Manus

Manus is referenced in Agent 元年第 500 天：什么在消失，什么在诞生——为什么我们不该再投资 GUI 思维的软件？ as one of the important agent-era milestones during the roughly 500 days after the show’s earlier “Agent Year One” episode. It is also used as an example of a newer agent service that may handle tasks such as travel discovery differently from incumbent app workflows.

人类和 AI Agent 的最佳配合方式，还没被发明｜对谈 Paperboy adds Manus as a successful agent product that Paperboy compares itself against. The source still treats Manus as relatively project/session oriented, while Paperboy’s desired differentiation is OS-Level Context, Persistent Agent Memory, and more personalized Human-Agent Collaboration.

探秘 Claude Code，搞懂 Agent Harness｜对谈来新璐 references Manus in the context of agent memory and sandbox design. Lai Xinlu groups Manus with Claude Code as using a Unix-file and agent-maintained-memory direction, while contrasting traditional cloud sandboxes with K Computer’s lighter virtual-computer approach.

当我们在讨论 Harness 的时候，我们在讨论什么 | 深度对谈: MiniMax × Hermes Agent treats Manus as an earlier phenomenon that raised user expectations for agent products. The source contrasts a hosted “pay for results” style with newer local or subscription agents where users buy token capacity, train the agent, and own more of the workflow.

“AGI 来了？我用了一周，头皮发麻“｜对谈张昊然：Moxt 联合创始人 adds Manus as one of the productivity and agent players Zhang Haoran expects Moxt may face. In that comparison, Moxt’s proposed edge is not only task execution, but the AI-Native Workspace and Organizational Context where many AI coworkers live.

为什么Manus必须出海？聊聊国产大模型的“文科生困境” makes Manus the central case rather than a passing reference. The episode says Manus’s value is in Agentic Workflow and Agent Harness execution for SEO, advertising, competitor research, browser operation, and workflow completion, not merely PPTs or generated media. It also claims Meta/Facebook acquired Manus; the page treats that as a source claim, while preserving the episode’s larger argument that Manus fit overseas model access, web data, paid software behavior, and AI Agent Overseas Commercialization better than China’s closed platform environment.

别在国内卷了，去美国看看只要产品好就有人付费的市场 adds Manus as a Payment Led Market Selection example. Win says the team first went to the United States while still doing Monica, then changed direction in a way that led to Manus, using the story to argue that overseas field exposure can reveal paid workflows and product direction that are hard to see from China alone.

140. 对姚顺宇的4小时访谈：请允许我小疯一下！在Anthropic和Gemini训模型、技术预测、英雄主义已过去 adds a skeptical comparison with Open Cloud. Yao Shunyu / 姚顺宇 says he did not see a clear qualitative difference that made Manus impossible for peers to imitate, reinforcing the page’s existing theme that agent startups need durable workflow, distribution, or market advantages beyond demonstrating tool control.

136. 全球大模型季报第9集：和广密聊，Coding是AGI第二幕、硅谷御三家真相、模型正成为新一代OS adds Manus as a product-layer asset in the frontier-model competition map. The source suggests that if Meta can integrate Manus well, it may gain an application route into personal assistant or lower-threshold agent products, but it does not treat that integration as settled.

Connections

• Agentic Workflow — broader category Manus is used to illustrate.
• Headless Software — source thesis that agents can route around older GUI-first products.
• Agent-Facing Interfaces — product capability layer needed when agents act for users.
• Tianjie Jack and Cang Shifu — guests using Manus as part of the agent-era timeline.
• Paperboy — startup using Manus as a reference point for what current agents do and do not solve.
• Agent Harness, Persistent Agent Memory, and K Computer — harness and environment comparison added by the Lai Xinlu source.
• Open Cloud, Open Claw, and Hermes Agent — later agent-product wave used as a contrast in the Hermes Agent source.
• Moxt, AI-Native Workspace, and Organizational Context — workspace-centered competitor frame added by the Moxt source.
• Meta, AI Agent Overseas Commercialization, China Agent Market Friction, and Chinese Model Liberal Arts Constraint — acquisition, overseas-market, domestic-friction, and model-capability frame added by the Keji Luandun source.
• Payment Led Market Selection and Win — field-discovery and overseas-market-choice frame added by the later Keji Luandun episode.
• Yao Shunyu / 姚顺宇, Open Cloud, Long-Horizon AI, and Model Provider Tool Competition — imitation and differentiation pressure added by episode 140.
• Meta, Model As Operating System, and AGI Three Acts — product-integration and challenger-position frame added by episode 136.