Today's 14 level question.
Philosophy, in my view, is continuous not just with the sciences, which it's been throughout its history, but also with design, architecture, engineering, medicine and therapy - what you might call the professions. As you can see, I'm not talking only about the making professions, but also the healing professions. Some of these professions have emerged out of philosophy (which I will call P- from now on) and others have an independent history. The new philosophy - henceforth P+ - takes us back to the drawing board and asks us to take a fresh look at the philosophical impulse.
But it’s too much to ask to rethink philosophy from scratch. Is there a smaller, but still substantial idea-ideal-intuition that’s present in all these disciplines? I believe so: one commonality is the concept of a model, which:
In philosophy often takes the shape of a 'thought experiment.'
In physics, it could again be a thought experiment or a 'Gedanken experiment.'
In many parts of mathematics, applied mathematics in particular, and in the computational sciences, it would be explicitly called a model.
And so would it be in architecture or in design or engineering, where they might also be called 'prototypes' if they actually work.
In psychology in the form of ideal personality types such as in the enneagram.
I'm not saying that there is a unified theory of models, one that unifies modeling across history-philosophy-literature-architecture-physics-mechanical engineering-cognitive therapy. Far from it. But there’s an assemblage of models.
My vaguest of vague intuitions is that modeling should be to P+ that reasoning was to P-.
Building upon that intuition, I want to bring this more general concept of a model back to philosophy and ask what thought experiments might look like if we fleshed them out using modeling skills better known in mathematics or architecture or some other profession.
Thought Experiments Reconsidered
Thought experiments are among the oldest knives in the philosopher's drawer. They are among the easiest philosophical conundrums to remember and communicate to the lay person. We know them by name even when we haven't read them in the original - Plato's Cave is an example. The thought experiment as usually formulated is a mental exercises (Dan Dennett called them 'Intuition Pumps') typically present a carefully crafted challenge or dilemma that forces us to examine our deepest assumptions. Yet these classical thought experiments suffer from what we might call "static scenario syndrome" - they present fixed dilemmas without the dynamic complexity that characterizes real moral decision-making.
Computational tools (aka AI) offer unprecedented opportunities to revolutionize this ancient philosophical method. By transforming thought experiments from static puzzles into dynamic exploratory tools, we can create what might be called "possibility space modeling" - interactive frameworks that allow systematic exploration of moral and conceptual terrain.
Beyond the Single Dramatic Question
Traditional thought experiments typically present one dramatic scenario and ask for a binary response. The trolley problem asks whether you would pull the lever. Mary's room asks what happens when she sees color. These scenarios were brilliant for their time, but they reflect the limitations of pre-digital thinking tools. They assume perfect knowledge, present isolated decisions, and offer no mechanism for exploring variations or consequences, especially those that connect the current scenario with other related scenarios.
Technologically enhanced thought experiments can transcend these limitations through several key innovations. Dynamic parameter exploration allows users to adjust variables in real-time rather than contemplating fixed scenarios. Instead of asking "would you pull the lever?", an enhanced trolley problem might feature sliders for the certainty of outcomes, the ages of potential victims, their relationships to each other, and the probability that intervention succeeds.
This approach transforms philosophical inquiry from answering discrete questions to mapping continuous landscapes of moral intuition. Users can discover where their ethical commitments shift by sliding along different dimensions, revealing the topology of their moral reasoning - regions of clarity separated by boundaries of uncertainty, with phase transitions where small parameter changes flip intuitions entirely.
Then there's the architecture of trolley problems as a whole - what happens when one trolley problem affects another one? For that we need:
The Architecture of Moral Physics
The most sophisticated enhanced thought experiments would implement what we might call "moral physics" - systematic constraint propagation where ethical commitments in one region of possibility space create logical requirements elsewhere. When a user indicates that intention matters more than outcome in one scenario, the system can generate related scenarios that test whether this principle holds consistently across different contexts.
This creates iterative moral learning that mirrors how humans actually develop ethical frameworks - not through single dramatic moments, but through accumulated experience and reflection. Enhanced thought experiments can unfold over time, presenting initial choices, revealing consequences, introducing new information, then offering evolved scenarios that build on previous decisions.
The Trolley Problem Simulator
Having gotten this far, I did what any wannabe philosopher would do in 2025: I vibe coded a trolley problem simulator.
The application moves beyond the traditional binary choice of "pull the lever or don't" by introducing eight adjustable parameters that create a vast possibility space of moral scenarios. Users can modify the number of people on each track (1-20), adjust outcome certainty (50-100%), select different age groups and relationship types, set decision time limits, and even control the probability that their intervention will succeed. This parametric approach allows for systematic exploration of moral intuitions across thousands of potential scenarios.
Of course this won’t pass muster with a professional moral philosopher. That will take many more iterations of this app and - with a nod to experimental philosophy - the collection of Trolley Problem data from a vast number of people. However, it’s clear that such exercises can be done - the speed and ease of modeling lowers the bar for philosophers and designers and psychologists to work together on what can be called ‘moral engineering.’