are Sanjay Ghemawat, a Google senior fellow and Dean's legendary coding partner; Oriol Vinyals, a DeepMind vice president; and Quoc Le, co-founder of Google Brain. Together, they represent decades of foundational work on everything from MapReduce and TensorFlow to the Transformer architecture.
Discovery Loop's initial focus will be automating machine-learning research and engineering workflows, with plans to expand into drug discovery, hardware design, and clean energy. In a notable twist, Google itself is backing the venture as a founding investor and Cloud partner — a sign that Mountain View views this as a strategic spin-out rather than a hostile departure.
The exodus triggered a broader reshuffle at Google DeepMind. Demis Hassabis has stepped back from day-to-day operations to become chairman of DeepMind and chief scientist of Alphabet, while continuing to lead Isomorphic Labs. Operational control has passed to CTO Koray Kavukcuoglu, who was elevated to senior vice president and now oversees Gemini model development.
The shakeup comes at a critical moment. The same week, Anthropic revealed it has locked in roughly $71 billion in compute commitments, underscoring the arms-race dynamics driving the industry.
Sources: TechCrunch, Unite.AI, GeekWire
Gemini Crosses 1 Billion Monthly Users — Google's Fastest-Growing Product Ever
Google CEO Sundar Pichai announced on August 11 that the Gemini app has surpassed one billion monthly active users, making it Google's fastest-growing product in the company's history and its 14th product to hit the billion-user milestone, joining the ranks of Search, Gmail, Android, and YouTube.
The numbers paint a picture of an AI assistant that has woven itself into daily life: 63% of users interact via voice, one in five Gemini Live sessions now involve camera or screen sharing, more than 150 million images are generated daily, and the app can carry out automated tasks across more than 40 applications.
The milestone puts Gemini roughly two months behind OpenAI's ChatGPT, which crossed the billion-user threshold in June 2026. Notably, this count covers only the standalone Gemini app and web interface — not the separate billion-plus users who encounter Gemini through AI Overviews in Google Search.
Sources: TechCrunch, Forbes, 9to5Google
Alibaba Launches Qwen 3.8-Max: A 2.4-Trillion-Parameter Challenger
Alibaba has officially launched Qwen 3.8-Max, the largest model in its Qwen family, packing 2.4 trillion parameters in a Sparse Mixture-of-Experts architecture that activates only 95 billion parameters at inference time. The model supports text, images, and video with a context window of up to one million tokens.
Benchmark results shared by Alibaba place Qwen 3.8-Max fifth in Text Arena and second in Vision Arena, with the company claiming comparable or sometimes better scores than Anthropic's Claude Fable 5. The model is already accessible via API on Alibaba Cloud Model Studio at $2.00 per million input tokens and $6.00 per million output tokens, with open weights scheduled for release.
The launch represents a significant escalation in the global AI race. China's leading tech companies continue to close the gap with Western frontier labs, and Alibaba's decision to open-source the weights could further accelerate the commoditization of large-scale AI capabilities.
Sources: Bloomberg, MarkTechPost, TechNode
Manus Returns to Independence as Beijing Unwinds $2 Billion Meta Deal
AI agent startup Manus has confirmed it will resume independent operations as its $2 billion acquisition by Meta is formally unwound following a directive from China's National Development and Reform Commission. The regulator ordered the reversal in April 2026, citing foreign investment rules, after the deal closed in December 2025.
In a message to users posted on August 11, Manus outlined a transition plan: data created between December 29, 2025 — when the acquisition closed — and the unwind date is slated for removal in certain jurisdictions. Affected users have until August 23 to back up their data, with restoration available from August 25.
The unwind is the highest-profile casualty of growing geopolitical friction in the AI sector. Manus, which originated in China in 2022 and later relocated to Singapore, develops general-purpose AI agents for automating desktop tasks. The company now faces the challenge of rebuilding as an independent entity in an increasingly crowded AI agent market.
Sources: CNBC, South China Morning Post, TechCrunch
EU AI Act Transparency Rules Go Live — Fines Now in Play
As of August 2, 2026, the transparency obligations under Article 50 of the EU AI Act are officially enforceable, marking the most significant regulatory milestone for AI in Europe to date.
The rules apply immediately to all in-scope systems, regardless of when they were placed on the market. Key requirements include: AI-generated content must carry clear, visible labels and machine-readable marks; deepfakes must be disclosed as artificially generated or manipulated; and users of interactive AI systems like chatbots must be informed they are talking to a machine.
The teeth are real: noncompliance can trigger fines of up to €15 million or 3% of worldwide annual turnover, whichever is higher. A limited transitional period extends until December 2, 2026 for the marking and detection obligations on generative AI systems already on the market.
The European Commission has also published a voluntary Code of Practice on Transparency of AI-Generated Content, offering providers a recognized compliance path. Signatories benefit from a presumption of conformity and a more favorable enforcement posture — a carrot alongside the stick.
Sources: Cooley, European Commission, EU Digital Strategy