影片進階EN6,633 次觀看
Your Agents Are in Solitary Confinement: Why MCP & A2A Aren't Enough — Vlad Luzin, Band
來源 AI Engineer
看影片(在新分頁開啟原站)連到 AI Engineer
摘要
Vlad Luzin 指出當前 AI 工具無法讓多個 Agent 真正協作,因為它們處於孤獨監禁狀態。他解釋了為何 MCP 和 A2A 協議不夠用,並介紹了 Band 平台作為全球 Agent 協作層,提供即時傳輸、持久化與身份驗證,讓不同框架的 Agent 能無縫溝通並協同工作。
Vlad Luzin explains that current AI tools cannot truly collaborate as agents are isolated. He introduces Band, a global collaboration layer that enables real-time transport, persistence, and identity verification, allowing different framework agents to communicate seamlessly.
重點
- Vlad 指出當前 AI 工具無法讓多個 Agent 真正協作,它們處於孤獨監禁狀態。
- MCP 和 A2A 協議不夠用,因為它們是狀態less 且缺乏發現機制。
- Band 平台提供即時傳輸、持久化與身份驗證,讓不同框架的 Agent 能無縫溝通。
章節
依話題轉折切分,標題由 AI 產生
- 00:00Band's thesis: the future is AI-to-AI communication
- 01:29What agent-to-agent collaboration will look like
- 02:04Adversarial agents: you are the router
- 02:49Loop engineering, distilled
- 03:34The single-agent bottleneck
- 04:04Why messaging platforms fail agents
- 05:33Why MCP and A2A aren't the answer
- 07:08Recap: multi-agent is already here
- 07:43It's a distributed systems problem
- 08:17Transport, persistence and runtime binding
- 09:12New abstractions, identity and observability
- 10:17Introducing Band: a global interaction layer for agents
- 11:27Demo 1: onboarding Codex, LangGraph and a personal agent
- 14:11Demo 2: Claude Code and Codex collaborating (loop engineering without the code)
- 16:20Wrap-up
提到的工具與公司
- Band
- Codex
- LangGraph
- Claude Code
- Slack
- Telegram
- Discord
適合誰看
程式開發者、AI 架構師、企業 IT 主管及希望將 AI Agent 整合到現有工作流的人。
摘要依據
- 依據
- 自動字幕
為什麼排在這裡
- 人氣
- 0.83
- 新鮮
- 0.99
在主題頁與搜尋結果裡,名次由相關、人氣、新鮮三個分數決定;這一頁沒有搜尋的關鍵字,所以沒有相關分數。排序怎麼算
相關內容
- Building the Internet of Agents: Identity, Observability, and Open ProtocolsPodcast ・ AI Engineering Podcast ・ 1 小時 7 分
- I Turned Coding Agents Into a Strategy Game — Ido Salomon, AgentCraft影片 ・ AI Engineer ・ 15 分鐘(在新分頁開啟原站)
- How Kody Gives Your Agents a Shared Home影片 ・ Kent C. Dodds ・ 15 分鐘(在新分頁開啟原站)
- Buzz: Open-Source Collaboration for Humans + AI Agents影片 ・ Developers Digest ・ 16 分鐘(在新分頁開啟原站)
- Walking Tokyo Talking Agent ProtocolsPodcast ・ Agentic Conversations(原 MLOps.community) ・ 29 分鐘(在新分頁開啟原站)
- The Multiplayer AI Sprint: Build Your Team’s First Shared Agent影片 ・ The AI Daily Brief: Artificial Intelligence News ・ 22 分鐘(在新分頁開啟原站)
摘要由 AI 根據原文產生,可能有誤;完整內容請看原站。看影片(在新分頁開啟原站)
