The Gathering¶

Field	Value
Status	Active — seeding stage (2026-04-22)
Type	Recurring small-group forum for entrepreneurial and technically-curious peers
Format	Host-rotation and/or shared third-place venue; possible recorded / published output
Location preference	Seattle-area, in-person preferred; remote participants OK but not the default

Description¶

A recurring small-group forum (roughly 4–8 people) for long-form, honest conversation among entrepreneurial and technically-curious peers. Tone closer to a private roundtable than a panel — wide-ranging, non-political, values-aligned. The group meets in person on a regular cadence and may, optionally, record conversations and publish them as audio or video.

The name "The Gathering" is deliberately low-key — it signals a recurring meet, not a brand or product. It sits alongside rusty-cage (the band) and happy-body (the resistance-training group) as a real-world people graph — not a business in itself, but part of the substrate that real work gets done on.

Why this is a project, not just a wish¶

Recurring peer forums only happen if someone designs them on purpose. Listed here so the naming decision, format choice, and seeding progress have a single source of truth, the same way any other long-running personal project does in this repo.

The fleet thesis on goals argues that the agent era rewards people with deep context and trusted access; the corresponding observation in the people-graph is that deep context and trust compound over years in small, recurring groups — not one-off conferences. The Gathering is the deliberate investment in that graph.

Candidate formats¶

Three shapes are on the table, in rough order of overhead:

Host-rotation at homes. Each member hosts once per rotation (every 4–8 meetings depending on group size). Dinner or heavy appetizers + a loose topic. Low production cost; highest intimacy.
Shared third-place venue. Same group, same bar / coffee shop / coworking spot on a cadence. Less setup per meeting; less intimate than a home; easier for members who don't want to host.
Hybrid with media output. Either of the above, but recorded (audio and optionally video) and lightly edited for public release. Not a performance — a real conversation that happens to be audible afterward. Adds a permission / consent layer and a small production workflow (microphones, edit pass, hosting).

Decision between them happens after the first 2–3 meetings show what the group actually enjoys. Defaulting to (1) or (2) for the first few and only adding (3) once the chemistry is obvious.

Cadence (tentative)¶

Monthly meetings as the default rhythm. Quarterly is too infrequent to build momentum; weekly is too heavy for people with jobs / families.
Weeknight preferred (Tue–Thu) over weekends, to protect family weekends.
2–3 hour window per meeting.

Success criteria¶

First gathering: one meeting, ≥ 4 attendees, by end of 2026 Q3.
Group viability: ≥ 6 meetings with a stable core of ≥ 4 attendees across a year.
Optional media milestone (only if the group wants it): at least one published recorded conversation by end of 2026.

Explicit non-goals (for at least the first year): monetization, sponsorship, public-facing brand building, growing the group beyond a single table's worth of people.

Seeding¶

Step zero is assembling an initial 4–6 people. Criteria: entrepreneurial or builder-adjacent, technically curious (AI, agents, hardware, or similar), independent-minded rather than tied to a political ideology, comfortable with long-form discussion. Seattle-area preferred; a couple of remote-friendly exceptions fine if the chemistry is right.

The operative principle: channel choice does most of the filtering. A candidate pool sourced from costly-signal channels (warm intros, niche builder communities) is mostly pre-screened. A candidate pool sourced from a broad open call (Craigslist, generic Meetup post, Reddit) arrives unfiltered and forces every screen to be done manually on strangers. The cost curve strongly favors exhausting the former before touching the latter.

The actual seed list is not kept in this public wiki — it lives in local-only notes.

Seeding channels, ranked by pre-filtering strength¶

Warm intros from trusted friends. The highest-signal channel. Direct ask: "who's the most interesting builder-type you know that I don't — not a networker, someone you'd actually want at a dinner?" Friends' judgment does most of the ideology and chemistry screening for free.
Second-degree from existing anchors. Adjacent surfaces in the personal graph — rusty-cage audience, happy-body regulars, existing peer connections from the wiki's project surface. These people already self-selected into something values-adjacent, so overlap is non-trivial.
Builder communities with skin in the game. Seattle AI Tinkerers, PyData Seattle, Cascadia.js, Indie Hackers Seattle, local LLM / agents meetups, hacker-house events. Attend regularly and spin off a smaller dinner ("four of us are getting food after the next one, want in?") — turns the meetup into a funnel without ever posting a public ad.
Content-as-beacon. A public post — even a 600-word version of the fleet thesis, or a pointer from goals — and let people opt in after reading. Anyone who reads long-form strategy writing and still reaches out has already passed a filter that couldn't have been designed manually.
Tight-topic one-off workshops. Host a single-session thing ("three-hour Saturday on agents and context engineering, 6 people, my place") and invite the two best attendees to the recurring forum. The workshop itself is the screen.
Broad open call (Craigslist / generic Meetup / Reddit). Worst signal-to-noise. Not strictly zero value, but channels 1–5 should be exhausted before this is on the table at all. For a 4–6 person seed, this tier is probably never reached.

Screening¶

Regardless of channel, layer these in order. All of them apply before anyone gets invited to a recurring meeting.

No jump from intro to invite. Always a 60–90 minute 1:1 first — coffee or a walk. This is the single most important gate. It screens the candidate, and it screens whether the conversation is actually enjoyable for an hour-plus.
Invites framed as a single dinner, not membership. Low commitment in both directions. If chemistry is off after one meeting, no one has to un-join anything.
Probe questions that reveal orientation without asking about it. Ideologues give themselves away in 30 minutes of normal conversation without needing to be asked directly:
- "What are you building or working on right now?" — filters non-builders immediately.
- "Who do you read or listen to regularly?" — strong signal. Builder / long-form / generalist figures correlate well; primarily-political pundits on either side are a yellow flag.
- "What have you changed your mind about in the last year or two?" — the strongest single filter. Ideologues genuinely struggle to answer; builders and genuinely curious people answer it instantly and often with pleasure.
- "What's a take you hold that most of your friends disagree with?" — similar test for independent thinking without being partisan bait.
Steelman check. In the first conversation, mention a position the candidate likely disagrees with and see if they can represent it fairly. Independents steelman naturally. Tribalists can't help themselves.
The "everything is political" tell. When the non-political framing of The Gathering is described, watch the reaction. A healthy candidate nods or says "yeah, I'm tired of that too". A candidate to pass on pushes back with "well, everything is political" — a reliable ideologue tell. Not always disqualifying on its own, but a strong yellow flag.
Right-of-refusal, both directions, made explicit. Norm from meeting one: anyone can bring a guest to a future dinner, and the organizer can quietly not re-invite. No votes, no membership, no drama.

Operating principles¶

Chemistry ceiling is set by the weakest member, not the strongest. One ideologue or one tedious conversationalist lowers the quality of every future meeting. Tolerating a marginal candidate is more expensive than leaving a seat open.
Err on the side of 4 great people over 6 mixed ones. The instinct under time pressure is to fill seats; resist it.
Warm intros over open calls, always. If a broad ad is being drafted, that's a signal channels 1–5 haven't actually been worked yet.
Screen for what the group is for, not just what it's against. Non-political and non-ideological is a floor, not a ceiling. The positive criteria — actively building things, technically curious, willing to change their mind — are what make a conversation worth showing up to.

What-Im-Working-On — lives under Context / other lines; not an income project.
goals — aligns with the fleet thesis by building the in-person peer graph that trust and context compound across; does not occupy a fleet-roster seat (it's people, not an agent).
rusty-cage, happy-body — existing people-graph anchors, same category of "recurring real-world surface."

Meta¶

Created: 2026-04-22.
Trigger: Strategic review named the absence of a peer-forum surface as an unstaffed gap, and endorsed the name "The Gathering."
Next review: 2026-07-18 (end of current focus window) — seed list progress, any meetings that happened, first cadence signal.