1
Declarative Schemas for Simpler Database Management
Supabase 发布声明式模式，用 .sql 文件定义数据库最终状态并版本控制，简化复杂库表管理
Simplify managing and maintaining complex database schemas
2
Show HN: Layrr – Point Click and Edit any site
点击网页元素并描述需求，Layrr自动调用编码代理修改UI，告别繁琐描述。
I made Layrr because I got tired of describing UI changes to coding agents. Layrr opens your web app in the browser. You click the part and describe t…
3
国轩高科拟量产硫化锂及固态电解质，目标推动固态电池成本降至 1 元 / Wh
聚焦固态电池核心材料量产计划，目标成本降至1元/Wh，关键突破点在于硫化锂成本控制。
IT之家 5 月 17 日消息，据界面新闻今日报道，在 5 月 17 日举办的国轩高科科技大会上，国轩高科宣布，为解决固态电池价格高难题，计划实现固态电池关键材料的量产。 据国轩高科高级总监潘瑞军介绍， 硫化物固态电池成本的 70%-80% 来自硫化物电解质 ，而硫化物电解质成本的 70%-80% …
4
绿源推出 S86 GT-Max / Ultra 电动摩托车，5399~6199 元
绿源新款电摩极速80km/h，配ABS+TCS+智能功能，5399元起
IT之家 5 月 17 日消息，绿源本周推出了全新的 S86 GT-Max 及 S86 GT Ultra 电动摩托车，京东显示为 5399 元和 6199 元，5 月 31 日 20 点开售。 京东 绿源 S86GT Max 高性能电动摩托车 5399 元 直达链接 2026 年数码家电政府补贴持续…
5
Show HN: Agnt – Free open-source CLI to run any public or MIT-licensed AI agent
免费开源命令行工具，一键运行各类公开AI agent，开发者的实用利器
https://github.com/meajsinghk/agnt Comments URL: https://news.ycombinator.com/item?id=48170395 Points: 2 # Comments: 0
6
An AI Hate Wave Is Here
民意调查揭示AI反感情绪正全面蔓延，技术与社会的摩擦加剧。
Article URL: https://www.axios.com/2026/05/17/ai-backlash-polling-sentiment Comments URL: https://news.ycombinator.com/item?id=48173318 Points: 72 # C…
7
蔚来总裁秦力洪：大部分超豪华车就不该卖 100 万以上，在欧洲卖 5 万欧元为什么到中国就卖 100 多万？
蔚来总裁直言：大部分超豪华车在中国卖100万以上不合理，欧洲仅5万欧元，高价源于进口税和品牌溢价，国产崛起正打破这一怪圈。
5 月 16 日晚间，乐道 L80 上市发布会后，蔚来联合创始人、总裁秦力洪，蔚来高级副总裁、乐道汽车总裁沈斐与媒体对话。 秦力洪表示，把今天在中国卖百万以上的车回到它的原产国，看它卖多少钱，中间多交的还不仅仅是智商税，是真正的税，一辆百万以上的进口车有超过 50% 都是各种税。 他指出，今天有两个…
8
openJiuwen社区开源新招：重磅发布JiuwenSwarm，拉开群体智能“养蜂”序幕
华为支持的开源AI Agent平台openJiuwen发布蜂群智能体JiuwenSwarm，从单Agent走向群体协同，开启“养蜂”序幕。
底层范式变了
9
Solvita: Enhancing Large Language Models for Competitive Programming via Agentic Evolution
论文提出一种基于智能体进化的方法，显著提升大语言模型在竞赛编程中的推理表现。
arXiv:2605.15301v1 Announce Type: new Abstract: Large language models (LLMs) still struggle with the rigorous reasoning demands of hard competitive pr…
10
Look Before You Leap: Autonomous Exploration for LLM Agents
针对LLM Agent在陌生环境中“过早利用”导致失败的问题，提出自主探索能力量化评估框架，助力构建更自适应的智能体。
arXiv:2605.16143v1 Announce Type: new Abstract: Large language model based agents often fail in unfamiliar environments due to premature exploitation:…
11
Is One Score Enough? Rethinking the Evaluation of Sequentially Evolving LLM Memory
现有LLM记忆评估靠最终准确率，但会掩盖关键失败模式，本文提出新视角
arXiv:2605.15384v1 Announce Type: cross Abstract: Memory plays a central role in enabling large language models (LLMs) to operate over sequential task…
12
Domain-Independent Game Abstraction using Word Embedding Techniques
用词嵌入技术实现领域无关的博弈抽象，大幅提升复杂游戏策略的可迁移性
arXiv:2605.15543v1 Announce Type: cross Abstract: Many games of interest in the real world are often intractably large, thereby necessitating the use …
13
A Few GPUs, A Whole Lotta Scale: Faithful LLM Training Emulation with PrismLLM
用少量GPU忠实模拟千卡级LLM训练环境，降低开发调试成本与复杂度。
arXiv:2605.15617v1 Announce Type: cross Abstract: Large language model (LLM) training today runs on clusters spanning thousands of GPUs. While this sc…
14
Interaction-Aware Influence Functions for Group Attribution
打破传统影响力函数局限，提出考虑样本交互作用的群体归因方法，提升机器学习可解释性。
arXiv:2605.15675v1 Announce Type: cross Abstract: Influence functions approximate how removing a training example changes a quantity of interest, call…
15
GenShield: Unified Detection and Artifact Correction for AI-Generated Images
统一检测与伪影校正框架，直面扩散模型生成图像的逼真度挑战
arXiv:2605.16122v1 Announce Type: cross Abstract: Diffusion-based image synthesis has made AI-generated images (AIGI) increasingly photorealistic, rai…
16
Bounded-Rationality, Hedging, and Generalization
从有界理性视角解析学习器的样本依赖与泛化权衡，揭示训练损失与对冲成本的内在关系。
arXiv:2605.15340v1 Announce Type: new Abstract: A learner does not only fit data; it also determines how strongly the training sample may shape its ou…
17
$\phi$-Balancing for Mixture-of-Experts Training
新框架直接优化群体级专家平衡，解决MoE训练中负载均衡偏差问题。
arXiv:2605.15403v1 Announce Type: new Abstract: Mixture-of-Experts (MoE) models rely on balanced expert utilization to fully realize their scalability…
18
A Multi-Layer Cloud-IDS Pipeline with LLM and Adaptive Q-Learning Calibration
云安全新思路：将LLM与自适应Q学习结合，构建多层云端入侵检测流水线，应对未知攻击。
arXiv:2605.15889v1 Announce Type: cross Abstract: Security in cloud computing has become a major concern due to several factors such as layered cloud …
19
Density Estimation via Binless Multidimensional Integration
新方法BMTI通过无箱多维积分实现非参数密度估计，数据高效且鲁棒。
arXiv:2407.08094v3 Announce Type: replace-cross Abstract: We introduce the Binless Multidimensional Thermodynamic Integration (BMTI) method for nonpar…
20
A Split-Client Approach to Second-Order Optimization
提出拆分客户端新思路，突破二阶优化中海森矩阵计算与分解的同步瓶颈，提升收敛效率。
arXiv:2510.15714v3 Announce Type: replace-cross Abstract: Second-order optimization methods offer superior convergence rates but are often bottlenecke…