This page is about juan.al content; for information about me, see Me⁠.

This site contains entries dating back to 2001, when I started blogging. In these two decades, I’ve written about a wide variety of topics: travel, hiking, living around the world, books, movies & TV series, anime, video games, D&D adventure logs, doodles, self-experiments, yoga, learning, and expertise. As of August 2022, I’m in the process of cleaning up the site, removing useless entries, and improving the existing content before writing new one.

I write to understand things better, and I write mainly for myself in a public way, an approach called digital garden nowadays. I hope that people reaching this site finds value in what I write, of course, but right now that’s not my priority. I also revise my writing constantly as I revise my thinking, which may leave pages in an inconsistent state despite my efforts to maintain readability throughout revisions.

What you see on these pages is my best understanding of what I write about. If you think I’m wrong (factual errors, invalid or incomplete conclusions, etc.), please let me know; you will be helping me.

How to be notified of new content

Email

RSS

• https://juan.al/index.xml (you will see all entries)
• https://juan.al/tags/CHOOSE_TAG/index.xml (see available tags)

Entry metadata

At the top of each page, you will see metadata information, including two uncommon attributes that I have copied from gwern.net:

• Certainty: how sure I am that what I’ve written is true.
• Importance: how important I think what I’ve written is to humankind.

I write to be less wrong about how the world works, and I would rather focus on important rather than non-important. Keeping these two attributes in mind helps me stay focused on my goal with the site.

Forecasts

The site has a section with predictions about the stock market to help me improve my ability to think probabilistically. I record those predictions using a tool I wrote for this purpose based on ideas from Superforecastering by Philip Tetlock. Other similar but publicly available prediction market tools are:

• The Good Judgement Open: community with skilled forecasters and high learning potential, but you can’t register your own events. It’s connected with P. Tetlock and Superforecasting.
• Prediction Book: you can register your own events, but many events are personal, making it harder to learn from others.

Comments

I used to have a comments section on this site. In the beginning, it was invaluable for me to get motivated to keep writing and to feel like I was contributing valuable ideas and observations, but as the visits grew, so did the spam, questions, and hacking attempts, so I decided to remove them. Thank you to all who left comments during that time, and I apologize for having deleted them.

Spanish pages

I started blogging in 2001 in Spanish for a Spanish audience. Most of what I wrote in those days wasn’t valuable and is now gone, but I couldn’t get myself to delete what was left, and those entries are now mixed with English ones. I haven’t found a better way to deal with this situation than to pretend that mixing languages isn’t a problem, so I have done nothing about it. I apologize to those of you who don’t speak Spanish.

Site technology

The first versions of the blog were entirely static HTML files made in Macromedia Dreamweaver. Later I found out how to create dynamic pages, converted the blog to PHP, and started using Vim. After that, I discovered Wordpress.org and used it self-hosted for many years. Eventually, I got tired of being hacked and moved to Hugo, a static site generator, and Vim. Nowadays I start writing drafts in Obsidian, usually after sketchnoting on a Moleskine or iPad. The drafts get exported to Hugo through a chain of Git post-commit hooks that ultimately triggers a Netlify deployment.

Image format

On Aug 24, 2022, I converted all site images from JPG and PNG to WebP because it seems like a widely supported format, and preliminary tests were promising: significant bandwidth savings with no discernible loss of quality. I used the following script which relies on Google’s WebP Converter.

# Create WebP files
$ find Public/ -depth -name "*.png" -exec sh -c 'cwebp -preset picture -q 95 -hint picture "$1" -o "${1%.png}.webp"' _ {} \;
$ find Public/ -depth -name "*.jpeg" -exec sh -c 'cwebp -preset picture -q 95 -hint picture "$1" -o "${1%.jpeg}.webp"' _ {} \;
$ find Public/ -depth -name "*.jpg" -exec sh -c 'cwebp -preset picture -q 95 -hint picture "$1" -o "${1%.jpg}.webp"' _ {} \;

# Update entries references to use WebP
find Public -type f -name '*.md' -exec sed -i '' s/\.jpg/\.webp/g {} +
find Public -type f -name '*.md' -exec sed -i '' s/\.png/\.webp/g {} +
find Public -type f -name '*.md' -exec sed -i '' s/\.jpeg/\.webp/g {} +

The sum of all images occupied 650Mb of space before the conversion; after the conversion, they occupy 412Mb, a 47% decrease. I manually inspected several images, and I couldn’t tell JPG from WebP apart. Not every day can you improve things by almost 50% with so little effort. Thanks Google!

Museum

Since the early 2000s, the blog has had quite a few different looks. Here are some snapshots ¹ in chronological order from 2001 to 2022:

Footnotes