Shared Intelligence - Making Intelligence as Sharable as Knowledge

A New Approach to Information Curation

What is Intelligence?

We define intelligence not as what you know, but as how you come to know it: your process for gathering, filtering, and synthesizing information to achieve understanding. An optimal knowledge workflow for a specific task is what makes someone intelligent in that domain.

The Challenge

Information overload is one of the defining challenges of our time. While we have unprecedented access to knowledge, most of us consume information through recommendation algorithms designed to maximize engagement, not understanding.

Worse, the intelligence processes we develop—our personal strategies for finding and filtering information—remain locked in our heads, unable to be automated, shared, or improved upon collectively.

What if we could define our own knowledge workflows, automate them, and collaborate with others the same way we do with code or research projects?

Our Approach

Shared Intelligence is a research initiative exploring this question through three core capabilities:

Externalize

Express complex knowledge workflows explicitly—how you gather, filter, and synthesize information.

Automate

Execute these workflows automatically, saving hours of manual work.

Collaborate

Share your workflows with others and collectively improve them—just like open-source code.

By making intelligence (the process) as sharable as knowledge (the product), we enable collective improvement of how humans interact with information. Instead of everyone independently fighting information overload, communities can share, adapt, and refine the workflows that work.

The Research Opportunity

Understanding current limitations reveals opportunities for innovation. Our research addresses three interconnected challenges.

Centralized Control

Sophisticated recommendation algorithms currently live within large platforms—institutions with the infrastructure to build and maintain complex curation systems at scale. While technically impressive, these systems serve millions with generalized approaches.

The Gap: Individual users lack personal recommendation engines they can control, customize, and refine for their specific goals.

Research Question: How can we return agency to individuals in defining their information diet?

Isolated Intelligence

Everyone independently develops their own information curation strategies—what sources to follow, how to filter noise, when to deep-dive. These personal intelligence processes are:

• Isolated: Developed independently, not shared
• Manual: Time-consuming and repetitive
• Implicit: Locked in people's heads
• Unoptimized: No feedback for improvement

The Gap: We lack mechanisms to measure, optimize, and collectively improve these intelligence processes.

Research Question: Can we design systems where optimization aligns with user learning and growth?

Infrastructure Gap

We have mature infrastructure for knowledge artifacts (storage, search, sharing), but no equivalent for intelligence processes:

• Defining: How to express multi-step workflows
• Executing: How to automate information gathering
• Versioning: How to iterate curation strategies
• Sharing: How to transfer workflows between people
• Discovering: How to find and adapt others' approaches

The Gap: Without tools to externalize workflows, expertise remains trapped and communities cannot develop best practices.

Research Question: What infrastructure enables knowledge workflows to be shared and collaboratively refined?

Research Timeline

A three-phase research agenda toward collective intelligence

In Development

Phase 1: SIQL

Open-source query language for expressing and executing knowledge workflows

✓ Core parser and executor
✓ Academic sources
🤝 Seeking contributors for new data sources
🤝 Seeking contributors for filtering & synthesis mechanisms

View on GitHub →

Planned

Phase 2: No-Code Interface

Visual workflow builder making SIQL accessible to non-programmers

• Drag-and-drop workflow builder
• Template library
• Scheduled execution
• Interactive results exploration

Join Waitlist →

Planned

Phase 3: Collaboration Platform

GitHub-like platform for sharing and forking knowledge workflows

• Public workflow repository
• Fork & collaborate
• Community ratings
• Marketplace for sources & agents

Coming Soon

Let's see how it works...

SIQL: A Query Language for Knowledge Workflows

Think of it as "SQL for knowledge"—instead of querying databases, you query any information source with AI-powered filtering and synthesis. Create workflows that run on schedules, monitor for changes, or execute on-demand with custom inputs.

Automated Intelligence Workflows

Build multi-step workflows that run automatically—gathering, filtering, and synthesizing information from multiple sources on your schedule.

ai_safety_monitor.siql

// Define extraction and formatting criteria
$key_debates = "Ongoing discussions, controversial viewpoints, emerging consensus"
$format_description = "Markdown with sections: Executive Summary, Key Papers, Community Insights, Industry News"

WORKFLOW "AI Safety Research Monitor" {
  SOURCES {
    academic: [arxiv, semantic_scholar],
    forums: [lesswrong, alignment_forum],
    industry: [openai.blog, anthropic.blog]
  }
  
  PIPELINE {
    PARALLEL {
      $papers = SEARCH ["AI safety", "alignment"] FROM academic
               | FILTER date > 7_days_ago,
      $discussions = FETCH FROM forums 
                    | FILTER date > 7_days_ago
                    | EXTRACT $key_debates,
      $announcements = FETCH FROM industry
                      | FILTER date > 7_days_ago
    }
    SYNTHESIZE "Weekly digest: Research breakthroughs, 
       community debates, industry developments" WITH [$papers, $discussions, $announcements]
    | FORMAT $format_description
  }
  
  OUTPUTS {
    schedule: WEEKLY(day: "Sunday", time: "8:00am"),
    format: email
  }
}

Result: Complex research monitoring that would take hours of manual work runs automatically every week, synthesizing insights from academic papers, community discussions, and industry announcements.

Real-Time Intelligence Monitoring

Monitor any source for changes and trigger intelligent analysis only when something important happens—no noise, just signal.

sec_filing_monitor.siql

// Monitors SEC filings for key competitors
$competitors = ["Tesla", "Rivian", "Lucid"]

// Define extraction criteria as natural language instructions
$financial_highlights = "Key financial metrics, revenue trends, profit margins"
$risk_factors = "Material risks, regulatory concerns, competitive threats"
$strategy_changes = "Shifts in business strategy, new initiatives, market positioning"
$significant = "material impact on competitive position, major strategic pivot, or significant financial change"

MONITOR sec.edgar FOR $competitors {
  WATCH {
    filing_types: ["10-K", "10-Q", "8-K"],
    check_interval: 15_minutes
  }
  
  PIPELINE {
    EXTRACT [$financial_highlights, $risk_factors, $strategy_changes]
    | ANALYZE "Compare to previous quarter: 
       Material changes? Competitive implications? 
       Strategic shifts?"
    | FORMAT {
        type: "markdown",
        sections: ["executive_summary", "key_changes", "implications"]
      }
    | SEND TO slack WHEN $significant
  }
}

Result: Instead of checking competitor filings manually, get instant intelligent alerts with contextual analysis—only when something material changes.

Reusable Research Templates

Create parameterized workflows once, then reuse them for different topics. Share templates with your team or the community.

tech_landscape_analysis.siql

// Usage: RUN WITH {technology: "edge AI chips"}
$format_description = "Markdown report: Executive Summary, Technical Analysis, Patent Landscape, Market Players"

PIPELINE tech_landscape_analysis {
  INPUT technology
  $sources = [arxiv, patents, crunchbase, news]
  
  // Multi-source intelligence gathering
  PARALLEL {
    $research = SEARCH "{technology}" FROM [$sources.arxiv]
               | FILTER date > 90_days_ago
               | CLASSIFY BY "commercial_readiness",
    $patents = FETCH FROM $sources.patents 
              WHERE technology_area = "{technology}"
              | CLUSTER BY assignee,
    $startups = FETCH FROM $sources.crunchbase
               WHERE description CONTAINS "{technology}"
               | RANK BY funding_total
  }
  
  SYNTHESIZE "Investment thesis: Technical maturity, 
    patent landscape, key players, market timing" WITH [$research, $patents, $startups]
  | FORMAT $format_description
}

Result: A reusable research template that generates comprehensive technology landscape analysis in minutes. Run it for "quantum computing", "synthetic biology", or any emerging field—same workflow, different insights.

From simple queries to complex multi-source workflows—SIQL grows with your needs. Start simple, scale to sophistication.

Join Us

Help us create a world where individuals have absolute control over their information diet and can collaborate to grow collective intelligence.

Why This Matters

We envision a future where everyone understands how they gather, filter, and synthesize information—and can share these knowledge workflows with others. Instead of isolated, manual research processes, we're building infrastructure for collective intelligence.

The time is critical. We need to move fast and achieve wide adoption for this vision to become reality. Every contributor brings us closer to a world where intelligence is as shareable as knowledge.

What We're Building: Source Adapters

Phase 1 of SIQL is about making the language widely usable across different sources and use cases. To do that, we need source adapters—modular connectors that let SIQL gather information from any source.

The Opportunity

By building an adapter, you enable yourself and everyone else to use SIQL with new sources—expanding the language's power and reach.

The Impact

Each adapter unlocks new use cases: academic research, market intelligence, content curation, competitive monitoring, and more.

Skills We're Looking For:

• Agentic Systems — Experience with LLM-powered agents and structured outputs
• Web Scraping — Extracting data from websites, handling dynamic content
• API Integration — Working with REST APIs, handling auth, pagination, rate limits

Want to contribute to the future of collective intelligence?

Reach Out