AI Workshop
Episode 6 of 10Aired

Rules of intelligent communication with AI

The workshop manifesto point by point: clarity, context, verifiability, honesty, polite persistence — and no trust without checking.

Wednesday 11 November at 19:00 Moscow time· 90 min

Aired

The recording is being prepared — it usually appears within two days. Sign up for the next streams so you don't miss them.

What we'll show

  1. All twelve manifesto rules on real agent conversations — do and don't
  2. How to disagree: a specific mismatch instead of «wrong again»
  3. How to ask for honesty: «say if you're not sure» and docs links instead of guesses
  4. Running a long project with an agent: notes, decisions, history in the repo
  5. Where the agent's responsibility ends and yours begins

What you take away

Twelve rules that turn chatting with an agent into a managed project.

Manifesto: 12 rules of intelligent communication with AI

  • Yes: Clarity before speed
  • Yes: Context is work, not a luxury
  • Yes: Define done before you start
  • Yes: Plan first, code second
  • Yes: Trust, but run it
  • Yes: Targeted edits, not rerolls
  • Yes: Honesty both ways
  • Yes: Polite persistence
  • Yes: Least privilege, no secrets in the chat
  • Yes: The agent works inside a harness
  • Yes: Know when to stop
  • Yes: Responsibility stays with the human

Harness in this episode

The manifesto is the cultural layer of the harness: rules that hold the line before automated checks exist.

The full harness episode

What to have ready

  • Yes: A couple of your own failed AI conversations — we'll take them apart

Official documentation