Thriving Quest 03 · Public build

Family Fun

Quality time with my family is precious. We are fortunate to have (too) many good options in San Francisco. Finding the right thing to do, noticing it early enough, comparing calendars, buying tickets, and arranging childcare with enough lead time can be overwhelming.

This thriving quest is to create more meaningful and engaging family quality time with less work arranging it. Without any dogma, my target is one fun family event each week and one culturally stimulating date night each month.

Current problem

The best family experiences are scattered across venue calendars, ticket pages, memberships, and half-remembered ideas. Often, by the time I see good options they are no longer an option.

Thesis

Agents can deepen our family connection by improving the quality and quantity of fun family events.

Current progress

An owned possibilities calendar is live on my phone with direct-source discovery, preference filters, decision deadlines, and three planning cadences. The surfacing and decision cadence needs improvement.

Follow along

Subscribe to the . The build notes follow what the system finds, what we choose, what we enjoy, and what gets filtered out next time.

The vision

More meaningful and engaging family quality time; less work arranging it.

I want the planning to fade into the background. Good options that match our preferences should intuitively surface with enough lead time to discuss and arrange logistics. Admin and cognitive overhead is minimized, and time with my family doing cool stuff uniquely available where we live is maximized.

The quest has two lanes. The first is family time: storytimes, music, museums, nature, hands-on programs, and neighborhood discoveries that are age appropriate for our kid. The second is time with my lovely honey: world-class dance, music, theater, exhibits, and other cultural happenings that feel worth arranging childcare for.

The possibilities calendar sits between the event firehose and our commitments. It can be generous without being pushy. Every item is visibly tentative, and only a human choice turns an option into a plan.

Together as a family

One fun event a week

Local, engaging experiences that fit our child's age and interests, the household calendar, and the energy we actually have.

Together as a couple

One date night a month

Experiences that justify the tickets, coordination, travel, and childcare because they strengthen the relationship at the center of our family.

Design intents

  1. Connection over planning - reduce the overhead around time together
  2. Possibilities stay possibilities - suggestions never mark us busy or impersonate commitments
  3. Surface the decision, not only the event - tickets, childcare, and coordination get their own deadlines
  4. High signal on the phone - weak matches stay out of the visible calendar
  5. Trust official occurrences - first-party evidence beats inferred schedules and aggregator volume
  6. Learn from real life - attendance and enjoyment teach preference
  7. Protect the people in the story - family privacy outranks public detail

Success criteria

The Family Fun loop

  1. 01

    Discover

    Read approved first-party event sources.

  2. 02

    Filter

    Apply interests, availability, and logistics.

  3. 03

    Surface

    Put timely options and deadlines on the phone.

  4. 04

    Choose

    A human decides what becomes a commitment.

  5. 05

    Experience

    Be together. Let the system disappear.

  6. 06

    Learn

    Use what worked to sharpen the next set.

Status (2026-08-19)

The discovery and calendar system is live. Direct venue sources feed a preference-aware option set; weekly, monthly, and quarterly reviews separate quick choices from long-lead planning. The technical proof is ahead of the family proof. The next phase is sustained use against the weekly and monthly targets.

What's working

  • The option set is owned - portable calendar records sync to a normal phone client
  • Possibilities stay tentative - suggestions remain visible without blocking time
  • Two clocks are explicit - event dates and decision deadlines no longer blur together
  • Household constraints are real rules - unavailable times fail before preference scoring
  • Bad evidence fails closed - unsupported recurrence inference cannot publish

What's not (yet)

  • The target family and date-night rhythms have not yet run for three months
  • Human-reviewed candidates do not automatically teach the preference model
  • Several promising venues still lack narrow, tested first-party adapters
  • The option set needs continuing calibration so useful abundance does not become noise
  • The options and review cadence do not have a shared, intuitive surface

Next loop

  1. Use the weekly set - make choices, dismiss the rest, and record why
  2. Close the attendance loop - distinguish booked, attended, enjoyed, and regretted
  3. Sharpen together - make the option set useful to both adults in the family
  4. Prove the rhythm - sustain the weekly family event and monthly date night targets for three months

Key milestones

  1. Date-night preferences mappedCompleted milestone

    The first planning map captured the performances, venues, logistics, and calendar constraints that make a night out worth the effort.

  2. Technology-independent calendar deployedCompleted milestone

    A standards-based CalDAV service went live on an isolated host with private accounts, HTTPS, backups, and a clean path to replace any one client or server later.

    Changelog entry →
  3. Possibilities reached an ordinary phone calendarCompleted milestone

    The owned calendar completed a phone read/write round trip. The delightfully boring result: CalDAV basically works.

    Changelog entry →
  4. Possibilities separated from commitmentsCompleted milestone

    A tentative, transparent Family Fun calendar now holds options without marking time busy or pretending every suggestion is a plan.

    Changelog entry →
  5. Decision deadlines addedCompleted milestone

    Every possibility gained a second clock: the date of the experience and the earlier date when tickets, childcare, or coordination require a decision.

  6. Direct-source discovery expandedCompleted milestone

    Ten first-party venue sources began feeding a preference-aware discovery pipeline while routine programming and weak matches stayed out of the phone calendar.

  7. Unsupported schedule inference eliminatedCompleted milestone

    An irregular event series exposed a bad recurrence assumption. The adapter now requires explicit occurrence evidence and fails closed when a schedule cannot be verified.

  8. Weekly, monthly, and quarterly loops establishedCompleted milestone

    Separate reviews now surface quick opportunities, coordination decisions, and long-lead experiences on the clock each category needs.

  9. One family event a week for three monthsUpcomingUpcoming milestone

    The family rhythm sustains an average of one fun shared experience each week across a rolling three-month window.

  10. One meaningful date night a month for three monthsUpcomingUpcoming milestone

    The planning loop sustains one meaningful date night each month across a rolling three-month window, without a recurring logistics scramble.

Quest log

The possibilities calendar reached my phone

A standards-based CalDAV calendar put high-signal family outings and date-night options on my phone without turning them into commitments. The technical side quest worked; the real test is whether the options lead to more time together.

Essays from this quest

None yet - mature quest-log threads graduate here.