2026-10-21 当日要闻
当天 211 条
- ·OpenAI
Using thousands of AI agents to uncover new leads in biological data
- ·RadarAI
预训练之后为什么还要后训练:三种技术如何把基座模型变成助手
- ·RadarAI
WSJ 披露 Anthropic IPO 草案:上半年收入 163 亿美元、纸面亏损 807 亿美元、约 4500 亿美元算力承诺
- ·RadarAI
Databricks 员工调侃 DeepSeek 开源 DeepInfer 并公布推理收入数据
- ·Snyk Blog
AI Model Risk Depends on What the Model Can Reach
- ·arXiv AI
When Planning Problems Become Hard for LLMs: Separating Structural Complexity from Semantic Interpretability
- ·arXiv AI
LightCRS: Lightweight Multi-Domain Conversational Recommendation System
- ·arXiv AI
MAMath-Bench: Process-Aware and Robust Evaluation of Multimodal Multi-Agent Mathematical Reasoning
- ·arXiv AI
Fix, Not Just Resample: Bounded Compiler-Guided Repair in Lean 4
- ·arXiv AI
Industrial Agentic AI: LLM-Driven New-Generation Agents for Smart Manufacturing
- ·arXiv AI
DataPraxis: Scaling Analytics Tasks via Verified Workspace Evolution
- ·arXiv AI
MedPAO-Assist: Edge-Deployable, Modality-Aware Agentic AI for Radiology Report Structuring
- ·arXiv AI
Pareto-Optimal RTL Code Generation via Multi-Objective Reinforcement Learning with Large Language Models
- ·arXiv AI
Large language model and reinforcement learning based autonomous construction and progress updates
- ·arXiv AI
Navalia: Advancing Cowork Agents through Verifiable Task Synthesis and Long-Horizon Post-Training
- ·Google AI Blog
Southeast Asia’s digital economy to hit $368 billion
- ·RadarAI
消息称 NVIDIA 下代 Groq LPU 芯片转向 3D 堆叠 DRAM,放弃大 SRAM 方案
- ·RadarAI
Claude Code v2.1.306 发布,修复云会话登录与 MCP 工具报错问题
- ·RadarAI
Adobe 重新设计 Photoshop 滤镜机制:滤镜变为可移动、可调节的效果层
- ·RadarAI
Metaphor 云编码智能体发布
- ·RadarAI
DeepSeek 开源推理框架 DeepInfer 并发布技术报告,公开大规模部署细节
- ·RadarAI
小米 MiMo-V2.5-Pro 模型今日下线,替换为 V2.6-Pro 后支持多模态理解
- ·RadarAI
工程师分享用 Agent 状态图配合 pstack 形式化验证技能找 bug 的实践
- ·RadarAI
邵猛吐槽 X 平台 AI 回复泛滥,追问其目的
- ·量子位
飞书豆包工作AI先锋未来人才大赛:企业正在争抢会用AI做成事的人
- ·RadarAI
Kun Chen 四篇长文解析 AI 时代软件业的需求、价值与职业走向
- ·RadarAI
Dario Amodei 回应 Nathan Calvin 对 Anthropic 安全立场的批评
- ·RadarAI
Databricks 发布 MemCurrent:用在线读写加离线做梦给生产 Agent 装上长期记忆
- ·RadarAI
Uber 如何让 AI 安全地删代码:AutoClean 清理 1.6 万个过期 Flag 与百万行死代码的 Agent 工程实践
- ·RadarAI
如何用决策模型做 LLM Evals:从阈值调优到概率校准的实践解读