An eight-week course in setting direction, delegating useful work, and building an organization that can move forward without your attention at every step.
The shift
AI pair programming puts you alongside an agent, choosing and checking each next step. Autonomous work starts with a useful outcome: you establish intent and constraints, then let coworkers decide how to carry the work forward.
AAIO helps you develop that way of working. You’ll practice delegating, inspecting evidence, learning from failure, and changing direction when the result calls for it.
Build a new product, improve an existing repository, or bring the practices into your engineering organization. A team can be one person working with AI coworkers.
What you’ll practice
Bring something you want to accomplish. Each week, try the practices on real work and bring back what happened.
Develop foundational, strategic, and tactical Intents that connect daily work to what you’re trying to achieve.
Give coworkers clear responsibilities, organize their context, and redirect ongoing work without prescribing every step.
Trace what happened. Decide whether to change the intent, instructions, runbook, tools, or code. Then try again.
Evaluate useful results, test recovery, and leave enough recorded direction for a fresh session to continue.
The curriculum
Preparation before class, discussion and demonstrations together, and independent work on your project or organization.
Explore fleets and swarms, coworker communication, mission command, fractal organization, and PARA. Draft your organization’s foundational Intent, then develop useful strategic and tactical Intents.
Delegate a small assignment, inspect the result, and trace where execution diverged from your direction. Begin debugging and testing whether recorded context works from a fresh session.
Use blameless postmortems to understand disappointing runs. Introduce Superbuild’s build-and-review workflow: use the result, review it, fix findings, and verify the changes.
Divide related assignments, define boundaries and handoffs, and manage dependencies. Use stacked pull requests and small reviews to make integration easier to inspect.
Communicate changing priorities through intent and constraints. Practice deciding when coworkers should act independently, when they should escalate, and when you should intervene.
Investigate longer runs, grade useful results, and test interruptions and recovery. Identify what your setup needs before you can rely on it for production work.
Bring the practices together on your new-project or organizational-integration path. Collect evidence, examine the gaps, and improve the workflow ahead of your final demonstration.
Show useful autonomous work, explain how it was coordinated, and evaluate its limits. Aim for roughly 48 hours of autonomous work or 4–5 agents working toward one goal; these are alternative targets, not rigid thresholds. Human steering remains available.
Ways to take part
Work asynchronously, join the live cohort, or add more support for you and your team. Prices below are the planned course offer.
On your schedule
$1,000 upfront
or $1,250 across three phases
With the cohort
$3,000 upfront
or $4,000 across three phases
For you and your team
$5,000 upfront
or $7,000 across three phases
Live attendance is limited to 20 participants, including up to five Premium participants. These are course capacity limits. Enrollment and availability will be confirmed when registration opens.
Before you begin
You should be comfortable working in a repository, running an application, and evaluating code and tests. Prior agent experience is optional. Introductory programming is outside the course.
Use your own AI accounts and cover their usage costs. Agent choice is open. You can build something new or work on a real problem in your existing organization.
AAIO / Autonomous AI Organizations
Registration details and final scheduling will be added here. Enrollment is not yet available through this page.
Review the eight-week outline