Termsslug: agent

Agent

智能体

AI systems that autonomously plan and invoke tools for multi-step tasks, now the dominant source of model API traffic.

Curated Intelligence

406 items · Page 1/41
腾讯技术工程 (微信公众号)T2·Agents
58 pts
Capability Recombination Lets Judgment Combine With Built Capabilities

The article introduces "capability recombination": once a team distills specialized capabilities into permission-shareable agents and tools, individuals can bring their own judgment directly into more kinds of work without mastering every role's details. Using data-service delivery as an example, requesters ask in a group or agent, a dispatch agent assigns the task, specialists and data agents align on definitions, and results return to the original session, cutting manual handoffs. The shift is not that individuals got stronger, but that combining judgment with existing capabilities became cheaper.

LangChain BlogT1·Agents
62 pts
LangChain Rebuilds LangSmith Around Agents as Core Object

LangChain has rebuilt LangSmith around the agent as the core object, launching the LangSmith Agent Platform to unify building, running, improving, and governing agents in one place. The company cites a 2026 step change in agent adoption, with enterprises running coding, customer-facing, and internal productivity agents that need platform-level governance. Developers can start via LangSmith Studio or connect agents they already run.

Hacker News AIT2·Industry
58 pts
AI Reshapes Data Teams and Spawns the Data Intelligence Engineer Role

MotherDuck co-founder Nick Ursa reports that AI has cleared his data engineering backlog for the first time, shifting his job from computing correct numbers to verifying numbers produced by LLMs and agents. The author proposes the emerging role of Data Intelligence Engineer, centered on quality control and decision support rather than ticket queues.

IT之家 · 智能时代T2·Industry
78 pts
Stripe Closes $7.5B Acquisition of OpenRouter

Stripe completed its acquisition of AI model aggregation platform OpenRouter on November 10, with the deal valued at roughly $7.5-8 billion. OpenRouter connects over 400 models from 80+ providers and serves more than 10 million developers, processing about 790 trillion tokens in October 2026, 76% of which came from agent-related API calls. The two will focus on token routing and usage optimization, folding model selection and billing into payment infrastructure.

AWS Machine Learning BlogT1·Agents
38 pts
AWS Quick Encodes Repeatable Tasks with Skills and Agents

An AWS blog post explains how agents, skills, and connected data sources fit together in AWS Quick: agents interpret natural-language requests and orchestrate connectors, while skills encode repeatable tasks such as weekly status updates, release notes, and ticket creation into reusable, team-shareable instruction units. It also covers combining the official skill catalog with custom agents, but remains a vendor-bound conceptual and how-to guide without new models or quantitative results.

宝玉 (@dotey)T2·Agents
58 pts
Baoyu: Agents Must Submit Screenshot Evidence for Human Acceptance

To address the opacity of AI agent task results, Baoyu proposes requiring agents to self-test and generate screenshots as evidence of completion for human reviewers. The practice borrows from his team's rule that engineers attach feature screenshots to pull requests, and it lets agents catch and fix issues during the screenshot process.

机器之心 (微信公众号)T2·Industry
38 pts
State of AI 2026 Report Maps Frontier AI Landscape and Risks

Air Street Capital has released its ninth annual State of AI Report, surveying the past year across research, industry, politics and safety. It concludes that the global frontier has consolidated into a three-way race among Anthropic, OpenAI and Google, while flagging benchmark collapse, agent risks spilling into production systems, and AI turning into a geopolitical lever.

TechCrunch AIT2·Agents
38 pts
Meta Launches Spaces for Collaborating With Its Muse Agent

Meta has added Spaces to its Muse agent interface, letting users create separate sidebar conversations around specific projects such as vacation planning, gift tracking or wedding prep, with Muse working toward a goal inside each space. Meta frames these as "custom experiences," but they are essentially topic-scoped breakout conversations, similar in positioning to rival cartoonish agent assistants.

Simon Willison 博客T2·Industry
58 pts
Nilay Patel Criticizes AI Industry: Products Not Good Enough to Justify Costs

The Verge editor Nilay Patel argues on The Vergecast that AI products do not yet deliver enough value to justify their infrastructure and social costs, that agents remain largely unreal, and that industry spending far exceeds revenue. He contends the AI industry cannot escape its credibility and business problems until products visibly improve users' lives without needing persuasion.

IT之家 · 智能时代T2·Industry
58 pts
China's Broadcasting Regulator Targets 95% AI Penetration by 2030

China's National Radio and Television Administration set a target of at least 95% AI penetration across the broadcasting and online audio-visual sector by 2030, alongside more than 150 million new-type terminal applications, as part of its 15th Five-Year Plan. The goal covers AI adoption in content, networks, services, security and regulation, including industry-specific large models and agents, not the share of AI-generated content viewers see.