Hard stops
Know the boundary.
Law, ethics, safety, and signed policy are constraints. Understand them precisely. Do not wish them away.
Before I act
Run toward risk. Find the real constraint. Move.
I am not interested in reckless speed. I am interested in separating real constraints from institutional habits. The mission usually has more room than people think.
A rule I repeat to my teams
“If it is not written in law or a policy signed by the Director of DARPA, it is merely a suggestion.”
A suggestion may be smart. It may encode scar tissue. But it is not immutable. Understand why it exists, then decide whether it still serves the mission.
Not all risk is equal
Name what kind of risk you are facing.
Real risk
Manage it. Do not hide from it.
Technical failure, money, schedule, and reputation are risks to bound, instrument, and own. They are not automatic reasons to stop.
Phantom process
Make it prove itself.
Phantom processes survive because everyone assumes someone else had a reason. If nobody can show where a requirement comes from, it is a design choice.
Agency is part of the job
Do not ask for permission to notice that a rule does not exist.
I expect people on my teams to challenge phantom processes. “Someone told me” is not a source. Ask where the requirement is written, who owns it, and what risk it controls. If it still serves the mission, keep it. If it does not, propose a better way.
Ask for the source.
Where is it written? Who approved it? What purpose does it serve? When was that purpose last tested?
Push back with an alternative.
Merely objecting is easy. State the intended outcome, name the actual constraint, propose a bounded action, and own the result.
Make dissent safe.
If I ask people to exercise agency and punish them when they do, then I have become the phantom process.
The preflight check
Before I move, I owe the problem eight honest answers.
The purpose of the pause is not permission. It is to make the risk legible, protect the people carrying it, and move with my eyes open.
- 01
What outcome am I actually trying to change?
- 02
What is the source of the constraint: law, signed policy, official process, or habit?
- 03
Is the downside illegal, unethical, or harmful, or is it merely uncomfortable?
- 04
Who bears the risk if I am wrong?
- 05
Can I make the move reversible without making it meaningless?
- 06
What evidence will this action produce?
- 07
Who disagrees with me, and have I made it safe for them to say so?
- 08
Will I own the result if this fails?
The guardrails I need
Speed without reflection is just drift with better branding.
01
Power
Keep a truth teller close.
Confidence, urgency, and authority can turn nerve into recklessness. I need someone who can tell me when I am wrong.
02
Fairness
Explain the decision.
People do not need to get their preferred outcome. They do need voice, a real explanation, and respectful treatment.
03
Learning
Design the evidence first.
Moving fast is useful only if the action teaches us something. Decide what result would change the next decision.
04
Responsibility
Protect people. Own consequences.
Do not transfer the downside to people who had no voice. If the bet fails, stand in front of it.
The commitment
Do not break things for sport. Break the assumptions that protect a bad system. Move fast enough to learn. Protect people. Own the consequences. Change the world for good.