Deep Thinking
Why Learn to Pay Attention?
Deep Thinking and Warm-up Time
How right can PG be! It is only after a few hours of working on an essay or program that you start getting the really good ideas. Just doing it for an hour or so doesn’t seem to be enough. I would have missed the above insights had I just gone with my original plans and not pushed through my resistance.
To do something Hard, Give it your Full Attention
Apartment search lessons: It’s like running a startup (though I haven’t done that): you have to give it your undivided attention. I was hoping to do it on the side while the others focussed on it for a day or two and got the job done beautifully. Man, people have a lot of information about apartment search. This guy has a list of 7 property sites and links to the FB groups for Indian grad students at Purdue.
How to Pay Attention
Definition: Notice Diffs
Hypothesis: One way to “pay attention” is to get only pertinent exemplars.
If you get exemplars in close succession, you can see the diffs in their inputs and outputs and induce relevant factors for the category hierarchy. You will also be able to form property associations for that category.
If you wait too long between exemplars, say by moving on to exemplars for some other category hierarchy, you won’t be able to see the diffs and won’t induce the relevant factors. You won’t form property associations either.
In general, you need to remember the previous exemplar so that you can notice the diff.
Test: Positive exemplar - was thinking about this hypothesis nonstop for the last fifteen minutes. Got rid of the hypothesis that it was about “the rate of associations per unit time” (though that is relevant). The key idea is that you should not get distracted by irrelevant exemplars, that you should not be “suckers for irrelevance”. – was paying attention.
Test: Negative exemplar - was thinking about this hypothesis. Then started thinking about a Hangouts call I need to make. Lost my train of thought. – Wasn’t able to induce or eliminate any factor.
Hypothesis: Another way to “pay attention” is to remember past exemplars so that you notice diffs.
Test: Sherlock - “you’re wearing lipstick” – remembered the past exemplar and noticed the diff.
Corollary: One way to measure the “depth” of your attention is to count the number of factors you accept or eliminate per minute. If you’re inducing only a few factors, then that means you’re not seeing exemplars at a high rate.
Necessary Condition: Lack of Network Tools
Hypothesis: Network tools can be a relevant factor for distraction. I think this is because they are usually more fun (or more reinforcing) than work.
Test: [2019-01-15 Tue] If the internet is on, I spend hours on YouTube or Reddit. However, when it’s off, I don’t.
Test: [2019-01-15 Tue] It’s not always a relevant factor, because when I was focused on my compilers project, I didn’t surf the web even though it was there. Coding was more fun.
Test: Negative exemplar - the weather app is a network tool, but I don’t keep refreshing it because it’s not all that reinforcing.
Necessary condition: Lack of Distraction
I think a lack of distractions is key to deep focus. Hotel lobbies and airports have this quality. So do libraries and study rooms. On the other hand, the simplest of things can cause distractions. Like your internet-enabled smartphone, tablet, laptop, or even your music player.
You must have nothing else whatsoever to focus on but your subject matter. Nothing to take your mind off.
Unexpected places have these qualities. And other places that you would expect to have these qualities don’t. For example, some of my deepest moments of focus have come on aeroplanes. Unless you give in to the crappy movie player in front of you or dive into your own smartphone movie player, there is literally no distraction.
Obstacle: As Time flies, more Distractions creep in
Here’s a rough hypothesis: The more time we spend in an environment, the more distractions we manage to bring in.
Consider the first time you move to a new house or office. Initially, nothing is set up. You probably don’t have wifi working, or a TV, or anything. You don’t have comfortable furniture around, most probably. So, you have no choice but to focus deeply on whatever you’re doing. But, over time, you figure out various ways of distracting yourself - what the nearby movie theaters are, what restaurants deliver food at your home, what you can do in your relaxation time, etc.
Soon, your place abounds with distractions. You can’t get anything done.
To Pay Attention, Categorize something from Scratch
Hypothesis: You pay full attention when you have to actually categorize something.
For example, they tell you to pay attention to your mistakes when you’re practicing. What does that mean? I made a mistake when I typed “washingon monument” instead of “washington monument”, but I didn’t realize it. It was only after I’d ruled out a problem with the dictionary program that I started to look at the words letter-by-letter and found the missing ‘t’. In other words, you can’t assume that you’ve done it right when you feel you’ve done it right. You can’t trust the labels given. You have to start categorizing from the bottom.
Losing Focus: Unable to See Diffs
Hypothesis: If you can distinguish between instances and induce relevant factors, you will consider that “paying attention”.
If you can’t distinguish between instances, you will consider that “losing focus”.
Test: “Pay attention” to something - can keep forming associations. However, at some point, you “lose focus” and can no longer form any associations. – It all starts to seem the same, like the third section of a new research paper I’m reading.
Limits of Attention: Shallows In, Shallows Out
Hypothesis: If you have been taking in only “shallow” material, you will lose focus very soon because you will not be able to catch subtle differences between instances and thus the material will start seeming the same.
If you have been taking in “deep” material, you will be able to keep focus and build long category hierarchies because you will be able to catch subtle differences between instances.
Maybe the difference between “shallow” and “deep” material is that you get highly distinct instances in the first case, with very little context, so that you can easily differentiate them, whereas you get subtly-different instances in the second case with a lot of shared context. Basically, in the latter, you will be lost unless you keep the context in mind. In the former, you can get by with just a little context.
Test: [2018-10-10 Wed] Reading a novel after a long time - keep checking how many pages I’ve read. Don’t feel like continuing for a long period. Can’t imagine reading all the way to the end, which seems so very far off. There is a lot of difference between scenes. But I suspect that the magnitude of the difference depends on your “imagination”. When would you have a large difference and when a small difference?
Suppose you introduced a bunch of characters and said a bunch of stuff about them. I wouldn’t be able to “follow along” because I wouldn’t be able to store “he’s a retired soldier” on a particular name like “Ronald” or “Moran”. I’d be forced to drop that piece of information for the lack of a unique cubbyhole to put it in. Basically, if you want to save the output factor “retired soldier”, you’ll need to diff the input with existing inputs for “retired soldier” like John Watson or Si. But if you don’t know the person’s name or other characteristics, you can’t diff him and thus can’t store his information.
So, to insert this new piece of information in its proper place, I need to know the other input factors for that person. I need to know the “context”.
I suspect that the “hooks” or cues on which you associate new pieces of information are all created in your head. The words on the page are usually not distinctive enough to give you a unique cue (unless maybe it’s poetry; maybe that’s why they say that it paints a picture in your mind - it gives you a unique cue). You have to keep your mental model in mind at all times. For example, this is the part where he talks about the rents of coffee shops in railway stations, as part of a chapter on scarcity power. That was a pretty distinctive example, actually. Easy to remember.
Sections about “response competition” or “cue-dependent forgetting” are much harder to remember. Unless you have the context of “overall chapter on forgetting, where we’ve already seen decay theory, consolidation theory, and unlearning”, you will be screwed. You won’t be able to place “cue-dependent forgetting” as an alternative to the above “trace-dependent theories of forgetting” and you won’t be able to figure out its unique cues.
In short, books are harder to learn from because you have to keep the structure of the book in your mind so that you can insert new ideas appropriately.
Test: [2018-10-10 Wed] Watching AOE YouTube video - could watch it forever. Little context (player A is doing pretty well). You can clearly tell the difference between scenes over time because time passes, people advance through the ages, they build greater economies, they fight, and they ally.
I suspect that the “hooks” or cues on which you associate new pieces of information are formed uniquely by the scene in front of your eyes. For example, this is the “Baahubali jumps off a cliff scene”.
How to Spend your Time Well
Cut the Umbilical Cord: Turn off the Internet
Principle: When you’re working, turn off the internet.
As Cal Newport recommends, schedule your distractions for specific periods. Apart from that, just turn it off.
Weaken Strong Habits by Changing the Environment
My aim is to weaken habits that are undesirable but strong and strengthen ones that are desirable but weak.
Which habits are strong right now? Watching movies and surfing online (that’s how I spent my day). Which ones are weak? Practicing skills and summarizing my old notes.
A change of environment would help. Reading books on my tablet would weaken my habit of sitting in front of my laptop for hours on end watching movies or videos. Similarly, practicing with pen and paper would help stave off on-screen distractions. Turning off the internet during my work hours would certainly work too.
(Technically, I’m not weakening the habit, just removing the cue that triggers it. I end up doing it less, which is what matters.)
Comparative Advantage: Calculate the Per-Hour Value of a Task
When I started my business I thought it would be amazing if I eventually earned $100 per hour working on it. (This is a princely sum for 20-something programmers in Japan.) Its funny, people generally get very good with practical experiences of the mathematical properties of averages in school and then totally forget about that experience in business. If you want a 92 average, you’d better not routinely get 60s on your homework. If you want to earn $100 an hour, you’d better not busy yourself with $5 an hour tasks.
Question: Am I getting $50 worth of value every hour out of this?
Schedule Deep Work and Punish Dawdling
Experiment: Set aside 6pm-9pm every day for deep work. Don’t surf or distract yourself in any way.
Hypothesis: Based on Cal Newport’s Deep Work, this should help me get into the habit of doing substantial amounts of deep work every day.
Measurement: Number of hours of deep work, i.e., zero distraction and peak cognitive capacity.
Consequential Deadlines make you spend Time
Observation: I’ve had severe deadlines for the past two weeks. I couldn’t afford to take my eye off the game for any period of time (to indulge in either fantasies or YouTube binges). Guess what? I did a ton of deep work. Respected the LeechBlock limits on my web surfing. Actually read books after a long time. Life’s good.
Observation: Contrast that to all the times where I did the exact opposite and couldn’t muster the will to even close a YouTube video.
Hypothesis: Deadlines cause motivation (due to low delay) and thus automatically eliminate shallow work, which comes about because low-value actions are more tempting.
Corollary: The deadlines need to be for high-value tasks; otherwise, you’ll have high motivation for nonsense like some random bureaucratic forms you need to fill.
Lesson: So, if you’re bitching and moaning about not being able to do deep work, ask yourself if there’s any valuable deadline you’re trying to meet.
Don’t give YouTube control of your Time
When you open Youtube or Twitter or Reddit for whatever purpose, you get a ton of incidental information that grabs your attention. An enticing video, a “thought-provoking” tweet from a famous person, or a bunch of new posts with a lot of votes. Those stimuli impose themselves on your mind. You get side-tracked from your goal.
You should be in charge of what thoughts come into your mind. If you really want to get more ideas from some luminary, you should be in control - you should read a book written by him or maybe watch a talk he gave. The only thoughts that enter your mind should be those explicitly approved by you, to the extent possible.
For example, on a news website, my eye is automatically drawn to the “Top Stories” sidebar - there’s some scare story about a terrorist, a review of the upcoming Batman vs Superman movie, a video about Virat Kohli - how can I not click any of them or at least not think about them?
It’s the same thing at the end of the news article. There’s an “Also read” column at the bottom showing intriguing pictures - some guy threw a slipper at a politician, some female celebrity was playing Holi, and so on.
Deep Work Environments: Paying Attention for Long Periods
Nothing like an Airport for Deep Work
Observation: Thought deeply for hours on end - on the shuttle ride, in the airport, and on the plane itself.
Hypothesis: Nothing like it.
Observation: Zero distractions. Nowhere else to be. Nobody to distract you. Unfamiliar environment, so I don’t revert to my usual web-surfing habits.
Hotel Room
Observation: Things are neatly organized in a hotel room. There’s no residue from your past work over there. It’s all been reset to the default. Contrast that to your own house. There are the loose sheets from a previous assignment, clothes that need to be laundered, books lying here and there, an old lamp you’ve been meaning to fix for the last six months, and so on.
Observation: You don’t have to worry about anything else.
Hypothesis: Wow. It’s a strict interface. All you can do is your deep work (and, of course, other enjoyable stuff).
Notes
Steve Yegge writes about the ideal conditions for deep thinking in his post Innovation 101.
The pop-neuroscience book Your Brain at Work has a great chapter on distraction.
comments powered by Disqus