Hello, I have today been informed about a method for writing content that is relatively easy and painless. The idea is that by default a human will when reading predict what word is going to come next. Studies show that humans can predict what letter is going to come next and get pretty good entropy scores. That means that even the process of reading involves generating plausible next words.
And so it is possible to similarly do the same when writing. When writing a sequence of words, after the end there will be a feeling that some word will come next. It is unexpected and abrupt for the sentence to immediately end, so there is a feeling of continuation, that something will come after. One can while writing simply look at the space right to the right of the text that has already been written and there will be a feeling there about what word will come next.
Do not worry if this is difficult and slow at first. A practice exercise is to write a few words and then take some time to get a feeling of what word is going to come next. With meditation it is possible to notice experiential phenomena that are in one’s experience and sometimes have a spatial location. Well, maybe someone who knows how to find these phenomena is going to be able to find some sense of “next word” ness when focusing on the end of a sentence fragment. That will tell them what word they think is going to come next.
It is possible to simply, when noticing what word is going to come next, type that word. That is in fact what I am doing right now and what I have been doing for this entire post. It is rather easy and painless to do. It is not anxiety-provoking because it is a meditative, instruction-following exercise where I simply get a sense of what word will come next and type it. I am not responsible for the output of my text predictive model. I am not to blame. I am not unworthy.
This shortcuts various anxiety loops that are possible when writing in a more traditional manner. At that point, while there are many sentence continuations that are invisibly generated, all are filtered out because it would feel somehow bad to have written that sentence. Maybe the sentence would be cringeworthy or stylistically bad or offensive or embarrassing or I-don’t-know-what, it might just be off, and someone with enough of an aesthetic taste might simply be repulsed by the idea of actually writing out that sentence. There is no filter this way, because the sense of what word is going to come next is always present even when reading text, and so it is not specific to writing and doesn’t indicate that the person having such a predict is, themselves, responsible for that text.
I know this is a somewhat silly way of thinking about moral responsibility, but it is emotionally compelling enough that it can increase writing output by a lot. And output is not always good, quality matters too. It would be to be susceptible to Goodhart’s law to say that writing more words is always better. However, writer’s block is a pain, it is really bad if it isn’t possible to write anything at all, or to only be able to write very short sentences and paragraphs. That way it is possible to go months without making any progress on a writing project.
I prefer to have something written even if it is low-quality. That way, it is possible to edit or re-try later. That will increase the level of quality. And I know that that seems like a confident statement, that it will increase the quality, not that it might. The difference is that right now I do not care about seeming overconfident, since I am simply reporting on what word I consider most likely, and not worrying that the word I think is most likely to come next is not actually the most likely, I am not even considering alternatives (more than 1) because I am writing so fast, so the words just come out and I am not worried about the implied overconfidence.
This is nearing the end of this essay. There is a job to conclude everything. I am not sure how to do that except by rating the quality of what I have already written. I have done practically no editing so far, only spelling and that kind of thing, and maybe pressed backspace a few times to erase the last word. But there is basically no editing here. And is it high quality? I will leave that for readers to judge.
In any case, this method has produced an artifact that explains the very method and provides a way of judging the method by the very artifact produced by the method in a rather recursive loop that might ground out somehow in some kind of probabilistic or utility theoretic calculation made by someone who is deciding whether to use this method to write, or not and to instead use their previous writing method exclusively at the risk of hitting writer’s block.
The end.
[ED NOTE: This took about 6 minutes to write. I previously described this method on Twitter. I made a single edit after, to add a missing word.]
To me, the difference between the colloquial term “brainstorming” and this site’s term “babble and prune” is the intentional choice to split the activity into two phases: an unfiltered idea generation phase followed by a filtering/editing phase. Emphasis on “unfiltered”, for the anxiety-reducing and writer’s block circumventing reasons you gave.
I’d be grateful for an update down the line, if you come across any unexpected benefits/shortcomings.
Writing an article seems more difficult to me because it involves a choice on two levels—what to write, and how to write it (the outline vs the actual words). How to put these two levels together?
One option is to simply start writing, so both the outline and the actual words are generated in the same pass. You can edit some words afterwards, but you can’t really edit the outline… unless is means identifying some superfluous parts and removing them. Adding a new part would require switching to the generating mode (for that part) again. Reordering parts? Not sure if the text remains fluent.
So, what else is possible? Decide the outline first, and then generate the text with the idea that “I am trying to progress to part B” in the background? Is or isn’t this substantially different from the original version? The difference is that you are having a goal, instead of just writing and seeing where it goes. The similarity is that in the original version you still at some moment need to finish the article, which is also a kind of a goal?
Another risk is that your generator will fluently travel between the predetermined topics A, B, C, D, E, you create a lot of text, on the specified topic, but… it will somehow lack the conclusion? It will be just “a stream of text that ended at some point” rather than “a stream of text that culminated in a punchline”. Unless you maybe think about the punchline first, and then set up the A, B, C to include the prerequisites.
Just don’t change your mind about the outline in the middle of writing the text. Like, at the end of this comment I realized that this way of writing, from the perspective of the bicameral mind, is to simply shut up and keep writing what the gods are telling you. And the version with the outline is… making plans, and then praying to gods to make it happen… and then accepting their verdict, whatever it is?
Or do the change of outline on the level of articles. Like, finish the original article with the original outline, and then use your “this should have been done differently” thought to set up the outline for the next article.
There’s a known divide among two styles of writing fiction, one of which is to plan the plot out in advance, and the other is to just start writing and see where it gets you. The latter method may produce events that feel more organic, but has a known failure mode where the author may realize that the story they’ve written ended up having no clear ending and then they have to throw away the last hundred pages of text and try to take the plot in a different direction from an earlier point. (Or in the case of writers such as G.R.R. Martin, have the whole project just sputter out after the plot has just gotten bigger and bigger with no satisfying end in sight.)
This generalizes to actions.
Lots of people I know, including me, take way fewer actions than is optimal because we are trying to avoid making mistakes instead of trying to get what we want. (In some cases that’s just a cover story and we’re actually trying to avoid revealing our location or preferences, so that we’re not a target.) But in lots of contexts if you just do things, instead of trying to supervise your intentions so you only do things you’ve preapproved, you can get lots more done that you want to do.
If you’re worried that this is risky, you can think about what’s likely to go wrong and how important that is in the specific context where you’re considering to just do stuff, and plan ahead to mitigate the risks worth mitigating.
I am attempting to write a response in the same manner.
This reminds me of automatic drawing. Automatic writing is an obvious correlate. The dramatic tension of the underbelly of fleeting meeting gets drawn up into the universe and doesn’t very much write the thing itself.
A pause is suggested.
When the dramatic moment is over, there’s something left behind. It accepts a period. Differences in text suggest differences in author.
Moreover, the strategy presents itself as golden.
I suppose I will finish here.
Oooo thanks for this, just used it to write a post for my blog and it was more fun and easier than usual: My Anxiety Keeps Persisting Even Though It Has Been Diagnosed, Rude
Interesting!
Since about 2-3 weeks back, this is how I write.
Based on some kind of writer-instinct. Markov chain completion. “Babbling”. Going with the flow. I simply output words, until I reach the point where it feels right to stop.
Corrections and edits are the same ways. Look over the text and go with the feeling. Add more text where it feels incomplete, where the transitions feel abrupt. Sometimes this type of corrections redirect the text on a new trajectory, making the old version obsolete. Sometimes not.
Having never written fiction before, I wrote an 8k word story in 3-4 days (and abandonded, because it was kinda cringy). Still, I was really happy with the attempt.
There’s also Ultraspeaking, which is a set of exercises for learning to do this but for spoken speech. If you are the kind of a person who finds it difficult to make conversation, you can try to just start saying something and then let your brain “autocomplete” the rest, by essentially the same kind of process as described in this post. I definitely feel like I’ve been able to level up some of my conversational skills by practicing that process.
EDIT: This was my attempt to do what you suggested.
EDIT2: I feel like what I’m talking about is something more generally applicable. You can use it for decision making.
Another reason I like it is because I tend to have mental filters in place, that prevent me from thinking about something that seems difficult. This helps because I think I just let my self occupy the thought that feels like it would be too hard to think and trust in my intuition to lead me somewhere useful instead of trying to force things? Probably not true. Anywho, for people with mental scars like me, this kind of thing is useful. Don’t know why. But it also reminds me of what SquirrellInHell is talking about in BeWellTuned. Like you’re like a CEO or maybe someone who sends ratings to various thoughts/plans that come up, and your brain decides what to do as a result. If you just ignore your job, then parts of your brain will get fed up and start not trusting you to rate good plans as good and you become lazy, time incosistent and hate yourself. I feel BeWellTuned has a much better description than I do.
__________________________________________________________________________________________
What an interesting way to write. But I think it is more like babble than anything else. If you don’t have good babble, you’re probably not going to produce anything useful by doing this. My internal experience seems more like a series of suggestions, potential next ideas and corrections to make. On occasion, I might get a feeling that says “no, that’s not right” and have to iterate next part of the sentence a bunch. And my feelings tend to be right, so I don’t try to “forge ahead”. Honestly, I have been doing what I describe for a while when making my notes. It results in a great deal more productivity, though I don’t do it nearly as often as I should.
I was viewing it as something like a linked list of suggestions, where you don’t know what will happen if you follow the next suggestion, but invariably something useful will come of it. Your description of it as a seris of max likelihood predictions kind of matches, except I’ll often have like 7 or 8 things that I think are worth writing, and they can apply to various parts of the text I’ve written or even be continuations of thoughts I’ve not written down for whatever reason. So I’m paying attention to ealier parts and that’s helping me generate stuff. So I guess what I’m saying, is that it is less linear in some sense than what I think you’re talking about.
Also, sometimes I will have to FOCUS to catch the feeling I’m thinking of, or break through some reluctance. Which isn’t that hard when I’m in this frame of thought, because I notice the reluctance is just another suggestion, not something I have to follow. I feel more aware, and deliberate despite just following suggestions that seem like they’re not unpleasent.
Thank you for this, I resonate with this a lot. I have written an essay about this process, a while ago: Always go full autocomplete. One of its conclusions:
I suspect that after doing this enough times, you will internalize that generating words like this leads to them being published under your name and thus your filters will adapt and make this way of writing anxiety-inducing too.
Well, at some point in the writing process there is an opportunity to edit the text, I specifically didn’t for this post because I wanted to demonstrate the raw output of the process so as to make it easier to judge the process itself. Also my discount rate is such that short term gains in writing output quality-weighted are somewhat valuable.
Did you write this reply using a different method? It has a different feel than the original post.
Partway through reading your post, I noticed that reading it felt similar to reading GPT-3-generated text. That quality seems shared by the replies using the technique. This isn’t blinded so I can’t rule out confirmation bias.
ETA: If the effect is real, it may have something to do with word choice or other statistical features of the text. It takes a paragraph or two to build and shorter texts feel harder to judge.
I didn’t write that reply (or this one) using the method. IMO it’s more appropriate to longform.
Thanks for sharing!
There’s also Ultraspeaking, which is a collection of exercises for learning to accomplish this but for spoken communication. If you are the sort of a person who finds it difficult to create conversation, you may attempt to simply start saying something and then allow your brain “autocomplete” the rest, via basically the same kind of process as detailed in this piece. I really feel like I’ve been able to level up some of my conversational abilities by using that technique.
For those who wonder how to edit and structure the content after you’ve autocompleted your way to completion, I suggest The Minto Pyramid Principle by Barbara Minto (available eg. here: https://archive.org/details/mintopyramidprin00mint).
Beware local optima.
This approach is essentially a maximum-likelihood sample, and like normal sampling this can have issues where there are very good regions of state space “down in the tree”, but no good way to get there.
While I can see the value in the suggestions you make, writing essays is not something I look forward to. I’d rather have a team of experts https://www.trustmypaper.com/ do the work for me if I could just give them a call. I know I’m not the only student to make use of resources like this.