Over the years I’ve come to realize that the way I operate—both in business and in life—can be visualized as a pyramid.
At the top are mental models. Beneath those sit the systems that operationalize those models. And forming the foundation are the tools that allow those systems to run efficiently and, when possible, automatically.
The pyramid matters because it enforces something simple but powerful:
Tools should never drive thinking. Thinking should drive systems, and systems should determine the tools.
Too often organizations start with tools and hope good outcomes emerge. I prefer the opposite approach.
The Top Layer: Mental Models
The top of the pyramid is the smallest but most important layer. These are the mental models that shape how I interpret problems, make decisions, and allocate effort.
I first encountered many of these ideas through Charlie Munger and then spent more than thirty years collecting, testing, and refining them through experience.
Some of the models that influence how I operate include:
-
First-principles thinking
-
Pareto optimization (80/20)
-
The entourage effect
-
Inversion
-
Compounding
-
Second- and third-order thinking
-
The Five Whys root cause analysis
-
Risk = Probability × Impact (and sometimes × Novelty, borrowing from Taleb)
-
Creating more value than I harvest
Together these form what Munger described as a latticework of mental models.
They influence everything I do—from cybersecurity architecture to business strategy to personal productivity.
Mental models are powerful because they allow you to reason from principles rather than reacting to symptoms.
But by themselves they are abstract.
Which brings us to the second layer.
The Second Layer: Systems
Mental models shape thinking.
Systems turn that thinking into repeatable behavior.
Over time I’ve developed several systems that embody the mental models above.
TaskGrid
One of the most important is a task and project management system I built called TaskGrid.
It’s based loosely on the Eisenhower Matrix, but evolved into something closer to a personal operations dashboard across the planes of my life.
Each day TaskGrid tracks three types of activity:
-
Things I must do
-
Things I should do
-
Things I want to do
The system keeps me focused on high-value tasks while also revealing patterns where urgency and importance diverge.
One unexpected benefit is psychological.
TaskGrid signals when the day is finished.
When the items on the grid are complete, my brain gets a clear signal that it’s time to stop working and return to full optionality—the freedom to explore, learn, or simply disengage.
That boundary is incredibly valuable.
AI-Driven Knowledge Distillation
Another system focuses on information analysis.
The modern information environment produces far more content than any human can realistically process. Yet buried inside that flood are small amounts of extremely valuable insight.
To deal with that, I use AI to analyze large volumes of articles, research, and news.
But the goal isn’t just summarization.
The goal is to apply models like Pareto, inversion, and second-order thinking to extract the few ideas that actually matter.
Often the most valuable insights are the ones that are uncommon, overlooked, or hidden inside noise.
AI helps surface those signals.
Risk Analysis Systems
Risk has always been central to my work in cybersecurity, but I apply the same thinking more broadly.
Over the years I’ve built systems—initially using traditional analytics and now increasingly using AI—that monitor and evaluate risk across multiple areas:
-
Information security
-
Financial decisions
-
Business operations
-
Personal life decisions
These systems analyze probability, impact, and occasionally novelty to produce actionable insights rather than just dashboards.
The goal is simple: better decisions under uncertainty.
The Foundation: Tools
At the base of the pyramid are the tools.
Tools are important, but they are also the least important layer conceptually.
They exist to support systems—not the other way around.
I primarily operate within the Apple ecosystem, using multiple devices that are often configured for specific types of work such as AI experimentation, automation, research, or communication.
One principle I try to enforce aggressively is asynchronous operation.
Optionality disappears when your time is constantly interrupted.
So I try to push as much of life and business into asynchronous workflows as possible.
That includes things like:
-
Automated scheduling and calendar management
-
Routing unscheduled calls to voicemail that becomes email
-
Automated email management that surfaces only meaningful messages
-
Time-boxing tasks, research, and projects on my calendar
In many ways, I live and die by my calendar.
Both local AI and cloud AI have also become central tools in this layer. They help automate routine work, accelerate learning, and simplify repetitive tasks.
But automation itself requires judgment.
To help decide what should and should not be automated, I rely on a framework I developed called FRICT, which I described previously on notquiterandom.com.
FRICT helps identify tasks that benefit from automation while protecting areas where human judgment still matters.
Why the Pyramid Matters
Many organizations invert this pyramid.
They start with tools, bolt on processes, and hope good decisions emerge.
But tools alone rarely create good outcomes.
Instead, I think it works better in this order:
Mental Models → Systems → Tools
Start with the models that shape how you think.
Build systems that embody those models.
Then choose tools that make those systems easier, faster, and more automated.
When the layers align, something interesting happens.
Complexity decreases.
Optionality increases.
Decisions improve.
And over time, the entire structure begins to compound.
Support My Work
Support the creation of high-impact content and research. Sponsorship opportunities are available for specific topics, whitepapers, tools, or advisory insights. Learn more or contribute here: Buy Me A Coffee
* AI tools were used as a research assistant for this content, but human moderation and writing are also included. The included images are AI-generated.
Not Quite Random 