2026-06-09 · AI 资讯日报
今日新收录 71 条公开资讯，按模型 / 产品 / 行业 / 论文 / 观点 自动归类汇编（非 AI 生成，点击可溯源原文）。
本日报由系统根据当天新收录的公开资讯自动汇编、按方向归类，非 AI 生成；点击任意条目可溯源原文。
今日精选6 条
- 1.
Draw a store, generate LLM personas, and watch them shop — an isometric 3D sandbox for synthetic-consumer experiments.
- 2.
AZMX AI — The sovereign agent platform.
- 3.
Hi HN, we’re open-sourcing ktx. It’s an executable context layer that makes agents reliable on your data stack. We built it after going through the experience of building productio…
- 4.
Repository-level coding benchmarks such as SWE-bench have driven a rapid surge in the capabilities of coding agents. Yet they usually treat coding tasks as a holistic, binary prediction problem (e.g.,…
- 5.论文研究ResearchClawBench: A Benchmark for End-to-End Autonomous Scientific Research— HuggingFace Papers
AI coding agents are increasingly used for scientific work, but their end-to-end autonomous research capability remains difficult to verify. We present ResearchClawBench, a benchmark for evaluating au…
- 6.
Hi HN! We’re Jimmy and Ray. Jimmy is a Thiel Fellow with a Ph. D. from MIT who has worked on programming tools for 15 years; Ray became VP of Sales at a $2B company when he was 19…
模型发布/更新1 条
- 1.
AI 导读 · 开源生态信任危机，黑客直击代码供应链，微软Azure成攻击跳板。
产品发布/更新6 条
- 2.intellicia-public/parastore— GitHub
Draw a store, generate LLM personas, and watch them shop — an isometric 3D sandbox for synthetic-consumer experiments.
- 3.AzmxAI/azmx— GitHub
AZMX AI — The sovereign agent platform.
- 4.SumanD18/sentinel— GitHub
Open-source observability and trust layer for AI agents: trace every step, score every output, catch hallucinations and runaway loops in real time. Self-hostabl…
- 5.SoumilBhandari/nanoDLM— GitHub
The simplest masked-diffusion language model you can actually train, debug, and learn from — ~1100 lines of plain PyTorch, char-level, with an honest head-to-he…
- 6.GoodQ02/goodq4all— GitHub
Local-first multimodal epistemic memory for scene-level video, audio, and text intelligence.
- 7.
AI 导读 · 国产4B模型实现端侧部署，验证了小参数大模型的可行性，开辟了AI轻量化新路径。
行业动态6 条
- 8.
Hi HN, we’re open-sourcing ktx. It’s an executable context layer that makes agents reliable on your data stack. We built it after going through the experience of building productio…
- 9.
AI 导读 · 聚焦AI编程工具质量，MIT博士与销售奇才联手，技术实力与商业经验兼备。
- 10.Demo跑通了，然后呢？带你摸透AI创业的4个底层逻辑— InfoQ
AI 导读 · AI创业从Demo到产品化，必须直面技术、市场、资本的四大生死局。
- 11.
AI 导读 · 用工程闭环打通AI编码到验收，展现企业级落地的真实路径与关键挑战。
- 12.
AI 导读 · AI编程突破仍依赖人类高薪兜底，揭示技术落地与人工成本的现实博弈。
- 13.
AI 导读 · AI基础设施漏洞曝光，可能威胁大量AI应用安全。
论文研究6 条
- 14.SWE-Explore: Benchmarking How Coding Agents Explore Repositories— HuggingFace Papers
Repository-level coding benchmarks such as SWE-bench have driven a rapid surge in the capabilities of coding agents. Yet they usually treat coding tasks as a holistic, binary prediction problem (e.g.,…
- 15.ResearchClawBench: A Benchmark for End-to-End Autonomous Scientific Research— HuggingFace Papers
AI coding agents are increasingly used for scientific work, but their end-to-end autonomous research capability remains difficult to verify. We present ResearchClawBench, a benchmark for evaluating au…
- 16.On the Geometry of On-Policy Distillation— HuggingFace Papers
On-policy distillation (OPD) is increasingly used to improve large language model reasoning, but its training dynamics remain poorly understood. We characterize the trajectory of OPD updates in parame…
- 17.Latent Spatial Memory for Video World Models— HuggingFace Papers
Video world models that maintain 3D spatial consistency across generated frames typically rely on explicit point cloud memory constructed in RGB space. This design is both computationally expensive, r…
- 18.Human Psychometric Questionnaires Mischaracterize LLM Behavior— HuggingFace Papers
We examine whether human psychometric questionnaires can serve as reliable tools for characterizing and predicting LLM behavior in everyday user interactions. We analyze eight open-source LLMs by comp…
- 19.Echo-Memory: A Controlled Study of Memory in Action World Models— HuggingFace Papers
We present Echo-Memory, a controlled study of memory mechanisms in action-conditioned world models. These models generate multi-segment videos from a first frame, text prompt, and camera-action sequen…
技巧与观点1 条
- 20.Siri AI at WWDC 2026— Simon Willison
AI 导读 · 苹果Siri AI若在WWDC 2026兑现承诺，将扭转此前画饼争议，值得关注其实际落地。
快讯
- ·Demo跑通了，然后呢？带你摸透AI创业的4个底层逻辑— InfoQ · 刚刚
- ·蚂蚁数科Harness工程实践：从 AI Coding 到可验收的研发闭环｜AICon上海— InfoQ · 刚刚
- ·
- ·BadHost 漏洞使 AI 代理、评估器和 LLM 网关面临风险— InfoQ · 刚刚
- ·仅4B大小可端侧部署！卡帕西预言的「认知模型」被国产做出来了— 量子位 · 1 小时前
- ·被砍的魅族 22 Next“AI 小方块”工程机外观照片曝光：紫光展锐 T8200 芯片、4 英寸机身— IT之家 · 1 小时前