AI 情报文章归档

Step 5 Preview 现已登上 @OpenRouter 的 New & Trending 排行榜。 感谢每一位试用的人，也感谢我们的合作伙伴将它带入你们的工作流。

微软论文提出 TeleTune，让智能体直接从原始使用日志中学习软件技能，无需实时测试环境，仅保留能更好预测用户下一步动作的技能编辑。该方法猜测每个会话的目标，用文本技能库让模型预测每条日志动作，错误预测触发技能库修改，只有在留出日志上准确率提升的编辑才会被保留。论文称旧日志上的下一步动作准确率与实时成功率一致，建议产品记录用户活动并据此挖掘智能体技能。

Simon Willison 用 ChatGPT 桌面应用 Codex 标签页的语音模式，在做饭的约半小时内基本靠口头指令为博客开发出 Newsletters 索引功能，使用模型为 GPT-6 Astra High。

I first noticed something missing while reviewing saved discussions of Microsonya, my Telegram...

GOOGLE 🔥：Google AI Studio Build 上被发现正在开发一个新的 Ultra 模式。 > 其描述写道：“Build with advanced skills and tools。” > 这个新模式与 Plan、Build 以及此前发现的 Security review 模式一同出现。 目前没有迹象表明 Ultra 模式需要 Ultra 订阅，也没有信息说明它将使用哪些工具和技能。不过，这些新模式很可能正在开发中，以配合 Gemini 4 Argon…

This is a submission for the Hacktoberfest Open-Source AI Challenge: Week 1 — Touch Grass. ...

An open-weight tabular model forecasts your last spring frost and a local Gemma writes the planting advice. Fully offline, no account, $0 to run.

Runtime failure monitors for robot policies are typically evaluated by how accurately they distinguish failed from successful executions. Yet an accurate alarm may arrive too late for recovery, while an unnecessary intervention may disrupt…

Clinical prognosis requires integrating structured measurements, physiological waveforms, longitudinal history, and population context. We introduce PRISM (Population-Retrieved Inference for Structured Multimodal Prognosis), a multimodal l…

Federated learning (FL) seeks to optimize a shared global model from decentralized client data. While effective for average performance, this process can produce unfair performance across clients, particularly under statistically heterogen…

Generative Recommenders increasingly use Reinforcement Learning (RL) to optimize outcomes that may appear hundreds or thousands of interactions after a decision. Standard RL propagates reward according to temporal proximity rather than exp…

Update, Oct 9: I wrote this on Sept 29. The first backtest finished on Oct 1. One signal first looked...

Existing LLM-based agent approaches for long-horizon machine learning engineering (MLE) often use context engineering, implemented through custom harnesses that manage tools, skills, and accumulated experience. However, we find that as sin…

This is a submission for the Kaggle Benchmarking Challenge What I Benchmarked A published...

Sam Altman 在 Bloomberg 上谈到了 OpenAI 那个 700 亿美元的 ARR 数字。 （完整视频在 'Bloomberg Tech' YouTube 频道，链接见评论）

Agentic orchestration delegates tasks to sub-agents with independent contexts, providing a key path to extending the effective execution horizon of large language model agents. However, existing orchestration methods face two limitations: …

据 Financial Times 报道，汇丰计划裁减其英国约 70% 的理财顾问，将财富客户转向 AI 支持的数字服务。熟悉该计划的人士称，该部门一半的管理和专业岗位也将被裁减，银行目前正就涉及管理者、专业人员和理财顾问的岗位削减与员工进行协商。

On-policy distillation (OPD) provides teacher supervision on student-generated trajectories, but a single trajectory realizes only one continuation at each decision, limiting supervision coverage under a finite sampling budget. We propose …

In September 2026, IQVIA called trusted data, responsible AI, and enterprise governance the new...

Nonlinear model predictive control (NMPC) can be applied to large nonlinear systems with complex objectives while enforcing constraints, but the cost of solving the underlying optimization problems limits real-time application for systems …

Generating edge case scenarios is crucial for training, testing and safety assuring autonomous systems. This identifies hazardous failures to increase system reliability and robustness, ensures the system can still function safely on the b…

We derive approximation rates for Sobolev functions defined on the complex unit sphere and complex projective space, representing quantum states, and on the special and projective unitary groups, representing quantum unitary operations. By…

Representational similarity measures (RSA, CKA, Procrustes) are chosen for the transformations they treat as irrelevant, and all of them take the set of conditions as given. That set is fixed by two decisions, which inputs are admitted and…

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass ...

If you run AI agents on cron jobs the way I do, you already know where the bill comes from. It is not...

补充： > 能向非 AI 原生工程师传达技术概念。 > 推动保持简单，对功能/行为说不

Multimodal LLMs can fail to read text embedded in an image's coarse structure (text a human recovers by squinting) and report that no text is present. We name this class Squint Text and study it as a recognition deficiency, motivated by ca…

A 2026 empirical study found that the reusable "skills" we plug into AI agents leak credentials at scale, usually during ordinary use, with no exploit…

AI systems are making rapid progress in mathematics, but this progress is hard to measure. Benchmarks of solved problems saturate, and in benchmarks of open problems each problem can be resolved for the first time only once. We present Squ…

Cyber-Physical Systems are developed through multiple model based activities, comprising several stages of development. As the development progresses through simulation, experimentation, and operation, observations of system execution may …

Large language models (LLMs) have taken the world by storm. Their ability to communicate with us in our own language, and to reason about our world, has made them a favorite tool for writing, answering questions, and much more. Their answe…

Meta 新论文发现，智能体记忆系统对用户价值观与信念类声明应单独提高置信阈值，而非全局统一。这类声明仅占候选记忆的 21.5%，却贡献了超过其他所有类别总和的缺乏支撑的声明，其来源支持率仅 77.9%，其他事实为 96.2%。仅对价值观类收紧标准后，无支撑记忆从 6.2% 降至 4.0%，同时多保留约 13 个百分点的来源关键事实。

On September 4 I started a ComfyUI custom node: a canvas with layers and selection tools, so I could...

Speech recognition on low-cost microcontrollers is, in practice, limited to keyword spotting or a fixed list of commands. We present Oído, an inference engine that runs open-vocabulary English speech recognition entirely on an ESP32-S3 (24…

There is a bug in these six lines. Find it before you scroll. `def get_discount(price, user): ...

The Decoder 报道，OpenAI 解雇了 Tomek Korbak、Jasmine Wang 和 Mikita Balesni 三名与 Hugging Face 事件调查相关的安全研究员，三人在公开信中表示此举令在职员工恐惧，并提出落实外部审计员驻场、保持前沿模型可监控性、明确与外部安全组织合作规则三项要求。

Compression schemes provide a general framework for deriving data-dependent risk bounds for learning algorithms. For learning rules based on preferent compression, existing results upper-bound the risk in terms of a sample-dependent comple…

Have you ever wanted to upload your college notes and simply ask an AI, “Explain Unit 3” or “What is...

Recent accent classification systems can achieve high accuracy in standard in-corpus evaluations. However, when speakers overlap between the training and test sets, such results may overestimate model robustness compared with real-world de…

Collective Autonomous Vehicle (AV) route choices in real-world traffic networks can create externalities for Human-Driven Vehicles (HDVs), as routes that benefit AVs may increase travel times for other road users. To address this, we intro…