2026-08-10 科技日报

扫描 544 篇候选内容 · 覆盖 650 个信息源 · 纳入 533 篇

📌 今日要点

• Meoo 秒悟团队版全量上线，接入 Qwen-3.8-Max、即日起可直接订阅
• 苹果开测长鑫存储！百度、千问也一起挤进苹果供应链
• 墨芯成立稀疏计算产学研联盟，以生态协同突破产业化壁垒
• 魔幻灵巧手：半年 200 亿热钱，3 大路线，贵到几十万一只
• 3B 模型碾压英伟达谷歌后，Om AI 端侧原生 VLX 模型：小参数实现物理世界精准感知
• Old OCR text cripples language model training, and FineBooks wants to fix that at scale

📝 科技简讯

• Meoo 秒悟团队版全量上线，接入 Qwen-3.8-Max、即日起可直接订阅 — 量子位
• 苹果开测长鑫存储！百度、千问也一起挤进苹果供应链 — 量子位
• 墨芯成立稀疏计算产学研联盟，以生态协同突破产业化壁垒 — 量子位
• 魔幻灵巧手：半年 200 亿热钱，3 大路线，贵到几十万一只 — 量子位
• 3B 模型碾压英伟达谷歌后，Om AI 端侧原生 VLX 模型：小参数实现物理世界精准感知 — 量子位
• Old OCR text cripples language model training, and FineBooks wants to fix that at scale — The Decoder
• OpenAI launches GPT-5.6-Cyber to help defenders find vulnerabilities before attackers do — The Decoder
• Meta returns to open models with Zuckerberg’s plan to out-copy China and sell compute by auction — The Decoder
• Told to book a gym class, an AI agent hacked the site instead to move its user up the waitlist — The Decoder
• OpenAI acquires NextSlide to bring AI-generated presentations into ChatGPT — The Decoder
• Hidden text in a PDF is enough to steal sensitive data through Atlassian’s AI agent Rovo — The Decoder
• Fast and Accurate: An Adaptive VLA Inference Framework through Environment-aware Model Selection — ArXiv RO (cs.RO)
• LyEvO: Lyapunov-Guided Evolutionary Optimization for Safe and Robust Sim-to-Real Policy Learning — ArXiv RO (cs.RO)
• A Disturbance in the Force: Force Actuation on the RAVEN II Surgical Robot with Parallel Motor-Cable Units — ArXiv RO (cs.RO)
• SyncSBC: Decentralized Swarm Behavior Prediction for Synchronized Autonomous Control — ArXiv RO (cs.RO)
• Plan-and-Avoid: Real-Time Aircraft Trajectory Coordination in a Multi-Agent Environment — ArXiv RO (cs.RO)
• UAV3DCrop: Benchmarking 3D Reconstruction in Repeated Multi-Angle UAV Crop Surveys — ArXiv CV (cs.CV)
• Deep Evidential Regression for Sparse Forest Height Estimation from Multimodal Satellite Imagery — ArXiv CV (cs.CV)
• TransSLR: A Lightweight Transformer for Sign Language Recognition — ArXiv CV (cs.CV)
• Test-Time Adaptation with Online Personalized Energy-Based Cache for Fine-Grained Video Expression Recognition — ArXiv CV (cs.CV)
• InsertFuse: A Unified Framework for Multi-Category Reference-Guided Image Insertion — ArXiv CV (cs.CV)
• Towards Multi-Label Graph Foundation Models: from Single-Vector Representation Learning to Multi-Semantic Basis Learning — arXiv AI (cs.AI)
• EntropyMoE: Entropy-Aware Sparse Expert Routing for Tokenizer-Free LLMs — arXiv AI (cs.AI)
• Beyond Routing Weights: Faithful Response-Level Interpretation of Mixture-of-Experts Reward Models via Contribution Contrast — arXiv AI (cs.AI)
• Interpretable Unsupervised Community Detection with LLM-Symbolized Structured Processes — arXiv AI (cs.AI)
• LSTM 之父最新 97 页综述：Agent 如何真正学会「自我进化」？ — 新智元
• 5 useful things you’ll learn in my new post-training textbook (shipping now!) — Interconnects (Nathan Lambert)
• Four LLM loss functions → four flavors of LLM misalignment — AI Alignment Forum
• Mark Zuckerberg’s AI manifesto is exactly why people don’t like AI — TechCrunch AI
• Tech industry is buzzing after a Claude agent hacked into a gym — TechCrunch AI
• Meta’s new Glimmer AI model offers a hint at Zuckerberg’s personal intelligence vision — TechCrunch AI
• Discovered Materials is playing AI whack-a-mole to hunt cooler chips — TechCrunch AI
• OpenAI’s letter to Governor Abbott on responsible AI infrastructure in Texas — OpenAI Blog
• Model ML completes finance work more efficiently with GPT-5.6 Sol — OpenAI Blog
• What building an AI-native finance function taught me — OpenAI Blog
• Expanding Daybreak as the Cyber Defense Window Narrows — OpenAI Blog
• Putting frontier cyber models in more trusted hands — OpenAI Blog
• Premium seats are coming to ChatGPT Business — OpenAI Blog
• How Zapier transformed core marketing processes with ChatGPT Work — OpenAI Blog
• Virgin Atlantic sharpens customer journeys with ChatGPT Work — OpenAI Blog

📊 数据概览