Learn it → Apply it → Save as Skill
In a tightly coupled, nonlinear system, a small and often unremarkable change at the start can grow into a disproportionately large effect later. It is not a claim that everything is unpredictable. It is a prompt to ask where a system is sensitive enough that small inputs stop staying small.
The Dr. Wallace's Favorite Mental Models is designed to turn useful ways of thinking into reusable tools. Each model gives you a fast visual overview, deeper interactive learning, practical application exercises and an AI-ready Markdown skill you can save for future use.
Instead of merely reading about mental models, you can gradually build your own AI skills library: a collection of portable reasoning skills that can help ChatGPT, Claude, Grok and other AI assistants apply the mental models you value.
When a system has feedback loops that amplify rather than dampen, a tiny difference early on does not stay tiny. It compounds through each loop until the outcome looks nothing like what the small input would suggest.
Important: most systems are not chaotic. A small input usually produces a small, proportional result. This pattern shows up specifically in tightly coupled, nonlinear systems.
The model is most useful when feedback loops, dependencies or coupling mean a small change does not stay isolated from the rest of the system.
The method is about locating sensitivity before it becomes a crisis, not about predicting the unpredictable.
List the small, early inputs or decisions that feed into the system you are worried about.
Follow the feedback loops and couplings that could amplify, rather than dampen, a small change.
Ask whether a slightly different input changes the outcome proportionally, or disproportionately.
Add buffers, checkpoints or circuit breakers at the points where amplification is most likely.
Open a use case to see what the model changes.
Trace how a minor flag, default, or setting could interact with caching, load balancing or retries. Ask which of those couplings could turn a one-line change into a full outage.
Map the dependencies between institutions, instruments and liquidity sources. A shock that looks contained in isolation can cascade wherever leverage and tight coupling exist.
Identify which downstream tasks depend on the delayed one without slack. A short early delay can consume the whole schedule's buffer by the final milestone.
Map how a minor weakness could be chained with other small weaknesses to produce a major breach. Severity is often about the chain, not the single flaw.
A tiny routine can compound into a large change over time. This is related to, but not identical to, Compounding: here the focus is on sensitivity and feedback, not steady accumulation.
In weather and other chaotic systems, small measurement errors grow over time. This is why forecasts use ensembles of slightly varied starting conditions instead of one single run.
This is an illustrative example designed to teach the method.
Expand for the distinction between the scientific discovery and the popular metaphor.
In 1961, MIT meteorologist Edward Lorenz was rerunning a numerical weather simulation. To save time, he re-entered a mid-sequence value using a rounded, shorter decimal than the computer had stored internally. The rerun diverged completely from the original within a simulated month. Lorenz had found that some deterministic systems are so sensitive to their starting conditions that tiny, unavoidable measurement differences make long-range prediction impossible in practice, even though the equations themselves contain no randomness.
Lorenz published the underlying finding in his 1963 paper "Deterministic Nonperiodic Flow." The popular name arrived later: for a 1972 American Association for the Advancement of Science talk, the session convener, meteorologist Philip Merilees, titled Lorenz's talk "Does the Flap of a Butterfly's Wings in Brazil Set Off a Tornado in Texas?" The vivid title, not the original 1963 paper, is what made "the butterfly effect" the common name for sensitive dependence on initial conditions.
No. A chaotic system is still deterministic: the same starting point always produces the same result. What makes it chaotic is that two almost identical starting points can produce wildly different results, and no measurement is ever perfectly precise. That combination, not randomness, is what makes long-range prediction fail in these systems.
Most processes are stable and roughly proportional. Reach for this model only after checking that the system is genuinely tightly coupled and nonlinear.
"It's impossible to predict anything" is a way to avoid the harder, more useful work of finding exactly where a system is sensitive.
A line of dominoes is linear and predictable. The butterfly effect describes amplification through feedback, which is a different and less predictable mechanism.
Ensemble forecasting, stress testing and scenario planning exist precisely to manage sensitive systems. Do not treat sensitivity as a reason to stop analyzing.
Amplification depends on the structure of the system, not just the size of the initial input. The feedback loop is usually the real target for a fix.
Sensitivity typically grows with time. Near-term predictions in a chaotic system can still be reliable even when long-range predictions are not.
the system is simple and roughly linear, so a change in input produces a roughly matching change in output.
you need to quantify a likelihood rather than describe a qualitative sensitivity to initial conditions.
the effect you care about comes from steady, repeated accumulation rather than a one-time sensitive trigger.
This browser tool structures your thinking. It does not send your information anywhere or perform AI analysis.
Save as Skill converts the model into portable Markdown instructions that teach an AI when and how to apply the Butterfly Effect, what questions to ask, what evidence to seek, what mistakes to avoid and how to structure its analysis.
Save the skills you find useful and organize them into your own reusable library. The Markdown is intentionally model-neutral, so it can be used wherever an AI system accepts uploaded files, project knowledge, custom instructions or reusable text context. Platform capabilities differ, so this is a portable skill file rather than a claim of one-click installation.
No. It means precision degrades with time. Weather models stay useful for several days because the divergence from measurement error has not yet grown large. Past roughly one to two weeks, that divergence overwhelms the signal, which is why forecasters run many slightly varied simulations, called an ensemble, instead of trusting one.
It describes systems where the distance between two nearby starting points grows roughly exponentially over time, rather than staying constant or shrinking. A small, unavoidable rounding difference can therefore become a large difference given enough time steps.
No. Complexity and chaos are related but not the same thing. Some complex systems are well-damped and self-correcting. The question to ask is not "is this complex," but "does a small perturbation here grow or shrink as it moves through the system."