> ## Documentation Index
> Fetch the complete documentation index at: https://docs.actionmodel.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Training the Large Action Model

> Discover how Action Model trains the world's first community-owned Large Action Model. Learn the fundamental differences between LLMs and LAMs, and how millions of user journeys create the Action Tree that powers true AI automation.

<Frame caption="Training the Large Action Model - How Community Training Works">
  <iframe width="100%" height="400" src="https://www.youtube.com/embed/Cqm3BVnJMB0" title="Training the Large Action Model" frameBorder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowFullScreen />
</Frame>

<Warning>
  **The AI Revolution's Dirty Secret**: LLMs can write poetry, but they can't book your flight. They can explain quantum physics, but they can't fill out a form. Action Model changes everything by training AI to actually *do* things, not just talk about them.
</Warning>

## The Problem: LLMs Can't Act

### Language Models Were Never Trained to Use Platforms

<CardGroup cols={3}>
  <Card title="LLMs Training: Just Text" icon="language" color="#ef4444">
    LLMs were trained on language—books, articles, websites. They never learned how to *use* platforms, only how to describe them.
  </Card>

  <Card title="99.9% Behind GUIs" icon="window-restore" color="#f59e0b">
    The internet isn't text or APIs—it's graphical interfaces. Every button, form, and menu that humans navigate daily.
  </Card>

  <Card title="Can't Click or Navigate" icon="mouse-pointer" color="#ef4444">
    LLMs weren't trained to click, type, or navigate. The internet remains unusable to traditional AI.
  </Card>
</CardGroup>

## LLMs vs LAMs: The Fundamental Difference

<Tabs>
  <Tab title="Large Language Models">
    ### What LLMs Do

    **Generate and understand text**

    * Predict the next word in a sequence
    * Generate human-like responses
    * Process and analyze text

    **Training Data**

    * Books, articles, journals
    * Websites, blogs, social media
    * Scientific papers
    * Easy to scrape from the internet

    **Limitations**

    * Cannot interact with GUIs
    * Cannot perform actual actions
    * Often hallucinate interface interactions
    * Require APIs or integrations
  </Tab>

  <Tab title="Large Action Models">
    ### What LAMs Do

    **Perform real actions on interfaces**

    * Click buttons and links
    * Fill out forms
    * Navigate between pages
    * Complete complex workflows

    **Training Data**

    * Real user GUI interactions
    * Mouse movements and clicks
    * Keyboard inputs
    * Screenshots and DOM states
    * **Must be intentionally collected**

    **Capabilities**

    * Direct GUI manipulation
    * Platform-agnostic automation
    * Human-like task completion
    * No API requirements
  </Tab>
</Tabs>

## Why APIs Aren't the Solution

<Info>
  **The API Myth**: Less than 0.1% of web functionality is exposed via APIs. Major platforms like Instagram and Booking.com actively restrict or eliminate API access. APIs are built for developers, not users—and they never expose full functionality.
</Info>

### GUIs Are for Humans, LAMs Are for Humans

<CardGroup cols={2}>
  <Card title="API Limitations" icon="ban">
    * Restricted functionality
    * Developer-focused
    * Often blocked or rate-limited
    * Requires technical knowledge
    * Platform-specific integration
  </Card>

  <Card title="GUI Advantages" icon="desktop">
    * Full platform functionality
    * Human-friendly interaction
    * Universal approach
    * No integration needed
    * Works everywhere
  </Card>
</CardGroup>

## The Action Tree: Mapping the Interactive Internet

<Frame>
  <img src="https://mintcdn.com/actionmodel/NKX_PmMBtEBjlKwW/the-large-action-model-lam/images/actionTree.png?fit=max&auto=format&n=NKX_PmMBtEBjlKwW&q=85&s=eecdbc955457cb2384ebf0ad10c7cd0c" alt="Action Tree Pn" width="1000" height="400" data-path="the-large-action-model-lam/images/actionTree.png" />
</Frame>

### How User Journeys Become Intelligence

<Steps>
  <Step title="User Performs Task">
    A user completes a task naturally—booking a hotel, posting on social media, or managing emails—while the browser extension records.
  </Step>

  <Step title="Journey Recorded">
    Every click, type, and navigation is captured along with context: DOM elements, screenshots, and environmental state.
  </Step>

  <Step title="Path Mapped">
    The journey becomes a branch in the Action Tree, connecting with similar paths from other users.
  </Step>

  <Step title="Tree Grows">
    Millions of journeys interweave, creating a comprehensive map of how to complete any task on any platform.
  </Step>

  <Step title="LAM Navigates">
    When given a goal, the LAM traverses the Action Tree to find the optimal path, executing actions with human-like precision.
  </Step>
</Steps>

## Training Data Requirements

### The Complexity Challenge

<AccordionGroup>
  <Accordion title="LLM Training Data" icon="book">
    **Readily Available**

    * 175B+ tokens available online
    * Books, articles, websites
    * Can be scraped automatically
    * Synthetic data generation possible
    * Static text content

    **Collection Method**

    * Web crawlers
    * Database dumps
    * Public datasets
    * API access
  </Accordion>

  <Accordion title="LAM Training Data" icon="route">
    **Must Be Created**

    * Requires active user participation
    * Dynamic interaction sequences
    * Context-dependent actions
    * Platform-specific nuances
    * Temporal relationships

    **Collection Method**

    * Browser extensions
    * Desktop recording
    * User journey tracking
    * Active labeling
    * Community contribution
  </Accordion>
</AccordionGroup>

## The Training Process

### From Individual Actions to Collective Intelligence

<Info>
  **The Network Effect**: Every user journey makes the model smarter. When one person books a flight on a new airline website, millions can now automate that same task. This is the power of community training.
</Info>

### Data Collection Methodology

| Component             | What's Captured                      | Purpose                                 |
| --------------------- | ------------------------------------ | --------------------------------------- |
| **DOM Elements**      | HTML structure, element IDs, classes | Identify clickable/interactive elements |
| **Screenshots**       | Visual state at each step            | Understand visual context and layout    |
| **Mouse Coordinates** | Exact click positions                | Precise action replay                   |
| **Keyboard Input**    | Text entered, keys pressed           | Form filling and navigation             |
| **URL Navigation**    | Page transitions and routes          | Understand site structure               |
| **Network Requests**  | API calls and responses              | Capture dynamic content                 |
| **Timing Data**       | Delays and load times                | Realistic action pacing                 |
| **Error States**      | Failed attempts and recovery         | Robust error handling                   |

## Community Training at Scale

### The Resistance Builds Together

<CardGroup cols={3}>
  <Card title="Global Trainers" icon="users">
    Active community members training the LAM across websites daily.
  </Card>

  <Card title="Millions of Actions" icon="mouse">
    Individual actions recorded, labeled, and integrated into the Action Tree.
  </Card>

  <Card title="Growing Platform Coverage" icon="globe">
    Websites and applications mapped with expanding workflow coverage.
  </Card>
</CardGroup>

## Real-World Example: Multi-Platform Workflow

<Note>
  **Complex Task**: "Find trending news on X/Twitter, create a graphic in Canva, and post to Instagram"
</Note>

### How LAMs Execute Complex Chains

<Steps>
  <Step title="Understand Intent">
    Parse the user's goal into a sequence of sub-tasks across multiple platforms.
  </Step>

  <Step title="Navigate to X/Twitter">
    Use the Action Tree to find the path: Open browser → Navigate to X → Login if needed
  </Step>

  <Step title="Find Trending Content">
    Click explore → Identify trending topics → Extract relevant content
  </Step>

  <Step title="Open Canva">
    Navigate to Canva → Select template → Insert extracted content
  </Step>

  <Step title="Create Graphic">
    Use design tools → Apply styling → Download image
  </Step>

  <Step title="Post to Instagram">
    Navigate to Instagram → Click create post → Upload image → Add caption → Publish
  </Step>
</Steps>

## Why Community Training Wins

<Tabs>
  <Tab title="Diversity">
    **Millions of Perspectives**

    * Different workflows for same goal
    * Cultural and regional variations
    * Platform-specific optimizations
    * Edge case coverage
  </Tab>

  <Tab title="Scale">
    **Exponential Growth**

    * 24/7 global training
    * Parallel data collection
    * Rapid platform coverage
    * Continuous improvement
  </Tab>

  <Tab title="Quality">
    **Human Validation**

    * Real user workflows
    * Natural interaction patterns
    * Practical task completion
    * Active labeling and verification
  </Tab>

  <Tab title="Ownership">
    **Community Owned**

    * Contributors earn tokens
    * Shared value creation
    * Democratic governance
    * Aligned incentives
  </Tab>
</Tabs>

## The Training Paradox

<Warning>
  **Big Tech's Dilemma**: Training a LAM requires massive-scale user interaction data that even Google and Microsoft struggle to collect. Why? Because they can't watch every user's screen. But we can—with permission, transparency, and rewards.
</Warning>

### Why Action Model Will Win

| Factor              | Big Tech                   | Action Model                  |
| ------------------- | -------------------------- | ----------------------------- |
| **Data Collection** | Limited to their platforms | Every website, every platform |
| **User Incentive**  | None (they take your data) | Earn tokens for contribution  |
| **Training Speed**  | Slow, corporate processes  | Rapid, community-driven       |
| **Coverage**        | Their ecosystem only       | The entire internet           |
| **Ownership**       | Shareholders               | Community members             |

## Technical Architecture

### The Action Loop

<Frame>
  <img src="https://mintcdn.com/actionmodel/NKX_PmMBtEBjlKwW/the-large-action-model-lam/images/actionLoop.png?fit=max&auto=format&n=NKX_PmMBtEBjlKwW&q=85&s=4fc25561d06a9ddf9628c0c544b78e6c" alt="Action Loop Pn" width="1000" height="400" data-path="the-large-action-model-lam/images/actionLoop.png" />
</Frame>

<Steps>
  <Step title="Observe Environment">
    Capture current screen state, DOM, and context
  </Step>

  <Step title="Search Action Tree">
    Find relevant paths based on current state and goal
  </Step>

  <Step title="Predict Next Action">
    Determine optimal next step with confidence scoring
  </Step>

  <Step title="Execute Action">
    Perform click, type, or navigation action
  </Step>

  <Step title="Verify Result">
    Check if action succeeded and goal is closer
  </Step>

  <Step title="Repeat or Complete">
    Continue loop until goal achieved or timeout
  </Step>
</Steps>

## Join the Training Revolution

<CardGroup cols={2}>
  <Card title="Install Extension" icon="download" color="#9333ea" href="/the-large-action-model-lam/browser-extension-overview">
    Start training in 60 seconds and earn tokens for your contribution
  </Card>

  <Card title="ActionFi Bounties" icon="trophy" color="#f59e0b" href="/actionfi/Earn-on-ActionFi">
    Earn up to 50x multipliers by training on high-value partner platforms
  </Card>

  <Card title="View Progress" icon="chart-line" color="#10b981" href="https://train.actionmodel.com">
    Track your training contribution and earnings in real-time
  </Card>

  <Card title="Security & Privacy" icon="shield" color="#3b82f6" href="/the-large-action-model-lam/security-and-privacy">
    Learn how your data is protected during training
  </Card>
</CardGroup>

## The Future of AI Training

<Info>
  **Projection**: By 2026, the Action Tree will contain paths for every significant task on every major platform in every language. This isn't just an AI model—it's a complete map of human digital interaction.
</Info>

### What Happens Next

* **Phase 1**: Platform Coverage (Current)
  * Mapping major platforms
  * Building core workflows
  * Community growth
* **Phase 2**: Deep Personalization
  * Individual preferences
  * Company-specific workflows
  * Cultural adaptations
* **Phase 3**: Universal Automation
  * Any task, any platform
  * Cross-platform chains
  * Natural language to completion

***

**You're not just training an AI. You're building the future of work.**

**Train it. Own it. Control it.**