AI 情报文章归档

„Penny? Anwesend?“ — Der Auftrag, der verschwand, bevor er fertig war Penny-Days, Folge 3. Entwurf,...

Google is rolling out direct checkout powered by the Universal Commerce Protocol (UCP) to shoppers in...

October 2026 was my first month of hackathons. I wanted a few real projects on GitHub, something...

Elvis Presley 音乐权利人及 Eminem 部分作品的出版方起诉 AI 音乐公司 Suno，指其非法生成模仿这些艺人的音乐和歌词。原告律师主张 Suno 违反田纳西州 ELVIS 法案，该法案禁止 AI 公司克隆歌手声音。

Most resume tools will stretch you to fit any posting. Paste a nursing job next to a warehouse resume...

On November 2, 2026, a security researcher sent an uncomfortable email to a database vendor. The message told the company that its own admin-level cloud API key

Building Beanstick: Autonomous Fiat-to-Crypto Settlement on Solana with Anchor PDAs &...

队友可以帮你撰写、出版并在 Kindle 上销售书籍

We describe the system submitted by Telefónica Innovación Digital to the Speech-COSER Challenge 2026, which features spontaneous speech from elderly speakers recorded in rural Spain. For ASR, we perform CTC-based pseudo-forced alignment sc…

I let strangers try to talk AI agents out of real Solana devnet vaults. Same attack, same model, same prompt: one drained, one locked by an on-chain spend policy.

This post was written as part of the Kiroween 2026 hackathon (#KiroweenHackathon26). A customer...

We study stochastic bandits with reward-dependent delays, possibly infinite, whose reports omit the originating round. For each $c\in[0,1)$, Balanced Historical Optimism (BHO) has a delay term controlled by the largest suboptimal-arm quant…

Predictive algorithms are increasingly used to guide hospital discharge decisions, yet these models can exhibit censoring bias for vulnerable patients due to insufficient historical data or past discharge decisions. When biased readmission…

Imagine usar a tecnologia não apenas para criar mais um aplicativo no mercado, mas para oferecer uma...

Many studies of animal vocalizations require clean recordings, but animals are usually recorded in the field, with background noise from weather, vehicles, and other animals. Animal vocalization denoising is challenged by the scarcity of c…

Existing evaluations often test persona expression rather than whether persona information persists in extended decision-making. We introduce THEATRON, a controlled benchmark that uses financial replay to evaluate persona persistence in lo…

We present MicroSim, a system for binary code similarity detection that represents each function as a graph built from IDA Pro Hex-Rays microcode and compares such graphs with a contrastive Siamese graph neural network (GNN). We evaluate o…

On November 8, 2026, Bloomberg's Mark Gurman reported that Apple CEO John Ternus held a closed-door...

In multi-institutional biomedical data analysis, comprehensive disease profiles are often fragmented across different data silos. Furthermore, local data distributions typically exhibit varying levels of granularity or even complex nested …

Swarm Learning improves resilience and removes the need for a single trusted coordinator by distributing model aggregation and validation across participating nodes. However, it also delegates the validation of model updates to peers whose…

AutoGen's archive warning and John Langford's AI Frontiers exit account expose separate questions about maintenance and the route from research to products.

Josefina Tersoli, October 2026. The code is at github.com/JTersoli/bigpickle-brain, MIT. Every...

The small open embedder EmbeddingGemma-300M, used as released, finds the right passage first for only 51.7% to 60.8% of Belebele questions in Tajik, Kazakh, Uzbek and Kyrgyz. We present Saidzoda-CA-EmbeddingGemma-300M, one model for all fo…

Generating accurate CAD models requires substantial engineering expertise and careful control of geometry, topology, and mechanical constraints. Text-to-CAD aims to simplify this process, but current LLM-based methods often generate CAD pr…

Large language models (LLMs) perform well on general coding tasks, but their capabilities in application-layer embedded software development remain underexplored. Existing embedded benchmarks primarily focus on hardware-in-the-loop validat…

Google 宣布可在 Google AI Studio 中直接构建由 Google Maps Platform 驱动的应用，用提示词描述想法即可，AI Studio 自动完成配置和 API 密钥管理，无需信用卡。

Anthropic 的 Fable 5.2 让 Kieran Klaassen 在三天内做出融合 Premiere 剪辑、After Effects 特效与 Ableton 混音的 AI 智能体视频编辑器 Everycut，一周后在公司内部演示。这是其测试过的首个让专业人士不读底层代码就能构建生产级工具的模型，能接手整个产品并持续构建、测试和打磨。

Zyphra 发布 ZONOS3-Mini，一个 Apache 2.0 开源的零样本语音克隆 TTS 模型，采用稀疏 MoE 架构，总参数 8 亿、活跃参数 1 亿，在 Pixel 9 浏览器端 Time To First Audio 低于 570 ms，在 M3 MacBook Air 上低于 150 ms。

Google Cloud 开源 JAX 训练框架 MaxText 新增长上下文训练支持，通过 All-gather、A2A、P2P 及 A2A+P2P 混合四种上下文并行策略，在 Cloud TPU 上实现超长序列训练。

Google 宣布用于训练 Gemini 等自家大模型的分布式系统 Pathways 在 Google Cloud 正式可用（GA）。Pathways 用单个客户端把整个计算描述为图，由分布式运行时完成放置、张量传输和调度，支持 RL 循环等多程序重叠执行、主机故障后重映射实现弹性训练，以及通过 Shared Pathways Service 让团队共享 TPU 集群。

Preferred Networks 以开放权重形式在 Hugging Face 发布金融特化小语言模型 PLaMo 3.0 Fin，采用 PLaMo 社区许可证，年收入或销售额低于 10 亿日元的企业及个人可免费商用。

Artificial Analysis 推出 Endpoint Accuracy Index v1.1，将 AA-LCR-25 每题重复次数从 10 提升至 25，使该评测平均 95% 置信区间从 ±4.7 收窄至 ±2.6 分。

Artificial Analysis 评测 Meta 最新高效档模型 Muse Ember，其在 Intelligence Index 得 36 分，是达到该分数中最便宜的模型，每任务 $0.14，与 Claude Haiku 5.5 (max) 同分但成本约三分之一。

字节 Seed 发布新一代图像创作模型 Seedream 5.5，带来三项升级：非编辑区域像素级保持、支持围绕同一作品多达 100 轮连续编辑；新增原生 RGBA 透明素材生成，直接输出带 Alpha 通道的图像，省去抠图；改善光照材质与皮肤纹理表现，并显著提升文字渲染能力。

Hark 官方宣布其服务接入 iMessage 和 RCS，用户可在短信中直接与 Hark 对话，内容会同步进 Hark Pro 既有线程。

百度文心团队发布自动研究系统 ERNIE Evolve，用双层循环组织研究方向选择与实验迭代，由 Director、Selector、Executor、Evaluator 和 Reviewer 分工推进研究。

Anthropic 与 Diode Computers 合作，让 Claude Fable 5.2 从简单提示词出发完成免提音箱、语音助手中枢和电子宠物三种设备的设计，包括 PCB 布局、CAD 建模、固件编写和 DFM 检查，三块电路板直接由工厂按设计文件制造，原型经数周使用仍正常工作。

Anthropic 发布 Claude Fable 5.2 和 Claude Mythos 5.2，两者为同一底层模型，Mythos 需通过 Cyber 或 Life Sciences Verification Program 申请访问。

Audio deepfake detectors trained with known synthesis attacks can learn generator-specific shortcuts that fail as synthesis systems and acoustic conditions evolve. We investigate an attack-free alternative: a normality model trained exclus…

The CoRL 2026 RoboEval challenge has eight bimanual tasks, and it ranks a policy not only on whether it succeeds but on how: every stage of the task, done safely, efficiently and with both arms in coordination, scored against human demonst…