1
博云新材：未与蓝箭航天签署任何框架协议，也并非蓝箭航天“朱雀三号”可回收火箭热防护系统的供应商
36氪获悉，博云新材公告，2025年度公司营业收入9.10亿元，其中商业航天产品营业收入约200多万元，在主营业务收入中占比低（不足1%）。2026年以来公司商业航天产品尚未形成销售收入，也没有“快舟系列”配套碳/碳喉衬产品的销售收入。相关传闻严重失实。公司未与蓝箭航天签署任何框架协议，也并非蓝箭航…
2
行云科技：签订3.22亿元企业级SSD硬盘销售框架协议
36氪获悉，行云科技公告，公司控股子公司行云存算与浙江甚湖签署《设备销售框架协议》，行云存算拟向浙江甚湖出售企业级SSD硬盘，含税总金额3.22亿元。若合同顺利实施，预计对公司未来的经营业绩产生一定影响。本订单销售的产品为公司自有品牌企业级PCIe SSD产品，主要面向AI服务器相关业务，未来可配套…
3
HarnessAPI: A Skill-First Framework for Unified Streaming APIs and MCP Tools
LLM工具部署不再分裂：一个框架同时搞定HTTP端点和MCP注册，舍弃重复造轮子，拥抱统一流式API。
arXiv:2605.22733v1 Announce Type: new Abstract: Every Python function deployed as an LLM tool must today exist in two forms: an HTTP endpoint for huma…
4
VRPRM: Process Reward Modeling via Visual Reasoning
通过视觉推理提升过程奖励建模精度，为复杂任务训练提供新思路。
arXiv:2508.03556v3 Announce Type: replace Abstract: Process Reward Model (PRM) is widely used in the post-training of Large Language Model (LLM) becau…
5
Heterogeneous Agent Collaborative Reinforcement Learning
突破异质智能体协同难题，提出HACRL框架，让类型各异的智能体高效协作学习。
arXiv:2603.02604v2 Announce Type: replace Abstract: We introduce Heterogeneous Agent Collaborative Reinforcement Learning (HACRL), a new Reinforcement…
6
外钛内铝：三星被曝正研发新型“双相”框架以提升折叠屏手机耐用与散热性能，应对苹果液态金属创新
IT之家 5 月 23 日消息，爆料人 Schrödinger 在 Telegram 上发布消息称，自去年以来，三星一直在尝试融合钛与铝两种金属的优异性能，尝试打造出一种兼具坚固性与散热性的新型“双相”机身框架。 该框架将用于三星未来的折叠屏手机产品线，其外部为坚固耐剐蹭的钛合金，内部则采用航空级铝…
7
Unified Data Selection for LLM Reasoning
提出统一数据选择框架，为LLM推理任务高效筛选高质量训练数据，显著提升推理能力。
arXiv:2605.22389v1 Announce Type: new Abstract: Effectively training Large Language Models (LLMs) for complex, long-CoT reasoning is often bottlenecke…
8
Agentic CLEAR: Automating Multi-Level Evaluation of LLM Agents
提出Agentic CLEAR框架，自动化评估LLM Agents多层次能力，提升评估效率与客观性。
arXiv:2605.22608v1 Announce Type: new Abstract: Agentic systems are becoming more capable: agents define strategies, take actions, and interact with d…
9
Two is better than one: A Collapse-free Multi-Reward RLIF Training Framework
破解多奖励强化学习中的模型崩溃难题，提出RLIF训练新框架实现稳定收敛。
arXiv:2605.22620v1 Announce Type: cross Abstract: Reinforcement learning with verifiable rewards (RLVR) has substantially improved the reasoning abili…
10
Show HN: SoMatic – Vision-based OS automation framework for AI agents
基于视觉的OS自动化框架，让AI代理精准操控电脑，突破多模态LLM定位瓶颈。
Hi HN, I'm Smyan and I enjoy building agents. Modern multimodal LLMs are great at vision and perception but are quite poor at localization. This natur…
11
Show HN: LoongForge-A high-performance training framework for LLM, VLM, VLA, Wan
百度开源高性能训练框架LoongForge，最高5倍提速，支持LLM/VLM/VLA，已服务企业生产。
Article URL: https://github.com/baidu-baige/LoongForge Comments URL: https://news.ycombinator.com/item?id=48221214 Points: 8 # Comments: 2
12
Text Analytics Evaluation Framework: A Case Study on LLMs and Social Media
提出针对LLMs处理社交媒体长文本的评估新框架，填补实践分析场景缺口
arXiv:2605.21338v1 Announce Type: new Abstract: LLMs have demonstrated exceptional proficiency in a wide range of NLP tasks. However, a notable gap re…
13
Beyond the Bellman Recursion: A Pontryagin-Guided Framework for Non-Exponential Discounting
用庞特里亚金最优控制原理突破传统强化学习折扣框架，为认知与经济学中的非指数折扣提供新解法。
arXiv:2605.20996v1 Announce Type: new Abstract: Most value-based and actor--critic reinforcement learning methods rely on Bellman-style recursions, ye…
14
TextSculptor: Training and Benchmarking Scene Text Editing
场景文本编辑新框架TextSculptor，训练与基准测试双突破，AI文字处理再升级。
arXiv:2605.21090v1 Announce Type: new Abstract: Recent advances in Multimodal Large Language Models (MLLMs) and diffusion-based generative models have…
15
VERA-MH: Validation of Ethical and Responsible AI in Mental Health
VERA-MH 提供标准化框架验证心理健康 AI 的伦理与责任，帮助开发者确保算法合规安全
arXiv:2605.13318v2 Announce Type: replace Abstract: Chatbot usage has increased, including in fields for which they were never developed for--notably …
16
Tailwind CSS v2.2
Tailwind CSS v2.2发布，高性能CLI、全面伪类与变体支持，堪称最丰富版本
An all-new high-performance CLI tool, ::before and ::after support, sibling selectors, selected text variants, and tons more.
17
BrainDyn: A Sheaf Neural ODE for Generative Brain Dynamics
结合层代数与神经ODE，实现高保真脑动力学生成，为脑科学计算建模开辟新路径。
arXiv:2605.19324v1 Announce Type: new Abstract: Efficient neural network models that generate brain-like dynamic activity can be a valuable resource f…
18
The Plan for React 18
React核心团队官宣React 18发布计划、工作组及Alpha版本，新特性主打渐进式采用，开发者不容错过的版本更新解读
The React team is excited to share a few updates. We’ve started work on the React 18 release, which will be our next major version. We’ve created a Wo…
19
React Conf 2024 Recap
React Conf 2024 官宣 React 19 RC、新架构 Beta 和编译器，回看大会所有重磅技术更新。
Last week we hosted React Conf 2024, a two-day conference in Henderson, Nevada where 700+ attendees gathered in-person to discuss the latest in UI eng…
20
Vanilla JavaScript support for Tailwind Plus
Tailwind Plus 所有UI组件现已支持纯 JavaScript，无需 React/Vue 框架即可使用，交互性与可访问性满分。
Every UI block in Tailwind Plus is now fully functional, accessible, and interactive, even if you don't use a framework like React or Vue.