VCN #49: THE SWARM - BEFORE DOORS
Run parallel coding agents in git worktrees
Wednesday 2026-08-12, 19:00-22:00 PT, Frontier Tower 10th Floor Annex, San Francisco

This file is agent-readable on purpose. Point your coding agent at it and let it
do the setup, or run the commands yourself. Five minutes, done before you arrive.


WHAT YOU NEED IN THE ROOM
-------------------------
1. A laptop. Laptops are mandatory for this one - it is a build sprint, not a talk.
2. A repo you actually work in, WITH MORE THAN ONE THING TO DO. This is the real
   prerequisite and it is not a formality: the whole night is about splitting work
   across agents, so a repo with a single task cannot be fanned out. Come with
   three things you have been meaning to do. Any language.
3. git 2.5 or newer. Worktrees have shipped since 2015, so you almost certainly
   have this already. Check with:  git --version
4. Claude Code installed and working.
5. Your ticket. It carries the two things that make the labs free for you:
       - z.ai + Claude Code access for the session
       - Nebius Token Factory credits, which run the fleet of model calls


FIVE-MINUTE PREFLIGHT (do this at home)
---------------------------------------
From inside the repo you are bringing:

    git --version                       # want 2.5+
    git status                          # want a CLEAN tree. commit or stash first.
    git worktree list                   # want exactly one row (your main checkout)

Then prove worktrees work on your machine, and clean up after yourself:

    git worktree add ../sw-test -b throwaway/preflight
    git worktree list                   # want TWO rows now
    git worktree remove ../sw-test
    git branch -d throwaway/preflight
    git worktree list                   # back to one row

If all five commands ran without error, you are ready. That is the entire setup.


PICK YOUR THREE TASKS (the part that actually matters)
------------------------------------------------------
Write down three things to do in that repo. Then apply the one test we care about:

    Do any two of them need to edit THE SAME FILE?

If yes, they are ONE task, not two. Merge them and pick another. Tasks that share
a file will collide at merge time and you will spend the lab untangling it instead
of learning the pattern.

Good splits look like:
    - add an endpoint  /  fix an unrelated bug  /  write the README
    - three independent bug fixes in three different modules
    - tests for module A  /  tests for module B  /  a dependency bump

Bad splits look like:
    - "rename this symbol everywhere"          (touches every file: ONE task)
    - "add the model, then add the endpoint that uses it"  (B depends on A)


IF YOU LINK A SHARED VENV OR NODE_MODULES INTO YOUR WORKTREES
--------------------------------------------------------------
Read this twice. It is the one thing tonight that can destroy data.

Linking a shared .venv / node_modules into each worktree is the standard trick for
making N trees cheap to create. It also arms a failure we reproduced on 2026-08-12:

    git worktree remove --force ../sw/01     # exit code 0, no error printed
    ls ../precious                           # EMPTY. contents deleted.

`git worktree remove` (and `rm -rf`, and shutil.rmtree) FOLLOW the link out of the
worktree and delete the target's contents. Always unlink first:

    rmdir ../sw/01/.venv          # Windows junction: drops the link only
    unlink ../sw/01/.venv         # macOS/Linux symlink: same idea
    git worktree remove ../sw/01  # now safe

`rmdir` / `unlink` remove the link. `rm -rf` removes what it points at.


RUNNING ORDER
-------------
    19:00  doors
    19:30  walkthrough - the conductor pattern, worktrees, what they do NOT isolate
    20:15  hands-on build - stand up the swarm on your own repo
    21:30  demos + social
    22:00  out

You leave with a setup that runs multiple coding agents in parallel on one repo.


TICKETS
-------
$10 early bird, $20 at the door. Frontier Tower members free - reach out to any
host directly rather than looking for a code.

    https://luma.com/vcn-49-the-swarm


VIBE CODING NIGHTS
------------------
SF's builder series for people who code with AI agents. Wednesdays 7pm at Frontier
Tower Floor 10; Vibe Coding Mornings Saturdays 10am on Floor 9. The "Off the Leash"
season (#41-#50) goes deep on the serious stack: local models, sandboxes, VMs,
agent loops, evals, and self-hosting your whole coding-agent rig.

    https://vibecodingnights.com

Hosted by Rayyan Zahid (Immersive Commons), Michalis Vasileiadis (Hacker Bob),
Eric Mockler (AI Geneticist), Devinder Sodhi (Learning Layer Labs).