ai-agent-book 精选快照（<2MB 代码与文档，来自 github.com/bojieli/ai-agent-book）

slides/lesson-40.md

@@ -0,0 +1,242 @@
+---
+theme: seriph
+title: "Lesson 40 — Who Should Coordinate Independent Agents?"
+info: "English video course for AI Agents in Depth"
+author: Bojie Li
+transition: slide-left
+mdc: true
+lineNumbers: false
+monaco: false
+aspectRatio: 16/9
+canvasWidth: 980
+layout: cover
+class: cover
+---
+
+<div class="course-kicker">Expand · Chapter 10 · Multi-Agent Collaboration</div>
+
+# Who Should Coordinate Independent Agents?
+
+<p class="course-subtitle">Peer review, managers, decentralized handoffs, files, and control planes</p>
+
+<div class="course-cover-meta">Lesson 40 of 42 · 17 minutes · Non-Shared Context; File System; Communication and Control; Collaboration Topologies; A2A</div>
+
+<!-- Presenter cue: Open with the concrete problem. Add personal context in your own words; do not read the slide. -->
+
+---
+layout: center
+class: text-center
+---
+
+<div class="course-kicker">The central question</div>
+
+<div class="course-big">When Agents work in separate contexts, what structure keeps their work coherent?</div>
+
+<!-- Presenter cue: Let the question sit for a moment, then state the failure mode the lesson will explain. -->
+
+---
+
+# Why this problem matters
+
+<div class="grid grid-cols-3 gap-5 mt-6">
+<div class="course-card blue">
+<h3>Peer loop</h3>
+<p>A proposer and reviewer iterate with independent evidence.</p>
+</div>
+<div class="course-card green">
+<h3>Manager</h3>
+<p>One planner decomposes, budgets, schedules, and integrates.</p>
+</div>
+<div class="course-card orange">
+<h3>Decentralized</h3>
+<p>Peers transfer work without a runtime central controller.</p>
+</div>
+</div>
+
+<!-- Presenter cue: Connect each card to a product or experiment consequence. -->
+
+---
+
+# Three ideas to keep in view
+
+<div class="grid grid-cols-3 gap-5 mt-6">
+<div class="course-card purple">
+<h3>Data plane</h3>
+<p>Workspaces, shared artifacts, external and system mounts</p>
+</div>
+<div class="course-card blue">
+<h3>Control plane</h3>
+<p>Spawn, message, status, terminate, and resource scheduling</p>
+</div>
+<div class="course-card green">
+<h3>Handoff package</h3>
+<p>Goal + evidence + artifacts + open questions + acceptance</p>
+</div>
+</div>
+
+<!-- Presenter cue: Define unfamiliar terms in plain language; the audience is new to ML training and RL. -->
+
+---
+
+# The book's visual model
+
+<img class="course-figure" src="/images/fig10-4.svg" alt="Manager architecture for sequential multi-Agent coordination">
+
+<div class="course-caption">Manager architecture for sequential multi-Agent coordination</div>
+
+<!-- Presenter cue: Trace the diagram in one direction and name the mechanism that matters for this lesson. -->
+
+---
+
+# Manager topology vs. Decentralized topology
+
+<div class="grid grid-cols-2 gap-6 mt-5">
+<div class="course-card orange"><h3>Manager topology</h3><ul><li>Global plan</li><li>Simple supervision</li><li>Planner is bottleneck</li></ul></div>
+<div class="course-card green"><h3>Decentralized topology</h3><ul><li>Local ownership</li><li>Flexible handoffs</li><li>Harder global consistency</li></ul></div>
+</div>
+
+<div class="course-caption course-caption-strong">Topology determines where planning errors and coordination costs accumulate.</div>
+
+<!-- Presenter cue: Explain the trade-off; avoid presenting the right column as universally superior. -->
+
+---
+
+# Pass artifacts; observe lifecycle
+
+~~~python
+job = manager.spawn(role, goal, budget)
+manager.message(job, evidence_paths)
+while job.running: manager.observe(job.status)
+result = manager.verify(job.artifacts)
+manager.cancel_dependents_if_satisfied(result)
+~~~
+
+<!-- Presenter cue: Walk through the executable idea line by line; keep implementation details for the terminal. -->
+
+---
+
+# Test the claim
+
+<div class="grid grid-cols-1 gap-4 mt-5">
+<div class="course-card blue">
+<div class="course-demo-head"><span>10-3</span><span>2 min</span></div>
+<h3>Rehearse a four-role translation orchestration</h3>
+<p><strong>Observe:</strong> Manager plan, glossary ownership, chapter budgets, proofreading, and integration</p>
+</div>
+</div>
+
+<div class="course-caption course-caption-strong">Demo budget: 2 minutes · one contiguous terminal block</div>
+
+<!-- Presenter cue: State the prediction before running anything. Name the observation that could disconfirm it. -->
+
+---
+class: course-terminal
+---
+
+<div class="course-kicker">Live demo</div>
+
+# Switching to the terminal
+
+~~~bash
+$ cd chapter10/book-translation && python demo.py --dry-run
+~~~
+
+<div class="course-terminal-watch">Run the command(s), narrate decisions, and point to the observation—not just the output.</div>
+
+<!-- Presenter cue: Switch windows now. Return to the next slide after every listed experiment is complete. -->
+
+---
+
+# What the evidence supports
+
+<div class="grid grid-cols-3 gap-5 mt-6">
+<div class="course-card green">
+<h3>Finding 1</h3>
+<p>Files handle large persistent artifacts; messages handle asynchronous coordination.</p>
+</div>
+<div class="course-card blue">
+<h3>Finding 2</h3>
+<p>The manager's decomposition quality caps the value of stronger workers.</p>
+</div>
+<div class="course-card purple">
+<h3>Finding 3</h3>
+<p>Explicit status and termination semantics matter as much as task prompts.</p>
+</div>
+</div>
+
+<!-- Presenter cue: Tie each finding to something viewers just observed; distinguish evidence from interpretation. -->
+
+---
+layout: center
+---
+
+<div class="course-kicker course-kicker-red">Where the claim stops</div>
+
+# Boundary condition
+
+<div class="course-boundary">A dry-run proves the orchestration graph and budgets—not translation quality or multi-Agent advantage.</div>
+
+<!-- Presenter cue: Say explicitly what this experiment does not establish. -->
+
+---
+layout: center
+---
+
+<div class="course-kicker">Engineering takeaway</div>
+
+# Design rule
+
+<div class="course-rule">Make ownership, artifact contracts, lifecycle states, budgets, and termination paths explicit before adding Agents.</div>
+
+<!-- Presenter cue: Present this as a reusable decision rule, then give one counterexample or trade-off. -->
+
+---
+
+# Continue the experiment
+
+<div class="grid grid-cols-3 gap-4 mt-6">
+<a class="course-link" href="../chapter10/autonomous-phone-registration/">
+<span class="course-link-title">Experiment 10-3: fixed baseline + autonomous phone registration</span>
+<span class="course-link-path">chapter10/autonomous-phone-registration/ (TalkAct baseline)</span>
+</a>
+<a class="course-link" href="../book-en/images/fig10-2.svg">
+<span class="course-link-title">Agent virtual file system</span>
+<span class="course-link-path">book-en/images/fig10-2.svg</span>
+</a>
+<a class="course-link" href="../book-en/images/fig10-6.svg">
+<span class="course-link-title">Manager parallel coordination</span>
+<span class="course-link-path">book-en/images/fig10-6.svg</span>
+</a>
+<a class="course-link" href="../book-en/chapter10.md">
+<span class="course-link-title">A2A and decentralized handoffs</span>
+<span class="course-link-path">book-en/chapter10.md</span>
+</a>
+</div>
+
+<!-- Presenter cue: Point viewers to the companion paths; do not walk through every extension. -->
+
+---
+layout: center
+class: text-center
+---
+
+<div class="course-kicker">Pause and apply</div>
+
+# Your turn
+
+<div class="course-big course-reflection">If the manager decomposes the task incorrectly, which independent gate can detect the mistake before workers waste their budgets?</div>
+
+<!-- Presenter cue: Invite viewers to pause the video. Offer your own answer after a short beat. -->
+
+---
+layout: center
+class: text-center
+---
+
+<div class="course-kicker">Next · Lesson 41</div>
+
+<div class="course-next">Test whether multiple Agents create new information or merely spend more tokens.</div>
+
+<div class="course-next-arrow">→</div>
+
+<!-- Presenter cue: Use this transition to make the course feel continuous rather than episodic. -->