This was clarifying for me. Some other techniques that fall under this umbrella: imagining what some particular person would say about your situation; rubber-duck debugging.
I have the sense that this kind of trick is very common, but I hadn't noticed before that it's basically the same thing as prompt engineering.
First I definitely agree that prompt is a powerfully simple technique to start my brain. I just follow its direction and almost automatically creates my own answers, reflecting upon past accumulated experiences or discovering new perspective I haven't considered before. Sometimes I experience a deep, focused state which is distinguished from my usual time: A prompt triggered my brain to answer and continue answer. My brain was heated as I wrote more, and if I stop, I feel pressure to continue the excited state of writing and thinking. When I read CFAR handbook, there was intuitive side of brain called System 1. I think accessing to system 1 is what is happening above.
This is the technique I have in mind to escape my bad equilibrium when I can't start work. However, I haven't been effective yet. In reflection, these prompts and my action being done are not necessarily cause and effect relationship. They can be an association, just happening at the same time without interaction. (Which is just a thought from a moment ago. I don't have any backup evidences).
Maybe prompt is helpful because it contains viewpoint different from me. But not every sentence is prompt, and not every prompt triggers my actions.
The nature of prompt is filling-in-blank, which is known for killing creativity. Wait, then how could prompting can be a method to pull my creativity? I thought about this for a min, and my conclusion is that filling-in-blank sets a rigid answer when it is not supposed have only one answer, and prompting does exactly opposite, allowing to have any answers at very particular question. Then it made sense to me, because I believe creativity appears when there is a opening context but not closing answers.
Lastly, I went meta and thought about how am I writing this comment, indirectly prompted by your posting. I wanted to make any comment because of the new good heart tokens. I tried other articles first but I had no idea what to tell them. Then I read your article, I had thought about prompting before, and now I am telling my story. I couldn't stop writing so I definitely had a system 1 moment. Before I open lesswrong.com, I wrote some responses and an email to ask a question. Before I open gmail, I was writing a comment to tell how I started watching one youtuber because he asked us to ask him anything for prize. I quite liked my comment. Right now, I am starting to feel tired even though system didn't require focus; it was a focus. And you may tell, I feel incoherent of this comment. Wait incoherency???! It prompted me something
Incoherency always appears when I am in system 1, and will be if I use prompting. It is a big obstacle but could be solved by editing and organizing. Sorry, here I won't. But I am glad I finally explicitly realized my system 1 is usually incoherent.
Of the three examples, only "rationality" came to mind : (
I'm ESL so that sorta ruins certain idiom based examples for me.
A stitch in time saves nine. As in, if you use a stitch to fix a small tear in your shirt now, you won't have to use more stitches to fix a bigger tear later.
and
The land of the free and the home of the brave. Last line of the US National Anthem.
Summary: You can prompt your own brain, just as you would GPT-3. Sometimes this trick works surprisingly well for finding inspiration or solving confusing problems.
Language Prediction
GPT-3 is a language prediction model that takes linguistic inputs and predicts what comes next based on what it learned from training data. Your brain also responds to prompts, and often does so in a way that (on the surface) resembles GPT-3. Consider the following sequences of words:
A stitch in time saves _____.
The land of the free and the home of the _____.
Harry Potter and the Methods of _____.
If you read these one at a time, you’ll likely find that the last word automatically appears in your mind without any voluntary effort. Your language prediction process operates unconsciously and sends the most likely prediction to your conscious awareness.
But unlike GPT-3, your brain takes many different types of input, and makes many different types of predictions. When we listen to music, watch movies, drive cars, buy stocks, or publish blog posts, we have an intuitive prediction of what will likely come next. If we didn’t, we would be constantly surprised.
You can take advantage of this knowledge by prompting your brain and causing it to activate the relevant mental processes for whatever you're trying to do.
Writer’s Block
When you stare at a blank page and struggle to find inspiration, the easy explanation is that you have no prompt. The brain has nothing to predict. That’s why one of the most common solutions to writer’s block is to put something, anything down on the page. Now the brain has a prompt!
If you’re writing fiction, you can start with a template like “my protagonist lives in _____ and wants to _____”. If you’re writing nonfiction, you can use an information-based template like “more people should know about _____” or “I wish I knew about _____ sooner”.
The need for creative ideas might seem obvious to your conscious mind, but often the rest of your brain just doesn’t know what you want it to do. The more context you provide, the more likely your brain is to give you a good answer.
Babble
The babble method (generate lots of output with low standards) works in part because it gives the brain a prompt for generating ideas.
A babble prompt might look like this:
50 ways to earn more money next year:
Before writing down the prompt, the brain isn’t sure it’s supposed to be generating these kinds of ideas. Now it has a clear task, and a format in which to generate the output.
Problem Solving
If you’re trying to solve a confusing or difficult problem, sometimes you’re just stuck in a loop of unproductive thinking.
There are many prompts you can use in this situation:
The obvious thing to do here would be to _____.
If somebody smart were in my situation, they would _____.
With this kind of problem, the trick is usually to _____.
I always get stuck here, and the solution always turns out to be _____.
These kinds of prompts get your brain to think in a very particular way, and it can help disrupt the cycle of thinking that got you stuck.
Your brain is not the same as GPT-3, but this kind of prompting works because you also have predictive processes that react to input. We often think of ourselves as abstract entities with total control over our cognition, but we’re really more like collections of mental faculties that operate in certain ways given certain inputs. These tricks might be somewhat embarrassing to use in that sense, but they can be very effective.