In the previous issue, I explored the "disagree" side of Have Backbone; Disagree and Commit. This time, let's tackle the harder question: how do we commit to a direction when everything changes every few weeks?
A new model drops. A competitor pivots. A capability you planned around becomes obsolete overnight. The temptation is to stay permanently in "disagree" mode - endlessly debating, waiting for clarity that never arrives.
But that's not what the principle asks of us. Once a decision is determined, they commit wholly. The question is: what does "wholly" mean when the ground keeps shifting?
Tool commitment vs. principle commitment
This is where most organizations get it wrong. They commit to tools - "We're a GPT shop" or "We're building on Claude" - and then feel betrayed when a better option emerges. That's not commitment. That's attachment.
Real commitment in the age of AI operates at a different level. It's not about which model you use. It's about why you're using AI at all, and what problems you've decided matter most.
I've found it helpful to think about this as a three-layer model.
Layer 1: First Principles (non-negotiable)
These are your decision rules. They don't change with the next model release. They might include:
- We only deploy AI where it creates measurable customer value
- Quality and trust are not negotiable - we don't ship what we can't verify
- A human is accountable for every AI-driven decision that affects customers
- Every initiative must generate measurable learning, not just output
These are the things you commit to wholly. When a new tool appears, when a competitor announces something flashy, when your board asks why you're not moving faster - these principles are your backbone.
Layer 2: Strategic Bets (quarterly commitment)
Below your first principles sit 2-4 priorities you've chosen for the quarter. Each one has an owner, a deadline, and - critically - kill criteria. What would make you stop?
This is where the "commit" part gets real. You've debated. You've disagreed. Now you pick a direction and go. Not because you're certain it's right, but because moving decisively with a good strategy beats endlessly optimizing for a perfect one.
The key: these are time-boxed. You're committing for a quarter, not forever. That's not hedging - it's disciplined experimentation.
Layer 3: Rapid Experiments (weekly iteration)
At the bottom layer, you run small-scale tests every week. These generate the evidence that either reinforces or challenges your strategic bets. A failed experiment isn't a failed commitment - it's exactly the kind of learning your first principles demand.
This is where speed lives. Not in changing direction every time something new appears, but in generating evidence fast enough to make your next commitment better informed.
Clarity of commitment creates speed
Here's the counterintuitive insight: the organizations that move fastest aren't the ones who stay flexible about everything. They're the ones who've been ruthlessly clear about what they will and won't change.
When your first principles are solid, you don't waste cycles debating whether to pursue a new tool. You ask: does it serve our principles better than what we have? If yes, switch. If no, ignore. Decision made in minutes, not months.
When your strategic bets are clear, your teams don't wait for permission. They know what they're building toward and can make hundreds of small decisions autonomously.
Clarity of commitment creates speed, not the other way around.
Your action step
Block 45 minutes with your leadership team. Create a one-page document with three sections:
- First Principles: 3-5 non-negotiable decision rules for your AI work
- Strategic Bets: 2-4 priorities for this quarter, each with an owner and kill criteria
- Rapid Experiments: What are you testing this week to generate evidence?
If you can't fill in all three layers, you haven't committed yet. And if you can't articulate kill criteria for your strategic bets, you're not committing - you're hoping.
The backbone isn't in picking the right direction. It's in committing clearly enough that everyone can move fast - and knowing exactly when it's time to recommit.