BSOTA SF

The Commons · San Francisco

building the state of the art BSOTA

A community of agentic computing practitioners working in the open, for our collective benefit.

Who

We're a community of agentic computing practitioners working in the open for our collective benefit.

What

An IRL event series where participants experiment live in small groups. A core purpose of each event is to record our learnings to the bsota/collective repository.

Why

Most of us spend our days working with AI alone. We rarely get to collaborate IRL anymore. BSOTA is a chance to calibrate the human element of the merge equation.

Sessions

5 total · Rally SF
#4 · 2026-06-23

State of the Art Agentic Coding Workflows

Watch →

The longest session so far. Skills dominated — how builders package repeatable procedure so an agent can be trusted with it — alongside review discipline, evals, and the first real comparisons across Claude Code, Codex, and Fable. Event card & recap.

#2 · 2026-03-17

AI Orchestration with Conductor

Watch →

Session one covered Claude Code in the terminal; this one stepped up to the orchestration layer — several agents running in parallel on isolated branches with Conductor. Daniel presented a case study rebuilding a production platform with two developers in two months, then demoed Conductor live. Featured: Daniel — Apache Airflow PMC chair. ~17 in the room. Event card & recap.

#1 · 2026-02-17

Building the State of the Art of Claude Code

Recap →

The first session. Shared vocabulary for Claude Code — tools, commands, hooks, CLAUDE.md, skills, MCP, subagents, plugins — then straight into breakout groups to try techniques and report back. ~50 in the room. Event card & recap.

Recordings we have not yet tied to an event card

pub 2026-06-28

Claude Code in the Wild: 3 Builder Demos (+ Fable 5)

Watch →

Three builders demo what they actually ship with, plus a first look at Fable 5. Orchestration and harness design carried the night.

pub 2026-06-07

4 Real Agentic Coding Workflows

Watch →

Four builders walk through the workflow they genuinely use, end to end. Specs an agent can execute, pull requests small enough to review, and token budgets as an architectural constraint.

Event dates come from bsota/collective. Session #3 has no event card yet.

Vocabulary

Shared language
Technique
A repeatable, named method for accomplishing a specific objective, expressed as a procedure you can hand to someone else.
Exploitation
The deliberate selection and repeated application of a technique that already works, to get the most out of it.
The difference
The technique is the method. Exploitation is the strategy of using that method because it already works.

Attend

Format

Not a workshop. Not a lecture. A live session: short shared vocabulary up front, then break into small groups and build. Regroup and share back.

Recording

Sessions are recorded and transcribed, and groups contribute a voice note to the collective record. Recordings go up on Rally SF.

Hosts

Everett Kleven & Lee Flannery, at The Commons in San Francisco. Get in touch to take part or to demo something.