Placeholder Image

Subtitles section Play video

  • In a world that is growing in distraction, the ability to focus is becoming increasingly

  • rare.

  • It's a skill that, simultaneously, is becoming increasingly valuable.

  • Its demand is rising while its supply is decreasing, to put it in economic terms.

  • In this essay, we'll establish a philosophy for focusing and learn how we can get better

  • at it.

  • But, before we discuss the tactical advice, let's construct a thought experiment that

  • will let us get to the heart of what it actually means to focus.

  • Imagine that our company - Robo Inc. - has built an extremely complex AI that works in

  • one of our company warehouses.

  • Let's call it Robo3000.

  • Robo3000 has been programmed to insert a key into a keyhole, turn it at a specific number

  • of rotations per minute (RPMs), and produce as many boxes as it can.

  • Because our AI is extremely complex and expensive, it's been programmed to sense any threats

  • in its surroundings.

  • A fire in the building, for example.

  • In the event of an emergency, the AI should leave the building safely so that we can avoid

  • the cost of replacing it.

  • However, the Robo3000 must also communicate with the people in the building.

  • It has to help them with lifting heavy objects, or in an emergency.

  • We hope that - in the case of a fire - it will be able to accurately gauge threat levels,

  • save as many people as possible, and save itself.

  • Our company has also invested in an even more expensive and complex AI: the Robo5000.

  • It can update all of our Robo3000's and make them more efficient.

  • Because Robo5000's are so expensive, we only have one.

  • Furthermore, it can only update one robot at a time and that process takes several hours.

  • So, it travels around the building and our Robo3000's have to interact with the Robo5000

  • to help maximize their productivity.

  • When a Robo3000 goes to interact with the Robo5000, it leaves its station and doesn't

  • produce any boxes.

  • So, it must be careful not to leave its station without good reason.

  • The Robo3000's have to periodically communicate with the Robo5000 to see if an update is worth

  • getting; they have to determine whether the benefits of an update outweigh the loss in

  • production.

  • At this point in the thought experiment, I would like to lay out some definitions.

  • Consider the action of the AI inserting its key into the keyhole, producing boxes, and

  • ignoring all of the environmental stimuli.

  • Let's call this a state of directed focus.

  • Directed Focus: Directing attention at a single thought or action.

  • A narrowed attention.

  • Providing an undivided attention while ignoring environmental stimuli.

  • The opposite state will be called generalized focus.

  • Generalized Focus: Broadly distributed attention.

  • Reacting to environmental stimulus.

  • In a state of generalized focus, our AI will do things such as analyzing if any one needs

  • serious help, if there are any imminent threats, and communicate with the Robo5000 to see if

  • an update is worth getting.

  • Clearly, maximizing the amount of time spent directing focus towards the production of

  • boxes would produce the maximum amount.

  • However, if the Robo3000 stays in a state of directed focus too long, it might not catch

  • what's going on in the environment.

  • This could lead to a disaster such as getting trapped in a burning building.

  • Clearly, there's a dilemma here; let's explore this problem further.

  • How can our AI maximize its production of boxes while also reacting appropriately to

  • its environment?

  • How does it decide how much time to spend in a state of directed focus producing boxes

  • and how much time to spend in a state of generalized focus analyzing its environment?

  • How can the AI separate a fire from someone cooking in the kitchen?

  • How can the AI separate a trivial request from an important request?

  • How can it run optimally?

  • As the coders, we have to decide what actions the Robo3000 prioritizes; there has to be

  • a system for operating and a hierarchy of priorities.

  • Here's an example of an operating system: for every 30 seconds the Robo3000 spends in

  • a state of directed focus, it must spend 5 seconds in a state of generalized focus.

  • During this period of generalized focus, the Robo3000 has to prioritize the stimuli it

  • detects in its environment and act accordingly.

  • Now, how will it do this?

  • What if we assigned it a point system based on actions & priorities.

  • For example:

  • 1.

  • Helping others with menial tasks - low to no priority - 20 points

  • 2.

  • Maximizing box production - medium priority - 40 points

  • 3.

  • Helping others with heavy lifting - high priority - 60 points

  • 4.

  • Saving its own life - higher priority - 80 points

  • 5.

  • Saving the lives of others - highest priority - 100 points

  • Now, if we code the Robo3000 to maximize the amount of points it achieves each day, it

  • should direct its focus appropriately.

  • It will maximize the production of boxes while also accomplishing more important tasks.

  • If the robot does not maximize its points, that means it has been inaccurate in its ability

  • to prioritize actions; there is a fault in its code.

  • How does this relate to us?

  • To change our ability to focus, we must - like the AI - change our code; we have to optimize

  • our action-priority point system.

  • However, I would like to borrow some terminology from the philosophy of hedonism - which we

  • talked about in the virtual reality video - for our condition.

  • Instead of maximizing points, you can think of humans as maximizing net pleasure.

  • In this case, pleasure refers to any state that we would enjoy being in.

  • The Robo3000 uses focus as a tool to maximize the amount of points it achieves over its

  • life.

  • Likewise, focus is a tool we use to maximize our pleasure over an entire life.

  • Like the robot, we do this through an action-priority system.

  • You would be correct in thinking that there is, however, an asymmetry between humans and

  • AI.

  • For the AI, the actions that would produce the most pleasure were very clear.

  • The AI knew exactly what it had to do to maximize its pleasure.

  • For us, this is not so simple.

  • Determining the net pleasure of an activity is not always intuitive, and it's highly

  • subjective.

  • Furthermore, as coders we could easily change how the robot operates and what it prioritizes

  • by altering its code.

  • As humans, some of our code is decided by nature or genetics, and what's left typically

  • has to be changed by ourselves.

  • We will explore how we might do this in a bit.

  • But first, a summary.

  • At it's core, this is what the AI Thought Experiment is about: a robot is trying to

  • maximize its pleasure in a very complex environment by appropriately directing its focus.

  • How can it accomplish that?

  • I claim that it can accomplish this through an accurate prioritization and understanding

  • of its actions.

  • Likewise, I argue that the same rule applies for us.

  • Before we discuss how to improve our focus, let's take a look at some reasons why we

  • can't focus.

  • There is another definition to discuss here: what does it actually mean to not focus?

  • Not Focusing: There is a difference between where an individual wants to direct their

  • attention and where it's actually being directed.

  • People direct their focus all of the time but it's not always towards what they want.

  • Let's discuss some reasons why.

  • Some people will find it difficult to direct their focus because they lead a high stress

  • lifestyle.

  • This is equivalent to the robot being in a burning building.

  • It will enter into a state of generalized focus, determine an imminent threat, and begin

  • saving people or evacuating.

  • It could not focus on creating more boxes amidst the chaos.

  • It would be more pleasurable for it to save the people and itself.

  • That's where it must direct its focus; that's the higher priority.

  • I will argue that many of us are programmed in the same way.

  • We lose the ability to direct our focus to what we want - homework, for example - when

  • our stress levels are very high.

  • In other words, high stress situations command our attention - like a black hole.

  • They are high priority events and rightfully so.

  • I think this is important to realize because some people need to deal with their high levels

  • of stress before they can focus on other tasks.

  • Furthermore, I hope that it will encourage some compassion for those who are suffering.

  • Sometimes, the students who perform the best are the ones that are the most stress-free

  • and not the ones that are the most intelligent.

  • I believe it is also in our interest to help alleviate the suffering of the less fortunate

  • so that they may focus on non-emergencies, improve their situation, and contribute positively

  • to society.

  • Directing our focus for extended periods of time is much easier when we enjoy the task

  • at hand.

  • Think about things you may enjoy doing such as: watching a movie at the theatre, playing

  • video games, having a good conversation, making music, doing arts and crafts, or reading.

  • Do you ever think about how easy it is to maintain a state of directed focus when doing

  • these activities?

  • This is likely due to the fact that we find these activities intrinsically pleasurable.

  • It's not difficult to determine that a good conversation with a friend will be enjoyable

  • or that playing a good game will be pleasurable.

  • We focus because we know that we will enter a pleasurable state; there is little to no

  • uncertainty.

  • I would now like to make a distinction between 2 kinds of pleasure: extrinsic and intrinsic.

  • Intrinsic: pleasure derived from an activity itself.

  • Eg - a good conversation.

  • Extrinsic: doing an activity that leads to pleasure.

  • Eg - working a job you hate for good money.

  • Activities that are extrinsically pleasurable are harder to focus on because the activity

  • itself is not pleasurable.

  • Not only do they not produce pleasurable states, it's often hard to determine whether they

  • will lead to future states of pleasure.

  • For example, let's say I work a job I hate and find it unpleasurable.

  • I also don't know if I'm going to get a raise, if that raise will bring me anymore

  • pleasure, if I'll get a promotion, or if that promotion will bring me anymore pleasure.

  • I originally took this job because it gave me the extrinsic pleasure of money.

  • Now, I'm more uncertain about the pleasure it will continue to produce for me.

  • This job will become very hard for me to focus on and, instead, I will begin to direct my

  • focus on activities that I know will produce pleasure such as