Pathfinding: how to navigate open-ended possibilities

Posted by


Our group currently faces a large possibility space.  We’re excited by the completion of the Exploratorium project. There’s an awesome network of people, and interesting projects that could come next. Before jumping ahead, I want to explore (and get feedback on) a few general strategies that a group can take to navigate a large possibility space. This post is less about our group in particular, and more about general strategy for making decisions. Later, we’ll discuss how to apply this approach to the cognitive technology field.


The early beginnings of most creative activities face a large possibility space, with many open-ended options. You might be looking for a good career, choosing a college major, designing a product for a start-up, or planning a new research project. You may also have very complex goals such trying to ensure humanity survives the next century, increasing the amount of information that can flow in and out of a brain, or being able to interpret the subtleties of the human genome.

There are many possible sub-goals, design decisions, and activities that could be explored, perhaps within some larger constraints.

When faced with many possibilities, a common failure mode is decision paralysis. Actual progress grinds to a halt. Mental resources are spent on “waiting to find more information” and “waiting for the pieces to crystallize in my head a bit more.”

These are important and necessary. The next step is to intentionally push forward the collection of information and crystallization of pieces in your head.

Eventually, you aim to reduce the space to more discrete options, where you are presented with path A, B, or C and can compare values between them. Once the problem is formalized, you can apply decision theory to decide between the options. In the large possibility space, we’re still defining what the options are.

I’ll start by looking at three (of many possible) general actions:

 1) Develop a better understanding of the possibility space.
  • A better definition of the game could come from:
    • more information about the opportunities and constraints that you face
    • a better understanding of your own possible moves and actions
2) Build up resources that will still be useful in most relevant future projects.
  • Resources could include social or technical skills, a social network, money, a portfolio of completed projects, etc.
3) Increase your intelligence and ability to solve problems. 
  • One way to approach this is to improve your tools/algorithms, and increase your speed/experience at using your set of tools/algorithms.

Pathfinding - Actions to Take When Approaching a Large Possibility Space - 2016Click the image to expand to full size. You can, share, and modify the original XMind file for this flowchart here

Chipping Away At Uncertainty

Question generation is especially interesting, so let’s explore that for a minute.

The general goal is to find questions that will give shape to the space of possibilities, and cast light on the constraints. The mental motion is: “Come up with a list of fairly specific questions, where if I knew the answer to those questions, I would have a clearer idea of what a better strategy would be.”

In other words, notice where there is uncertainty about your task or the task space. Come up with a list of specific questions that, when answered correctly, will make it better understood.

Example Questions:

  • What is the best possible outcome – what would be a “home run”?
  • If you accomplished nothing else, what is the most important outcome at the end of this project?
  • Are there metrics you can use that approximate the success of the project?
  • What is the worst possible thing that can happen?
  • What are the constraints? What directions can we definitely rule out?
  • What are the fundamental units of the system? What are they made out of? What is the lowest level at which we can work?
  • What activities would multiply / enable / build a springboard for projects in the future?
  • Has someone tried approaching this kind of space before? Are there concrete examples of success or failure? (See reference class forecasting).
  • Are there well-defined processes that could automate part of the plan?
  • What hasn’t been done before, that ought to be?

– If you’re looking for more concrete examples, I wrote two more examples of using this method here as a followup note.

– What else would you add to the list of questions? What other actions would you take in an open possibility space? I would love your feedback: email stephenfrey5 at gmail dot com.

Looking ahead

This was a first rough pass at describing the general approach to large possibility spaces. Specifically, I wanted to introduce the method of “generating and answering questions.”

(I’ll aim to provide more concrete examples in a followup post. This page may be updated once I have more feedback and information. If you know anything about decision theory and want to email me your thoughts, that would be lovely.)

Future posts will use method a tool to explore future actions the cognitive technology field.

Thanks to Madeeha Ghori for giving helpful feedback on this post. 


Cognitive Technology: a beginning and a question

Posted by

What is Cognitive Technology?
The current stated goal of the Cognitive Technology group is to “create tools to understand, extend, and improve human cognition.” This statement is quite broad (intentionally so). This post will discuss what this means so far, and how we can further refine the target.

The word “cognitive” represents an interest in a range of technologies, from low-level monitoring and stimulation of neural circuits, to higher-level interfaces such as virtual reality. As the technology becomes more powerful, I anticipate these levels will become more connected to each other, and I want to start a design conversation about how to pursue that in safe and extremely positive ways.

Cognitive Technology could be described as an extension of Cognitive Science. It asks a question: “How can we apply the knowledge from cognitive science to create tools that help improve mental ability?” Many people are working on different parts of this question, but we aim to intentionally pursue interdisciplinary research and development.

Understand, Extend, and Improve
How does the brain work, and how can we improve it?

We can create tools to understand the brain. Brain technologies are like telescopes that give us a window into the brain. As we create better telescopes, we will get a clearer picture of how the brain produces thoughts and feelings. In turn, more understanding will give us more ideas of areas we could extend and improve the brain.

We can create tools to improve the brain – to become better at problem solving, more focused, aware of cognitive biases, empathetic, and creative. The knowledge gained in the cognitive sciences and neuroscience can be applied to intentionally improve the way our brains work.

We can create tools to extend the brain. Through informational tools, haptic devices, immersive environments, robotics, and other actuators, we can amplify the amount of information coming out of the brain and use it to increase our abilities.

What’s Next?
The above writing expresses my current view of an “applied” approach to brain technology. In the future, I hope to get more information, find critical flaws in the way I think about this, and adjust course many times. Please, make suggestions or point out holes in these viewpoints – your feedback is incredibly useful.

In the next post, I will discuss (and look for feedback on) a general approach for taking actions in a large possibility space. Then we can think about how to apply this approach to future group activities in brain technology.

– Stephen