Skip to content

TL;DR - This All-Hands builds on our research series to explore the fork in agentic marketing: extraction vs. empowerment. We’ll use this framework to prototype “Agency Layer” systems - agents that optimize for human flourishing, not just clicks. The article below is your foundation for the Gemini 3 Hackathon, where you’ll apply these ideas to build next-generation, empowerment-aligned applications.


From Intention to Empowerment

In Phenomenology of Search we studied the geometry of meaning. In Memory & Agency we gave models a sense of continuity. In Geometry of Intention we learned to recognize human goals. Now we face the critical question: what do we optimize those goals for?

Thesis: Agentic marketing is a fork in the road. The same machinery that can maximize compulsion can maximize empowerment. The objective function decides. We’re building the empowerment path.

Read the full deep-dive: The Empowerment Imperative: Rewriting Agentic Marketing from Extraction to Human Flourishing


Build-In-Public Sprint - “The Agency Layer”

Goal: Prototype an empowerment-aligned agent that demonstrates the alternative to extraction-based marketing. You’ll connect theory to practice by building systems that optimize for human flourishing.

Track A - Core Agency Layer (Infrastructure)

StepDeliverableHints
1Goal Alignment EngineExplicit goals > inferred interests
2Consent Gates ModulePersonalization off by default
3Constraint CheckerHard stops on dark patterns
4Dual Metrics DashboardPerformance + Agency metrics side-by-side
5Demo Video (3 min)Show the fork in action

Track B - Empowerment Commerce Agent (Application)

StepDeliverableHints
1Full agent app (Gemini 3 API)Shopping assistant with agency layer
2Alienation DetectionFlag extraction patterns in real-time
3Trust Accumulation SystemOptimize for long-term satisfaction
4Capability Expansion ModuleHelp users become better decision-makers
5Demo Video (3 min)Show empowerment vs. extraction paths

Recognition: Submit to the Gemini 3 Hackathon hosted by Google DeepMind. Internal showcase winners get featured in our Labs Showcase and the open-source repository.


Hackathon Details - Gemini 3 (Hosted by Google DeepMind)

  • Deadline: 10 Feb 2026 @ 3:00am GMT+2
  • Prizes: $100,000 in cash + AI Futures Fund interviews
  • Register: Join the Gemini 3 Hackathon
  • Participants: 10,500+ builders worldwide

Prize Breakdown

PlaceCash PrizeBonus
Grand Prize$50,000Social promotion + AI Futures Fund interview
Second Place$20,000Social promotion + AI Futures Fund interview
Third Place$10,000Social promotion + AI Futures Fund interview
Honorable Mentions (10)$2,000 eachSocial promotion

Judging Criteria

  • Technical Execution (40%) - Quality development, Gemini 3 integration
  • Innovation / Wow Factor (30%) - Novel ideas, unique solutions
  • Potential Impact (20%) - Real-world usefulness, problem significance
  • Presentation / Demo (10%) - Clear problem definition, effective demo

Why join? Gemini 3 represents the next leap in AI reasoning and multimodal capabilities. This is your chance to build agents that serve humans using the Empowerment Imperative framework - and prove that ethical AI is also better AI.


Acceptance Criteria

  • Goal Alignment - Explicit user goals, not just behavioral inference
  • Consent Architecture - Personalization requires opt-in
  • Constraint System - Hard stops on exploitation patterns
  • Dual Metrics - Track both performance AND agency
  • Empowerment Demo - Show capability expansion, not just conversion

Bonus

  • Alienation detection with real-time alerts
  • Trust accumulation visualization over time
  • Integration with commerce platforms (Shopify, Stripe, etc.)
  • Comparison mode: extraction agent vs. empowerment agent on same task

The Code

We’re building this in public. The repository will contain:

  • Goal Alignment Engine
  • Alienation Detection System
  • Agency Optimizer
  • Constraint Checker with dark pattern detection
  • Dual metrics dashboard

Repository: Coming soon at github.com/ai-knowledge-hub


Discussion Prompts

  • Where is the line between “personalization” and “exploitation”?
  • Can empowerment optimization be as profitable as extraction? More profitable?
  • How do we measure “agency” in a way that’s rigorous and actionable?
  • What happens when an empowerment agent competes against extraction agents?
  • Could we build industry standards around the Agency Layer?

Add your thoughts in the comments or our internal #empowerment-imperative channel.


Reference Pack


Call to Action Use the Empowerment Imperative framework to build agents that serve, not exploit. Join the Gemini 3 Hackathon, ship working code, and show the world that the fork in agentic marketing leads somewhere worth going. The code is the argument. Let’s write it together.

Discussion & Idea Voting

Up-vote next week’s build idea by reacting with 👍 to any comment.

Published on Saturday, January 3, 2026