My coffee mug says: Discipline is simply remembering what you really want.

"Oooh!" I thought as I painted it up, "Every time I drink coffee from this mug, I'll remember what I really want, and then I'll only do the right things all day." Years later, I enjoy my coffee and chuckle. What I really want. Which me? The me that set goals in January, or the me that baked delicious brownies last night? Brownies go great with coffee.

In the moment when I’m deciding whether to write that next chapter or check my messages, I don't need to remember anything. I’m completely aware of my goals, and I’m still probably going to check those messages. Because even when I remember what the best version of myself "really wants", the right-now version is running the numbers. The numbers come out in favor of LinkedIn. That's business too, and who knows, right?

Those numbers. I have an internal bookie. So do you. (I call mine Harry.) It/Harry works 24/7. Harry has perfect recall of every bet I've ever placed, every return I've ever seen, and every time I've promised myself things would be different. Harry/It evaluates the options in play, then decides which actions are most likely to pay off.

The equation bookies run works something like this:

Internal bookies run this for every choice, constantly. It's also the root equation underlying team prioritization frameworks: effort vs. impact, cost vs. benefit, urgency vs. importance. Those frameworks ask the bookies to show their math.

Unfortunately for a lot of people and teams right now, the math isn't mathing. More teams are feeling overwhelmed, overbooked, and off-track. They’re doing everything "right" and wondering why it feels like a waste of time. To understand why, we need to look at where the bookie gets its data.

Your bookie doesn't calculate probability from first principles. It retrieves it.

Neuroscience calls it the default mode network. It's the brain's background processing system, most active when you're not focused on an immediate external task. Research published in Nature Reviews Neuroscience describes the DMN as specialized for "recall of past events and internal simulation of future events." It replays your history, projects likely futures, and pattern-matches the current situation against your experience. In short, the bookie's odds are pulled from your autobiography.

This is also why the brain is better described as a prediction machine than a reasoning machine. According to predictive processing research, the brain uses past experiences to build a mental model of the world. It then predicts the best course of action based on that model, and only updates the model when something in the environment violates the prediction strongly enough to demand revision. Small violations get explained away. Big ones force an update, but even then, the update is conservative.

The bookie changes its model as little as possible to account for the new data. Karl Friston calls this minimizing prediction error, and it's why deeply held beliefs are so resistant to change even in the face of contradictory evidence. The model protects itself. Your prior experiences shape what you notice, what you treat as a viable option, and what gets dismissed before it reaches conscious evaluation. You don't observe the world and then decide. You predict, act, and update.

Here’s one way this equation plays out between Harry and me.

I want to master the pull-up. My vision: I’m running through a field. Ahead, a mighty ancient tree beckons. I answer the call, leaping into the tree’s arms by pulling myself up onto a branch lovingly extended just above my head.

As a woman of a certain age, with a certain amount of heft to my…assets… and minimal upper-body strength, you are correct in assuming that achieving this goal will take some serious discipline.

But ok, whatever. I believe it’s possible. I did the research, and I have a plan.

But my bookie, Harry, is not taking that bet. His math:

Efficiency! The bookie keeps us from frittering minutes on things it doesn’t believe will pan out.

This is what good navigation looks like in a complex environment. You don’t know how things will play out, so you make your best guess with the information you have and move ahead.

Systems thinkers describe complex adaptive systems as environments with multiple interacting agents, shifting conditions, and outcomes that only make sense in hindsight. You can't optimize from first principles because the system is too dynamic to fully model.

We like to think that if we're just smart enough, if we do more research and consult more data sources, we can figure out the "right" thing to do every time. It's why kids truly believe that, if they had bank back in 2010, they would have bought way more bitcoin, Apple stock, and tickets to Prince concerts than their parents did.

But in complex environments, cause and effect are only legible after the fact. (New to this? See Dave Snowden's Cynefin framework for reference.) Good navigation in complexity depends on accurate pattern recognition and the ability to update those patterns fast enough that your next prediction is better than your last.

The brain's prediction machinery handles complexity well when the pace of change matches what our evolution prepared us for. Patterns are largely reliable. Updates are infrequent. The bookie has good data.

The problem we face now at work is that the update cycle can't keep pace with the rate of change. Patterns go stale before the next prediction runs. The bookie is still doing exactly what it's supposed to do, but it struggles when the game and the players keep changing.

The result?

Prioritization has always been hard. Two compounding factors make it harder now:

Optionality expanded faster than our frameworks expect. You can change careers every few years without stigma. You can launch a product, publish a book, or build an audience with tools that didn't exist five years ago. AI compresses the effort required for dozens of tasks that previously took years and specialists. Everything is, theoretically, on the table.

When options are limited — by geography, economics, social role, or plain logistics — the bookie has a short menu to evaluate.

When the menu explodes, the bookie has to evaluate everything. This is exhausting. And it leads to a particularly modern kind of failure: we keep adding commitments because each new option genuinely scores well on the equation in the moment we encounter it. The list grows. Nothing finishes.

And at the same time...

The bookie's data has become less reliable. No matter how much computing power you throw at it, bad data will always give you the wrong answer.

The DMN builds predictions from stable, repeated experience. When context shifts faster than the update cycle, those predictions become confidently wrong.

Software teams have spent years using impact/effort matrices to evaluate features. The effort side of that equation is now partially wrong in unpredictable ways: AI collapsed certain costs while leaving coordination overhead, legacy code, and technical debt largely unchanged. The impact side is foggier still because the market is shifting. Will people still pay big bucks for software when they believe they can write their own? The prioritization framework isn't broken. The priors that feed it are.

And at the individual level: when you can't reliably distinguish authentic information from sophisticated fabrication — when the media environment makes your pattern recognition about "what's real" actively untrustworthy — the DMN is retrieving priors from a context that no longer exists. The bookie is quoting odds on a race that's been cancelled.

Complexity scientists call this a turbulent environment: one where volatility has exceeded the system's capacity to adapt through normal pattern recognition. Your brain evolved for complexity. It did not evolve for this rate of change.

When you're navigating through uncharted territory, there's no way to optimize your route. Instead, the complexity folks seek to reduce the possibility space, establish clear attractors, and navigate with simple rules robust enough to hold under pressure. These are the documented mechanisms by which groups maintain coherent behavior when conditions exceed their prediction capacity.

Given all of this, here are three ways we can adapt these principles to help our bookies out, and make prioritization easier for our teams.

You don't have to weigh options you don't believe exist, and you don't believe in options that your environment makes genuinely unavailable. No brownies in the building means the bookie never runs the brownie calculation. In complexity terms, structural boundaries reduce the possibility space the system has to navigate. Fewer live options means less computation, faster response, and behavior more consistent with intent.

This is why structure beats individual competency for organizational performance. Explicit decision rules, transparent information flows, and predictable meeting schedules remove wasteful guesswork from the queue. The bookie never calculates the odds for a race no one's running.

Skills matter. But skills always fight the menu. Structure your environment to shrink the menu.

In complex adaptive systems, attractors are persistent pulls that orient a system's movement across varying conditions. Think of the flame to which the moth is drawn. The moth doesn't calculate a route, and a breeze can knock it off course, but the pull keeps reorienting its movement. So while the path may vary, the direction doesn't.

Of course, for the moth, 🔥=☠️. Which is a useful reminder that many of the attractors guiding our course are inherited; they’re set by our prior experiences of reward. My bookie has no prior experience of the thrill I’ll surely feel when I master pull-ups, but brownies! Oh yeah, I know all about good brownies.

Today, the things that signaled job success in the past (climbing the management ladder, building a software product from scratch) exert a powerful pull even when the conditions that made them valuable have changed. You may not be directionless, but you could be orienting toward a flame that no longer signals good things.

This is why naming your higher good explicitly matters, and why personal coaches will say that:
"You need to visualize it. Why do you really want it? What’s that deeper story - do you truly want to write a book, or is there something that you think the book will get you? What will that feel like? Who do you have to be for that to happen? You are that person. Maybe you weren’t before, but now you are. See it! Believe it!”

In the past, I found this advice perverse. It seemed to imply that people who fail just didn't want success hard enough, which is both cruel and nonsensical. Are you saying starving children don't believe in food hard enough? Heck with that.

But by dismissing the idea as a motivational slogan, I was missing the mechanism underneath.

Research published in Frontiers in Psychology found that linking actions to our beliefs about who we are and what matters to us–our identity–sustains behavior change more effectively than goal-setting, because identity operates as a persistent top-down prediction in the brain rather than a target fed into the bookie's calculation. A runner who has genuinely internalized "I'm someone who runs" isn't fighting the calculation every morning. They've changed the prediction model that generates the calculation. Goals are targets you feed into the bookie. Identity is a prior you install in the DMN. One operates inside the calculation. The other changes the calculator.

At work, this might look like Andersen Construction’s higher good. They seek to be their clients’ “Builder of Choice,” and prioritize actions that make their clients happy to call them again and again. The leadership team at Valdez Ports & Harbors believes they are “Positive Stewards” for everyone working with and for the city. They prioritize actions that ensure people are comfortable, respected, and welcomed at work. At Zingermans, they draft a “Vivid Vision” every few years that tells the story of who they are living into. They prioritize actions that make the story come true.

Your higher good — the value or outcome you're ultimately protecting — works at both the personal and organizational level. Name it clearly, make it concrete, and it stops competing with other preferences. It becomes the attractor: the criterion by which options are admitted to the menu or not.

In complex systems, agents don't navigate by running full optimization calculations at every decision point. They follow simple rules that produce coherent behavior without requiring perfect information or central coordination. The classic example is murmuration, where thousands of starlings move in fluid, coordinated waves by following three rules: maintain proximity, avoid collision, and match velocity. No bird leads, and none sees the flock.

Navigational heuristics work the same way for teams. Once you have an attractor and environmental boundaries in place, you can build the pre-decisions that mean the bookies run fewer real-time calculations. Some examples:

These pre-decisions make choices easier. The rules are simple enough to apply without deliberation, and robust enough to hold when conditions change. Once the higher good is clear, establish heuristics to keep the bookie from reopening settled questions every time something shinier arrives.

My mug is partly right. Knowing what you really want really does matter.

But it's neurologically naive to think we can rely on "remembering" to keep our priorities straight. Once you understand the mechanisms — how the bookie builds its odds, where the priors come from, what shifts the calculation — you can tilt those odds in your favor.

Structural boundaries shrink the menu before the bookie opens for business. A vivid higher good installs a persistent pull that keeps reorienting what the bookie treats as worth betting on. Navigational heuristics mean the bookie never has to run a full calculation on questions you've already settled. All give your bookie better conditions to work in, whether you're trying to master pull-ups or figure out what your team should be building this quarter.

Now, the more challenging implication. Every prioritization framework your organization uses (the impact/effort matrix, the cost/benefit analysis, the urgency/importance quadrant) is formalized bookie logic. These tools emerged as ways to simplify and crystallize the bookie's past outputs into repeatable processes. Which means they have exactly the same failure point: when the environment shifts faster than the priors update, the frameworks produce confident, coherent, wrong answers. Even the best tools fail when fed outdated data.

I’m not suggesting you abandon the tools, but I do think you need to understand what they can and can’t tell you right now. In a stable environment, your impact/effort matrix gives you informed answers. In a turbulent one, it’s a starting point for a conversation about whether the assumptions behind the numbers still hold. The teams navigating well right now hold their priority lists loosely and trust the attractor more than the matrix.

The bookie evolved for complexity. Give it the right conditions, and it's very good at its job. We just shouldn’t expect it to pick the winning strategy when it’s betting on a race it's never seen before.

Quick note: if you’ve followed us for a while, you may remember that we’ve shared 2 of 3 promised articles on Conversations We’re Not Having. Part 3 is delayed because I had more conversations, which made me realize the challenge was both harder and weirder than I believed. I’ll share those stories in article 3, but more noodling is needed.