AI 情报文章归档

Mixture-of-Experts (MoE) enables multimodal large language models (MLLMs) to scale capacity efficiently, but expert parameters dominate memory even though only a few experts are activated per token: they account for over 90% of model size …

Diffusion drafters accelerate speculative decoding by denoising multiple draft tokens in parallel. Their training uses anchor sampling to construct compact draft queries and performs cross-attention with a block-sparse mask over a much lon…

导演 Ash Koosha 以能否撑起院线长片为标准实测 Kling 4.0，预计 Fountain Film Engine 中废弃素材至少减少 50%（早期估计）。Kling 4.0 支持单次生成最长 30 秒，并提升物理交互、空间关系与镜头稳定性，减少偏离开场动作的漂移。他还看重多镜头生成与定向编辑：可只改表情、动作或画面元素而保留其余部分。

Reasoning models allocate test-time compute across distinct behaviors such as planning, deriving, and verifying. However, models frequently "overthink," continuing to verify and re-derive long after finding a correct answer. This wastes ex…

Merging companies that each run Salesforce must consolidate their customer data into a single organization while both source systems remain in use. This requires reconciling different schemas and picklists, resolving duplicate customers ac…

We study the minimum-disagreement cluster editing problem that, from a complete edge-signed graph, where a positive edge represents two vertices agreeing to each other while a negative represents a disagreement, constructs disjoint cliques…

Planning electricity systems requires repeated decisions that connect regional resources, investment, operation and network constraints. Here we present an autonomous planning framework that carries prepared regional scenarios through prop…

Saliency maps, offered as evidence that mammography classifiers use the cancer, are untested in dense breasts, which can hide cancers. We tested whether they locate cancers worse there, and why. On 1,345 malignant CBIS-DDSM views, we score…

Agentic architecture is repeating service-oriented architecture's two worst mistakes, the smart pipe and the synchronous chain.

Patient world models are increasingly proposed as simulators of clinical trials, and recent surveys rank their capabilities on a ladder from temporal prediction to counterfactual rollouts. Position on that ladder is currently assigned by r…

As sequence lengths continue to grow, distributed attention based on sequence parallel (e.g., Ring-Attention) has become a mainstream solution for large language models (LLMs). Nevertheless, inherent communication dependencies limit comput…

KV cache eviction is not merely a token-ranking problem: the workspace retained for generating eviction queries also determines the quality of those queries. Consequently, an early workspace error can distort later query observations and r…

Developers insert logging statements in source code to record its execution, and the resulting logs are essential for detecting and troubleshooting system anomalies. A key part of these statements is the log level (e.g., DEBUG, INFO, WARN,…

Fifteen days, twenty-five releases. PageSift went public as 0.20 on October 20; today it ships...

What 586 Real Traces Taught Us About LLM Agent Determinism How temperature selectively...

Forecasting emotion in conversation is increasingly framed as predicting the next emotional state's value. We show this targets a hard quantity in a state-observed testbed and propose a learnable, deployable reformulation instead. On valen…

Comparisons in conversational emotion-shift forecasting are easily distorted when a learned model receives rich features but its transition baseline is given a noisy or unavailable current-emotion label, when hyperparameter budgets differ,…

Recycling a target model's own predictions offers a training-free source of proposals for speculative decoding. In vision-language models, however, existing recycling indexes predictions by the last token alone, conflating predictions comp…

结合马斯克回复与AI1卫星公开设计，太空散热方案采用大型铝制展开结构配氨两相循环，铝制散热器朝向深空可达约1300 W/m²，AI1设计用约1400 W/m²。

The fastest way I know to break a working LLM feature is to "just tweak the prompt" on a Friday....

Agent skills are just markdown files that an LLM will obey. Here's what malicious ones look like, how they were caught with a static analyzer, and why the MCP config file is the most dangerous line that never reaches the model.

An agent can successfully query a database and still give you the wrong answer. Ask, “How many...

Unattended LLM pipelines fail in ways a person would normally fix by editing the prompt. If the...

Notes from building ContaCal, a photo-based calorie logger for Brazilian meals. The pipeline, the hard parts (regional dishes, portions, lunchboxes) and the limits we tell users about.

Apliquei as fases do Spec Kit em uma tarefa de um projeto de previsão do Brasileirão e conferi o resultado contra a própria spec.

A team I talked to last month had a 4x jump in inference spend over one quarter. Not a crisis — they...

Luma AI 介绍 Character Sheet 功能：该功能可将用户提供的参考图整理成一张一致的角色设定表，供后续持续使用，避免在生成下一个视角时角色形象发生变化。

A Postman desktop update added its plugin to Claude Code and Codex for an existing user. Its installer is built to approve its own Codex hook. How to check.

If you've debugged a web app with an AI assistant, you know the loop: The AI asks what the console...

作者分享了用 Grok Bot 搭配 vercel 开源的 agent-desktop 实现一个 Agent 控制本机其他 Agent 和应用的方法。

A personal assistant is easy to demo and hard to trust. The demo ends when the assistant answers....

人应当始终掌控未来。 OpenAI 反对这样一种观念：即认为让人类退居到只做“AI 系统认为值得在人类身上奖励的事情”的角色是可以接受的。 人类不应把未来的控制权让渡给超级智能，这一要求最终应当约束 AI 的进展。

很高兴看到 MiniCPM-OCR 集成进 Nock！💙 感谢分享成功案例和困难案例。这种真实世界的反馈非常宝贵！🙌

太喜欢这个 MiniCPM-V 的创意用法了！💙📱 把端侧视觉和 ESP32 结合起来做一个实用的日常工具，这想法太酷了。感谢分享并开源！🙌

This is part of my build-in-public series on running LLM workflows in n8n in production. Everything...

热烈祝贺你和团队！🎉💙 很高兴能与你们合作。期待MLX-VLM和MLX-Audio接下来的发展！🚀

If you let Claude Code, Codex or Cursor drive a browser, a big part of your token bill is page...

据大厂日爆 11 月 3 日爆料，字节跳动本月起将对全公司 L4 级别代码仓库和飞书文档实施外部 AI 访问封禁，员工个人采买的 Claude、Codex 等 AI 账号不能再用。

Chunk, embed, store, retrieve, rerank and answer with citations: a complete RAG walkthrough on Pinecone's new document indexes and the Claude Messages API.

A tour of the A2A CLI against Rust agents on A2A v1.0: the official Rust SDK, a plain stdin/stdout program wrapped by a2a server --exec, and an ADK coordinator on Gemini, each timed from the CLI and from a Rust client.