2026 年 10 月 17 日，DeepSeek 一条官方新闻让「许愿式编程」刷屏中文技术圈，但这个术语的身世远比一次发布复杂：它脱胎于 1985 年 SICP 的教学黑话，在 2025 年被 Andrej Karpathy 的 vibe coding 重新点燃，又花了大半年从贬义词变成官方卖点。本文按时间线拆解术语谱系，辨析它与 vibe coding、agentic engineering、规格驱动开发的分野，并用 METR、DORA、Stack Overflow 的数据与 Replit 删库等事故，回答「许愿」何时是真生产力、何时是灾难。

Repository-level software engineering (SWE) comprises heterogeneous task categories, whose progress under pooled agentic…

Google 比 OpenAI 早七周发布 Deep Research，但把「研究计划让用户改批」产品化的也是它——agentic 检索的通用循环是：规划、迭代检索、阅读、反思补漏、交叉验证、带引用报告。本文拆解这个循环的两种实现路线（端到端 RL vs 显式编排），用 BrowseComp 上「裸模型不足 10% vs deep research 51.5%」的差距说明多步浏览行为本身值多少分，也把「慢不等于对」的引用可靠性研究摆上台面。

从"控制流所有权"这一第一性原理出发，拆解 Workflow（DAG 编排、确定性执行）与 Agent（ReAct 循环、涌现式控制流）的技术原理差异；详解 Agentic Workflow"图做骨架、节点内自主"的三层混合架构，以及提示链/路由/并行化/编排者-执行者/评审-优化五种经典编排模式与工程选型经验。

Black-box optimization (BBO) arises in many scientific and engineering problems where objective evaluations are expensiv…

Building on nearly a decade of co-engineering, CoreWeave has built NVIDIA compute, networking and software into a cloud …

AI security is an engineering problem. That means defined security requirements, enforceable controls, named owners and …

At a glance Harnessed Agentic RL: Microsoft Research Asia introduces a training paradigm in which the same agent harness…

We propose BrickBench, a benchmark for agentic text-conditioned LEGO-set design. Given a prompt, an agent is tasked with…

Agentic reinforcement learning (RL) disaggregates training from rollout, so each policy update must reach the rollout cl…

At OpenAI’s DevDay event on Tuesday, the company announced the launch of Dots, a new personal agentic assistant powered …

Google Research has introduced an **AI video co-director** for long-form video generation. The suite of 4 agentic framew…

OpenAI announced on Wednesday that it is bringing voice-based agentic features to mobile, allowing users to trigger work…

Compositional 3D scene reconstruction has recently been explored from two directions: agentic reasoning that provides se…

Agentic memory is becoming essential for long-horizon AI agents, yet many existing systems rely on autoregressive LLMs t…

Professional graphic design is a long-horizon agentic task in which structured, editable artifacts emerge from many inte…

从控制流与上下文的所有权出发，给出四条可执行的 Agent 工程原则：一切外部接入以工具体系形式接入且不注入系统提示词；逐级披露贯穿技能、工具发现、工具执行与记忆四个环节；Workflow / Agent / Agentic Workflow / Graph 各有场景、不是替代关系；并逐层拆解四者的技术原理——DAG 与状态机、ReAct 循环、宏观图加微观循环的混合架构，以及 State/Node/Edge、超步执行、reducer 合并语义、checkpointer 恢复、interrupt 人审与递归上限。

SWE-Bench Pro has emerged as a standard benchmark for evaluating software engineering agents on challenging repository-l…

Visual generation is evolving from generative models used through a single invocation into agentic control processes tha…

StepFun has released Step 5 Preview, a sparse Mixture-of-Experts model with 600B total parameters and 27B active per tok…

As AI agents become participants in the development of their successors, they reshape both the production of intelligenc…

三种概念的系统辨析：预定义代码路径 vs 运行时策略驱动，附选型决策框架与混合架构实践

Companies in France are deploying agentic computing across industries spanning space, financial services, the public sec…

Reconstructing complete, scene-aligned 3D objects from casual images requires integrating sparse, uncertain observations…

Language model agents increasingly operate over long interaction streams, requiring online adaptation to distribution sh…

!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…

Agent 工程系统学习第 13 章（前沿专题）：三个高薪高门槛方向。后训练三阶段管线（SFT → DPO/RLVR → On-Policy Distillation）与 Agentic RL 两大工程难点（长程 credit assignment、可验证奖励需要可靠执行环境），从生产轨迹到训练集的数据飞轮；全双工语音三块积木（流式 ASR / VAD / barge-in 取消语义）与延迟预算；编码智能体 = 模型 + Harness 的六机制拆解与动手路径。

LLM agents generate intermediate reasoning and actions token by token, making extended interactions slow and computation…

Foundation models can serve as clinical agents through tool-use harnesses. However, conventional medical benchmarks asse…

Coding agents can now write, run, and debug programs with little human help. Robot tasks, however, are usually specified…

Last week, TypeSafe AI released Jev, its first **System One model**. Founder Diogo Almeida previously worked at OpenAI o…

Reliable AI safeguards require both control mechanisms that reduce unsafe behavior and monitoring mechanisms that detect…

Large language models (LLMs) excel at structured, verifiable tasks, but their low-entropy bias can produce homogeneous a…

Manuel Uth Sep 28, 2026 Harvard psychologist Steven Pinker thinks fears of AI wiping out humanity are overblown.** He ma…

Reinforcement learning (RL) for code agents often uses executable tests to provide binary rewards. With these rewards, G…

Language-model agents increasingly face long-horizon tasks with evolving state, interdependent decisions, and delayed ou…

To build and deploy sophisticated robotics applications that can perceive, reason and act in dynamic environments, devel…

SpaceXAI has released Grok 4.7, its new flagship model for coding, agentic tasks, and knowledge work. Grok 4.7 is built …

Alibaba's Qwen3.8-Omni-Flash understands audio and video, plans tasks, calls tools, and reports about 45.7% fewer tokens…

Training capable coding agents via reinforcement learning (RL) requires diverse tasks with reliable verifiers. Open-sour…

Multi-turn agents trained with reinforcement learning (RL) receive a single scalar reward per trajectory, which motivate…

把吴恩达 2026 年 8–9 月连发五封信构建的《AI 工程技能地图》从"看懂"变成"照做"：给出 20 格可打分自评表（四大顶层能力 × 全部细分项，每格配一句过关判定问题），逐格拆解"学什么—怎么练—什么算过关"，并附三条不同起点的学习路径、一张 12 周计划表、三个必做练手项目与七个反模式清单。全部细项定义来自吴恩达原文，判定问题与计划表为本文延伸并已标注。

Multimodal agents can create complex videos in software such as Blender by coding without relying on diffusion models. Y…

Video diffusion models are stochastic and hard to control: precise content often requires repeated sampling without guar…

In this work, we present ZGCM-1, a fully open 7B dense foundation model trained from scratch with extreme data, system, …

Recursive self-improvement (RSI) requires a concrete mechanism through which an AI system observes its capabilities and …

Diverse and simulation-ready indoor scenes are essential for interactive entertainment and embodied AI, yet their scalab…

Designing and authoring high-performance custom kernels for accelerators is a complex task that requires deep hardware-l…

“We’re at an inflection point in cybersecurity,” Jensen Huang told a sold-out crowd at CrowdStrike’s Fal.Con 2026 in Las…

Advertising video generation is not only a video synthesis task, but also a product-conditioned reasoning problem whose …