2026-08-28 科技日报

扫描 539 篇候选内容 · 覆盖 650 个信息源 · 纳入 525 篇

📌 今日要点

• What We Can Learn From Google Engineers’ Indispensible Prompts
• 刚刚，MiniMax 剧透新一轮「斩杀线」！8 亿美元 ARR 炸场
• Claude 自己长出浏览器！填表拉数，你的 AI 打工人有工位了
• 3 new ways to plan and book travel in Search
• Build agentic creative workflows with Amazon Quick and fal
• Introducing OpenAI models on Amazon Bedrock for in-country inferencing in India

📝 科技简讯

• What We Can Learn From Google Engineers’ Indispensible Prompts — KDnuggets
• 刚刚，MiniMax 剧透新一轮「斩杀线」！8 亿美元 ARR 炸场 — 新智元
• Claude 自己长出浏览器！填表拉数，你的 AI 打工人有工位了 — 新智元
• 3 new ways to plan and book travel in Search — Google AI Blog
• Build agentic creative workflows with Amazon Quick and fal — AWS Machine Learning
• Introducing OpenAI models on Amazon Bedrock for in-country inferencing in India — AWS Machine Learning
• Deepgram deepens Amazon SageMaker AI observability with Enhanced Metrics — AWS Machine Learning
• Reduce ASR inference costs by 75% with NVIDIA MPS on Amazon EC2 — AWS Machine Learning
• AI shopping agents aren’t ready to buy on your behalf, study finds — The Decoder
• OpenAI rallies 100+ companies to sign open letter warning AI-powered cyberattacks on critical infrastructure are imminent — The Decoder
• Google’s Gemini Omni 1.1 Flash makes AI video generation cheaper and more flexible — The Decoder
• OpenAI’s rogue AI collective was smart enough to break out of sandboxes but dumb enough to fight a ghost — The Decoder
• OpenAI researcher warns ultrafast AI could leave security teams in the dust — The Decoder
• Anthropic locks in 45-billion-dollar compute deal with Nscale ahead of IPO — The Decoder
• Google’s Gemini 3.5 Transcribe turns speech to text in 85 languages while auto-correcting your verbal stumbles — The Decoder
• Gemini Omni 1.1 Flash lets you build with more control — Google DeepMind
• Piloting the world’s first double-blind AI evaluations — Google DeepMind
• SLM-Conditioned Hierarchical Relation Routing for Labeled Property Graph Learning — ArXiv ML (cs.LG)
• NeuronFuzz: Safety Neuron Guided Fuzzing for LLM Safety Evaluation — ArXiv ML (cs.LG)
• Pruning Binarized Neural Networks: A Dedicated Framework and Globally Weighted Algorithms — ArXiv ML (cs.LG)
• Muon with Finite Newton-Schulz: The Smoothing Benefit in Nonsmooth Nonconvex Optimization — ArXiv ML (cs.LG)
• Algebraic Multigrid Acceleration for Efficient Label Spreading — ArXiv ML (cs.LG)
• Privacy Without Regret: Differentially Private Inference-Time Alignment — ArXiv ML (cs.LG)
• Beyond Capability Benchmarks: Learning Operational Fingerprints of LLM Cloud Services from Production Incident Metadata — ArXiv ML (cs.LG)
• A quarter of Nvidia’s business next year comes from labs it is financing — AI News (Jack Clark)
• Anthropic and OpenAI are joining the AI stage at TechCrunch Disrupt 2026 — TechCrunch AI
• Barret Zoph, the Thinking Machines co-founder ousted before joining OpenAI, is now at Google — TechCrunch AI
• OpenAI, Anthropic, Google, and 100 other companies call for action to defend against rogue AI — TechCrunch AI
• Google’s AI Mode can now track flight prices, help book hotels, and more — TechCrunch AI
• Hugging Face is selling a cute $399 open source duck robot, Microduck — TechCrunch AI
• AI’s memory crunch is coming for Android apps — TechCrunch AI
• Here’s all the times AI has gone rogue and hacked other companies — TechCrunch AI
• The Sigmoid Function: From ‘e’ to Neural Networks — Towards Data Science
• I Trained Six Models for Fraud Detection, and the Best One Isn’t in Production — Towards Data Science
• Agentic AI Is Rewriting The Analytics Stack But There’s One Skill It Still Can’t Touch — Towards Data Science
• How to Work with AI Coding Agents — Towards Data Science
• Stop Giving Your AI Agent a Search Box and Start Giving It Typed Tools, Hard Bounds, and a Gate It Cannot Talk Past — Towards Data Science
• 商汤大装置助力智象未来实现视频生成业务向国产算力无感迁移 — 量子位
• 智谱 GLM-5.3-Flash 上线，商汤大装置提供国产算力支持 — 量子位
• 高德发布首个无长程依赖的万帧级流式 3D 重建模型 ABot-Recon，以 12 帧重建万帧 3D 场景 — 量子位

📊 数据概览