Overview: In this post, I first go through what the entities listed at the top of each post on this site typically indicate. Next, I detail my considerations for the actual writing I host on this site. I then proceed to discuss what sorts of things might appear at the bottom of each post on this site and examine why this is the case. I try to explain my motivation for bringing or not bringing attention to different entities in this post. There may be many technical features of this site described in this post that are not yet implemented.

Disclaimer: Though I work on this page every so often, I do not expect that my edits will be finalized at least for the next few years. The content of this page serves as a reference for me to maintain some degree of consistency in my writing, i.e. what’s included on this page heavily guides how I write and what I write about, and I want to be especially careful in this regard.

## Top of the Page

An AI generated artwork, human generated artwork, or photograph precedes writing on most posts on this site. Typically, these visuals have some connection to the topic being covered, but it may simply be the case that I found the visual attractive and wanted to include it in the post to get a chance to look at it more often.

Each post has a title that comes after the visual, if there is one. Following the post’s title, a brief 1-3 sentence frame is used to conveyed what the post consists of, broadly speaking.

After the summarizing sentences, 1-10 content tags further indicate the nature of the post’s content; there are sometimes an additional 0-3 tags indicating topics that are specific to that post. These “Specifics” tags, as I call them, refer to topics that I do not believe I will write on extensively in the future. After the tags, between 3 and 12 (usually less than 6) labels for such things as reading time, word count, and/or importance are included to further assist the reader in understanding the overall nature of the post.

Finally, 1-4 paragraphs are allocated for things such as an overview, summary, or disclaimer, among other things, for the post. The Table of Contents is the last entity I that I believe should be considered to be contained in Top of the Page.

Here are some further thoughts on each of the aforementioned Top of the Page entities.

ART

Human generated artworks on this website typically come from one of the following: Unsplash, Birmingham Museum Trust, National Gallery of Art, or USEUM. At the time of writing this1, each of these resources have licenses that permit people to download and post some of their images freely.

My favorite artworks typically are Baroque, Realist, and oil paintings. I very much enjoy scenes of cities, natural landscapes, gardens, the night sky, space, and catastrophe. While I do not know of many artists by name, my favorite artist is Gustav Klimt, and my favorite painting is his Death and Life

Death and Life by Gustav Klimt

The AI generated pieces I’ve made via OpenAI’s DALLE-2 during the months of August and September 0002022. Hopefully I will be able to diversify the selection with some other DL methods soon, or even program them myself from scratch at some point.

Why use the art, anyway? My motivation is twofold, with the second option having more weight: (1) I look at this site more than any other person and enjoy looking at art, so I’ve viewed this site as an opportunity to do so, and (2) the art sometimes complements the post’s content in such a way that readers may find it easier to appreciate, contest, or absorb the information (e.g., a “timeless” artwork of space seems fitting for a post on longtermism).

One contention to including visuals on site might be the dark theme. Many programmers, writers, etc… on the Internet, including myself, enjoy dark themed websites, and bright visuals put strain on our eyes. I am sorry if this is the case for you, but I am going to retain the visuals for now. Some other contentions might be that the visuals distract too much from the post’s content or bias the reader in a way that distorts their capacity to engage with the post’s content accurately and with skepticism. I believe that the art is indeed distracting, albeit not enough to seriously prevent someone from engaging with the content of the post. I also believe that the art does bias the reader, but do not know the nature of this bias, and whether / how much of an issue this is. If it is learned that the visuals trap people in certain ways of thinking that are misaligned with reality, I may reconsider using the visuals, but I don’t expect the bias induced by visuals to be found to be this potent.

Credit for art or photos is provided under Cover Image in the Notes section at the Bottom of the Page. I have not yet decided whether to include the prompt for the artwork in the Cover Image section or anywhere for that matter if the artwork is AI generated. Perhaps it might be more intriguing to leave the prompt a mystery.

Some posts, especially short ones, do not have any visuals to accompany them. Given the lack of interesting content in these posts (e.g., a list of Wikipedia articles I want to read), I might opt to not include any art.

TITLES

I optimize post titles on this site for brevity and comprehensiveness.

FRAMES

Most posts on this site have a 1-3 sentence framing of the post’s content below the title. With this framing I try to succinctly answer, in the following sequence, these questions: (1) What is this post about?, (2) Why does this post exist?, (3) How do I proceed in exploring the “what” in (1)?, and (4) Why do I choose the “how” in (3) and not another “how”? I do not try hard to cover these four questions in each post. I would expect fairly inconsistent framings of post content across the site. Despite this inconsistency, I suspect people will mostly find the post framings useful for sorting through this site’s content.

Here is one example framing from my post My Life: Q3, 0002022:

This page (along with the other pages in Routine Personal Predictions) consists of my predictions about how some aspects of my life might change in the coming yearly quarter; these events serve as a partial record of my life and predicting their outcomes motivates me to get things done and make strong decisions. My predictions here can be perceived primarily as attitudes, given that few contain detailed justifications; not justifying most of my predictions here keeps the workload light, and I am willing to make this sacrifice for the time being.

Here is another example from my post Metaculus Track Record (Binary): [06/2021, 06/2022]:

Every prediction on Metaculus that I’ve inputted on binary questions between June 2021 and August 2022, my first year on Metaculus. These pages exist as demonstrations of my weaknesses and strengths in forecasting, and also serve as a record of potential growth I experience.

GENERAL TAGS

Below are all the general tags currently in use on the site.

Sometimes I find it difficult to choose which general tags to include in a post, but in most cases, I aim to include between 1 and 10 general tags. The criteria for this range is simply that I do not want to include too many tags. For me, > 10 general tags is too many.

SPECIFICS TAGS

These are special tags that convey more detailed information about the content of a post. “Specifics” exist because there are things I will likely not write about on this site again. Of course, there might be two or three posts that eventually come to include the same Specific tag, but that this tag refers to something so narrowly covered (e.g., the general tag biology versus the specific tag DNA) makes me disinclined to make it into a general tag.

LABELS

The following labels communicate aspects of the post as a whole. Some, such as Reading Time or Word Count, are straightforward, whereas others such as Certainty are not. The (*) next to the label indicates that I will cover it in greater detail outside of the following list. Some of these labels, such

• [Start Date] - [Date of Last Modification]: The 3 zeros in front of all dates on this site’s posts are personal reminders to think about where humans will be in 10, 100, 1000, 10000, 1000000, etc…, years, as this, along with how my personal life factors into the history of humanity, are things I want to be pondering quite frequently.
• Type - Status (*)
• Certainty (*)
• Importance (*)
• Impact (*)
• Word Count: The word count for a post, calculated using Liquid’s page.content divided by Liquid’s number_of_words.
• Reading Time: The word count divided by 280. The selection of 280 was somewhat arbitrary. I originally had 315 after, I believe, looking at list of college-level reading times. From the EA-Forum, this post was 998 words and had 4 minutes listed as the reading time, so ~249.5 words per minute. The average is approximately 280 words per minute.
• Durability (*)
• Extent (*)
• Views: [unimplemented] A button one can press to illustrate that they viewed the post. The idea behind this would be to avoid having to use Google Analytics and to remove extremely quick brief viewings from being counted.

TYPE-STATUS (Label)

This label is pretty self-explanatory in my opinion: “type” indicates which category or subcategory a post belongs to and “status” refers to how far along the post is. There are many combinations for type-status that I may use. In general, the format will likely consist of combinations of the following: [Forecast, Log, List, Review, Idea, Essay, Research, Learnings, Thoughts, Experiment, Report] - [In(complete), (Un)finished, Ongoing, Un(resolved), Speculation, Limited, Draft]. In some instances, the first grouping may have additional clarification (e.g., instead of just using Log for a log of my forecasts, I’ll use Forecasting Log). It should be noted that I do not presently feel fully satisfied with these categories and therefore might make significant revisions to them the future.

Here is some further clarification on the type of content that I host on this site. Some content forms may overlap a little.

Thoughts Exercises, ideas, or reflections on ideas or on facets of my life
Notes Notes, summaries, critiques, and/or comments on some work or idea
Forecast My forecasting process and predictions on some particular phenomena
Essay A discussion on some subject where I do not aim for complete com
Tutorial Explorations and/or explanations of a concept, positions, process, or set of ideas
Review Any lengthy piece of writing on this site that attempts only to evaluate the current state of research on some topic or question(s)
Report Any lengthy piece of writing on this site that might greatly influence people’s decisions or beliefs about some topic
Essay Any lengthy piece of writing on this site that is not a report or review
Experiment Observations and ideas relating to any experiments I conduct
List A static list or collection of something
Log Data about myself or relating to something I’ve done over time

CERTAINTY (Label)

My estimate of how accurate my claims

This refers to my estimate of the accuracy of what I’ve written in my post.

“Accuracy” here includes my sense of whether the distribution of feelings and thoughts engendered by reading my content matches what I intended to have engendered, how well the set of all claims in a particular post are synthesized, how evidenced my claims are, how well I evaluate how evidenced my evidence is, how thorough I am in revealing my uncertainty, and, among other things, how predictive what I’ve written is of phenomena related to the post’s content.

My use of these certainty tags was inspired directly by Gwern. I believe the following excerpt from Gwern’s website is likely illustrative of his thoughts on confidence tags:

The “confidence” tag is a little more unusual. I stole the idea from Muflax’s “epistemic state” tags; I use the same meaning for “log” for collections of data or links (“log entries that simply describe what happened without any judgment or reflection”) personal or reflective writing can be tagged “emotional” (“some cluster of ideas that got itself entangled with a complex emotional state, and I needed to externalize it to even look at it; in no way endorsed, but occasionally necessary (similar to fiction)”), and “fiction” needs no explanation (every author has some reason for writing the story or poem they do, but not even they always know whether it is an expression of their deepest fears, desires, history, or simply random thoughts). I drop his other tags in favor of giving my subjective probability using the “Kesselman List of Estimative Words”

1. “certain”
2. “highly likely”
3. “likely”
4. “possible” (my preference over Kesselman’s “Chances a Little Better [or Less]”)
5. “unlikely”
6. “highly unlikely”
7. “remote”
8. “impossible”

These are used to express my feeling about how well-supported the essay is, or how likely it is the overall ideas are right. (Of course, an interesting idea may be worth writing about even if very wrong, and even a long shot may be profitable to examine if the potential payoff is large enough.)

After skimming the word tables in the Kesselman List of Estimative Words, I chose to use this one, which I found on page 20:

Verbal Expression Probability
Impossible 0.00
Small Possibility 0.10
Small Chance 0.20
Somewhat Doubtful 0.30
Possible 0.40
Toss-up 0.50
Somewhat Likely 0.60
Likely 0.70
Very Likely 0.80
Quite Certain 0.90
Certain 1.00

While I have not yet implemented the following idea, I want to eventually include a “calculation” for each post’s aggregate “Certainty” in the Page Notes Header.

IMPORTANCE (Label)

An estimate for how important the topic I am writing on will be in the coming years.

IMPACT (Label)

I consider myself an Effective Altruist, which means to me that I should allocate some cognitive bandwidth towards assessing what impact I make on the world. The term “impact” here is nebulous, but I believe “the impactfulness of my writing content” should refer to some mixture of the following: effects on reader well-being, effects on collective human well-being, effects on reader decision-making (this includes behavior), and effects on preserving long-term human potential. This is not an exhaustive list and, ideally, I would like to come up with some way to measure each of these things, normalize them, and then use them together to create the Impactfulness score.

For measuring the impact of the content I generate or host, I could ask: How does this make the reader feel? Will any reader’s interests shift towards the topic of my writing? Might this post be hazardous in some manner to humanity’s long term potential, or might it result in suffering? How probable is it that this post changes people’s behavior, for better or worse? These serve as a first pass at gauging my impact, but I would like to do better, and will likely spend more time in the coming years thinking about what factors to consider when assessing how consequential a post of mine is.

At the present moment, I am not familiar with many “impactfulness” scoring rubics, and the only one the comes to mind was from Eli Lifland’s, yagudin’s, and sam_atis’s LessWrong post Impactful Forecasting Prize for forecast writeups on curated Metaculus questions. In this post, they incentivize people to explain their forecasts on a set of Metaculus questions that they scored using the following rubric

Impact Dimension Explanation
Decision importance The importance of the decisions which will be affected by this question. Should combine cause area importance + importance within cause area.
Decision relevance How much of an impact would this have on actual decisions if the forecast changed by a substantial amount? This factor is re-used from Nuño Sempere’s An estimate of the value of Metaculus questions
Ease of contribution ​​How easy will it be for a “median generalist forecaster” to make a contribution to the analysis on this question within a few hours? e.g. questions requiring lots of domain expertise or background reading would score low here.
Neglectedness of contributions How few contributions have there been on this specific question so far? How in need of attention is it? This should be subjectively evaluated using the existing count of forecasts and quantity + quality of comments/writeups.

They also write

A curation score was calculated, weighing decision importance at twice the other three due to it feeling like the most important factor. We chose a set of 25 questions based mainly on the curation score, but also including a diversity of cause areas and question types.

Their scoring system is appealing to me, and I expect myself to incorporate it into my writing somehow, at least until I develop a better system for measuring the impact of my posts. Since their measure of impactfulness mostly pertains to Effective Altruism oriented forecasting questions, I have modified the system to be more flexible, particularly for the content I post on this website. I use a 1-10 scale for each of the following entries in the table, and also weigh decision importance twice as much. Instead of summing the weighted scores, I multiply them and then divide by $20 \cdot 10 \cdot 10 \cdot 10 = 20000$ (the maximal impactfulness score) before multplying by 100 to convert this value into a percentage.

Impact Dimension More Flexible Explanation
Decision importance The importance of the reader’s decisions affected by this post.
Decision relevance If the content on of a post changed substantially, how much would the reader’s decisions change, relative to how the reader’s decisions before the change?
Ease of contribution ​​How easy will it be for a reader to make a contribution to the analysis on this post?
Neglectedness of contributions How much have people thought about the content in this post? How in need of attention is the content in this post?

EXTENT (Label)

A rough estimate of the amount of time I spent doing or thinking about something. The Extent label at the top of the page is usually an aggregate of the different extent measures throughout the post. For example, if in my post I reviewed 5 research articles and then thought and wrote about each of them for some amount of time, I would take the average “extent of my thinking/writing” across the 5 articles to get the final extent label at the top of the post. Note that “Extent” applies both to searching for something and to reading and thinking about something. .

Let IQRTWE mean: Internet querying and/or reading/thinking and/or writing/editing

• (1) Writing it to get it written/out of my head
• (2) IQRTWE ~5 minutes
• (3) IQRTWE 5 - 30 minutes
• (4) IQRTWE 30 - 60 minutes
• (5) 2 - 5, 60 minute IQRTWE sessions
• (6) 6 - 15, 60 minute IQRTWE sessions
• (7) 16 - 35, 60 minute IQRTWE sessions
• (8) 36 - 75, 60 minute IQRTWE sessions
• (9) > 75, 60 minute IQRTWE sessions

• Update (date):
• Disclaimer
• Epistemic Confidence: A standalone comment on my confidence in this post for times when I feel that a full canned transparency is not necessary.
• Overview
• Summary
• Takeaways: The major points in the post
• Canned Transparency (*)
• Contribution
• Acknowledgements
• Post Notes (*)

I want this site to be an example of strong epistemics. As such, I want to be as transparent as possible with my posts.

Here are types of support for claims, from OpenPhilanthropy’s Reasoning Transparency.

• another detailed analysis you wrote
• careful examination of one or more studies you feel qualified to assess
• careful examination of one or more studies you feel only weakly able to assess
• shallow skimming of one or more studies you feel qualified to assess
• shallow skimming of one or more studies you feel only weakly able to assess
• verifiable facts you can easily provide sources for
• verifiable facts you can’t easily provide sources for
• expert opinion you feel comfortable assessing
• expert opinion you can’t can’t easily assess
• a vague impression you have based on reading various sources, or talking to various experts, or something else
• a general intuition you have about how the world works
• a simple argument that seems robust to you
• a simple argument that seems questionable to you
• a complex argument that nevertheless seems strong to you
• a complex argument that seems questionable to you
• the claim seems to follow logically from other supported claims plus general background knowledge
• a source you can’t remember, except that you remember thinking at the time it was a trustworthy source, and you think it would be easy to verify the claim if one tried
• a combination of any of the above

General transparency on how what I’ve written is inadequate.

• Has the author presented a fair or biased presentation of evidence and arguments on this topic?
• How much expertise does the author have in this area?
• How trustworthy is the author in general?
• What are their biases and conflicts of interest?
• What was the research process that led to this analysis? What shortcuts were taken?
• What rough level of confidence does the author have in each of their substantive claims?
• What support does the author think they have for each of their substantive claims?
• What does the author think are the most important takeaways, and what could change the author’s mind about those takeaways?
• If the analysis includes some data analysis, how were the data collected, which analyses were done, and can I access the data myself?

Here are the entities I usually list in Canned Transparency:

How trustworthy are you? How/why is this post important? Who would you recommend read this post? What are your priors? How much external review has this post undergone? What is your epistemic confidence? What types of support does this post’s claims use? What shortcuts did you take? What are this post’s major inadequacies? How did you update your beliefs after writing this? How should I update my beliefs after reading this? Did you contribute anything notable by creating post?

Why does this essay exist? Who is this post for? How good is this essay? Can you trust me? What are my priors? Why? What is my confidence? Why? This essay’s claims? This essay’s inadequacies? This essay’s shortcuts? My updated beliefs? My sense of where you should update? My contribution?

• Are you an expert?
• How/why is this post important?: Why should you read this?
• Who would you recommend read this post?
• What are your priors here?: (my intuitions)
• How much external review has this post undergone?
• What is your epistemic confidence here? How incorrect I expect myself to be
• What types of support does this post’s claims use?
• What shortcuts did you take here?
• What are this post’s major inadequacies? (*)
• How should did update your beliefs after writing this?
• How should I update my beliefs after reading this?
• Did you contribute anything by creating post?

• To the reader: Critique this post, please. I want to be less wrong; please point out how I am wrong. This includes pointing out such things as improperly using statistics, generalizing when I shouldn’t be, poorly allocating attention to certain things (allocating too much or too little), omitting information (papers, datasets, mental models for thinking about something), misrepresenting the findings of a citation, framing speculation as being well supported when this is not clear, inadequately framing the distribution of something, be repetitive in my comments, not explicating when explication is due, miscalibrating how impactful or important the subject is, not updating my writing given new evidence on the subject.
• Importance Calculation:
• Certainty Calculation:
• Extent Calculation:
• Durability Calculation:
• Impact Calculation:
• My Reflections:

I thought this listing was interesting and potentially useful. I found it on Gavin’s Leech’s Argmin gravitas

Inadequacies is a catalogue of silly statements by scientists or journalists. It’s sort of the gag page of the Inference Review, a witty and arch journal for scientific prose. Entries fall into at least one of:

• Exaggeration (E)
• Irreproducible results (IR)
• Begging the question (BQ)
• Confusing correlation with causation (CCC)
• Plagiarism (P)
• Ill-conceived experiments (ICE)
• Ill-defined concepts (IDC)
• Conflicts of interest (CI)
• The numbers don’t add up (2 + 2 = 5)
• Purely ornamental mathematics (POM)
• Appalling prose (AP)
• Why did someone publish this? (WDSPT)
• Just plain dumb (JPD)
• Don’t touch our funding (DTF)
• We told you so (WTYS)
• Too close to call (TCC)
• Could be (CB)
• Stating the Obvious (SO)
• All of the Above (AA)

[pending]

## Notes

### Cover Image

Footnotes

1. Each time I edit this post, I will update this value. The present for me right now is: 08/23/0002022