Development Review
Artifact info
Title:

Hollow Movement Platform — Membership & Matching Development Sync

Engagement:

Client:

Meeting Date:
April 15, 2026
Next Meeting Date:
No items found.
People
James Redenbaugh
Ivan Gonzalez
Artifact Image
Meeting Summary

🎯 Session Context & Priorities

James and Ivan reconnected after a productive stretch — Ivan wrapping up a previous engagement and available to take on more work, James returning from a weekend trip to Tennessee with family. With the Hollow Movement Wave event deadline set for May 29th, the conversation focused on two major development priorities: getting the membership and subscription system fully functional, and designing the matching algorithm architecture.

---

💳 Membership & Subscription System

The immediate priority is getting a working pay-what-you-want subscription model implemented via Stripe [tag="stripe"]. PayPal was considered but deprioritized due to integration complexity — Stripe is the clear path forward.

The membership signup flow needs to be embedded directly into the profile creation process. Since users answer domain questions and configure their profile during onboarding, the subscription step should appear inline within that flow. For users who've already subscribed and return to edit their profile, that panel should simply not render — a conditional check against existing subscription state (06:38).

Account page updates are also needed so members can view their current plan, change their tier, and cancel if needed. A free tier option will be present in the signup flow as well, so users can explicitly choose to stay free rather than having it be a default (08:52).

The team still needs to define what's included in free vs. paid. Current thinking:

  • Free: Core directory access, basic profile, limited connection features
  • Paid: Courses, messaging, premium matching features, agent-generated connection profiles
  • Matching visibility may require completing the connection assessment, giving people an incentive to do it

[technology="Custom Membership System"]

---

🔗 Matching System Design

Stage 1 — Algorithmic Matching via Connection Assessment

The suggested connection map will be a grid visualization with two axes: proximity (Y-axis, calculated from latitude/longitude in Supabase [tag="supabase"]) and alignment (X-axis). For V1, alignment will be calculated using results from the connection assessment rather than the more complex tag/domain comparison — a deliberate simplification that also incentivizes more users to complete the assessment (13:12).

The connection assessment maps each user across 12 domains, each with three possible archetypal vectors (grounded, visionary, integrative). A first-pass binary/trinary comparison can identify users who share 11 or 12 out of 12 archetypes extremely quickly — practically in nanoseconds per comparison. A second pass then refines those top matches using the actual numeric scores for finer differentiation (32:34). For example, two users might share the same archetypes but one is meaningfully closer in their exact positioning across domains.

Computational strategy discussed: calculating every user against every other user on page load is likely too intensive at scale. The preferred approach is periodic background processing — pre-calculating and storing match scores in Supabase [tag="supabase"], then retrieving them when a user loads the connection page. Ivan suggested starting with a spike using a subset of users to get real performance metrics before committing to an architecture (21:18). Limiting the map to top matches (e.g., top 20–50) also reduces the calculation scope considerably.

Tag and domain-based alignment (V2) was tabled for now. It involves comparing four taxonomies — domains, custom tags, offering tags, and seeking tags — where seeking should be matched against offering for high alignment. The complexity of that cross-comparison is significant, and the connection assessment alone gives a cleaner signal for launch.

[technology="Intelligent Matching Algorithms"]

Stage 2 — Agent-Generated Connection Profiles

For deeper one-to-one matching, an AI agent will generate a detailed connection profile analyzing two users' assessments, text responses, tags, seeking/offering, and Holon memberships (37:45). The output — a structured paragraph explaining alignment and where collaboration potential lies — will populate a template on the connection page. Claude [tag="claude"] and N8N [tag="n8n"] automations are the planned implementation path, similar to a prototype James had already built.

Generated connection profiles will be saved as records in Supabase [tag="supabase"] so they don't need to be regenerated on every view. If User A generates a profile with User B, User B sees the same result. Users can trigger a regeneration if they've updated their profiles. This feature is positioned as a premium paid feature, though free users may get one trial run (41:49).

[technology="Communication Automations"]

Matching Gating & UX

Users who haven't completed the connection assessment won't see the connection map — instead, they'll see a prompt encouraging them to complete it. This creates a natural incentive loop. A loading state (similar to the existing profile creation loader, which takes ~45 seconds while N8N [tag="n8n"] generates the background image and tagline) will handle the wait time for on-demand agent runs (39:07).

---

🌐 Longer-Term Platform Vision

James outlined the three-level roadmap for the matching and collaboration engine:

  1. Individual-to-individual — connection profiles between two users
  2. Holon group analysis — skill alignment, blind spots, and gap identification within a working group
  3. Holon-to-Holon and network-level — cross-group recommendations, surfacing the right person from the broader network for a specific team's needs

Beyond matching, the platform's longer arc is about facilitating an engine for social good — helping people find grant funding, build project teams, and track and share impact. Gamification of achievements (micro-grants won, projects completed, outcomes reported) is planned as the platform matures over the next 18 months (45:30).

---

🛠️ Platform & Project Management Notes

The Join Holon button is currently non-functional and needs to be fixed before the conference. The project board has also fallen behind — James acknowledged that task tracking has become a personal pain point with too much living in his head rather than in a maintained system (47:26).

Ivan shared that he's been running a lightweight AI-chat-based project management approach using Claude [tag="claude"] — markdown files for the roadmap and per-ticket detail, with conversations to surface what's outstanding. He agreed to write up a summary of how his system works and share it with James (51:20).

James also briefly showed a personal site prototype (james.today) designed entirely by Claude [tag="claude"] — a clean, minimal aesthetic for real-time client connection. A separate domain, sensemaking.today, is a seedling idea for a collective AI-assisted sense-making dialogue project, bringing diverse perspectives together to understand rapid global change (55:30).

[technology="Collaboration Management Tools"]

---

Action Items

James Redenbaugh

  • Implement pay-what-you-want Stripe [tag="stripe"] subscription in the membership signup flow (06:38)
  • Build conditional rendering logic so the subscription panel is hidden for users who already have an active membership (08:11)
  • Add membership management to the account page (view plan, change, cancel) (08:52)
  • Grant Ivan access to the Figma matching UI designs (12:21)
  • Define free vs. paid feature tiers with the team ahead of May 29th (09:22)
  • Fix the Join Holon button functionality (47:08)
  • Share the collective sensemaking initiative write-up with Ivan (01:11:58)

Ivan Gonzalez

  • Spike the matching algorithm computational feasibility using a subset of users and Claude [tag="claude"] (21:18)
  • Research and document approach for the two-pass connection assessment comparison (proximity + trinary archetype match → score refinement)
  • Share a written summary of his AI chat-based project management system with James (51:20)
  • Recreate his platform profile and complete the connection assessment once the join Holon flow is functional (46:57)

Both

  • Continue async thinking on matching architecture and reconvene with individual findings to compare approaches (23:32)
Relevant Initiatives
No items found.
Transcript