Developers Say GPT-5 Is a Mixed Bag
Summary OpenAI’s GPT-5 has landed with mixed reactions from software engineers. Developers report the model is strong at technical reasoning, planning and helping think through complex problems, but often falls…
Agentic AI Use Cases for Security Soar, but Risks Demand Close Attention
Agentic AI Use Cases for Security Soar, but Risks Demand Close Attention Summary Agentic AI adoption is surging across security teams and wider industry, offering automation that speeds triage, groups…
AI helps assemble ‘brain’ of future quantum computer
Summary A team led by researchers including Jian-Wei Pan used an AI model to compute laser patterns (holograms) that rapidly rearrange rubidium atoms into ordered neutral-atom arrays — the potential…
Dodgy Huawei chips nearly sunk DeepSeek’s next-gen R2 model
Dodgy Huawei chips nearly sunk DeepSeek’s next-gen R2 model Summary DeepSeek attempted to train its next-generation R2 model on Huawei’s Ascend accelerators after pressure from Chinese authorities. Months of effort…
Forget Foxconn the iPhone factory. AI’s made it a server-slinger first and foremost
Forget Foxconn the iPhone factory. AI’s made it a server-slinger first and foremost Summary Foxconn’s Q2 results show a clear business shift: its Cloud and Networking Products division — driven…
Tencent doesn’t care if it can buy American GPUs again – it already has all the chips it needs
Tencent doesn’t care if it can buy American GPUs again – it already has all the chips it needs Summary Chinese web giant Tencent told investors during its Q2 earnings…
LLM chatbots trivial to weaponize for data theft, say boffins
LLM chatbots trivial to weaponise for data theft, say boffins Summary Researchers from King’s College London presented a paper at the 34th USENIX Security Symposium showing that system-prompt engineering can…
Are you willing to pay $100k a year per developer on AI?
Are you willing to pay $100k a year per developer on AI? Summary Steven J. Vaughan-Nichols argues that the economics of AI are being misrepresented: executives talk about AI-driven efficiency…
Little LLM on the RAM: Google’s Gemma 270M hits the scene
Little LLM on the RAM: Google’s Gemma 270M hits the scene Summary Google has released Gemma 3 270M, a compact 270-million-parameter language model designed for on-device use and rapid iteration.…
Oracle cuts cloud jobs with Seattle hit hard as AI spending soars
Oracle cuts cloud jobs with Seattle hit hard as AI spending soars Summary Oracle issued Worker Adjustment and Retraining Notification (WARN) layoff notices this week for more than 300 employees…