Tim Disney
How I’m Trying to Use BlueSky Without Getting Burned Again. I really like the principle of assuming the platform/service you are using will go away in three years. That's a long enough timeline to derive value without overcommitting.
Browser: Arc
Still a transformative browser UX but I'm worried about its future considering their recent pivot to a new "AI first" browser. We'll see how things go in 2025.
Notes: Obsidian
Continues to run my life.
Read it Later: Readwise Reader
Still the best thing ever. I have highlights sync to Obsidian, which is the best thing ever.
Task Management: OmniFocus
I tried to switch to Things last year but it didn't stick. Back to OmniFocus and it's great although the v4 update was a bit rough but it's mostly evened out now.
Code Editor: Zed
Zed is incredible. Fast, great VIM mode, AI integration is great (though I need to check out Cursor at some point too).
AI Assistant: Claude
This was the year I started to actually get productive with AI assistants. They aren’t always good but when they are they really are. I mostly use Claude at this point though I don’t really do extensive comparisons.
Launcher: Raycast
A bunch of my one-off AI assistants use is actually via Raycast. Super helpful to have it always a keystroke away.
Calendar: Fantastical
Continues to be great.
Podcasts: Overcast
Continues to be great
Email: Mail.app
Boring but I’ve come around on that being what I like out of email.
Finance: YNAB
Continues to be great.
AI Scaling Myths. As 2024 went on, more and more AI discourse started to come around to the idea that AI scaling was probably about to end. One things I learned from this article is that "scaling laws" have a more precise meaning than what I'd known.
I had a general understanding of "scaling laws" to mean that the bigger the model the more "capabilities" were supposed to pop out. But the original meaning in AI research is about the "decrease in perplexity" (perplexity being jargon I also learned from this article meaning a measure of word prediction uncertainty). A decrease in perplexity is not quite the same thing as new capabilities.
I've been playing a bunch of Palworld recently (along with everyone else apparently) and been having a great time. It's open-world survival crafting (think Valheim) mixed with monster collecting (Pokémon) mixed with some very light (at least in the first ten or so hours, might get more complicated later) automation (a little bit of Satisfactory).
I'm a little surprised how much it has hooked me since I've always bounced off the survival crafting and monster collecting genres. I understand their appeal but the tedium inherent in both have been too much for me (also the combat in Pokémon just straight up sucks, don't @ me). That said, I've sunk hundreds of hours into automation games (every few months I circle back to either Satisfactory, Factorio, or Dyson Sphere Program), and those games definitely have plenty of tedium baked into their design.
So what makes Palworld click for me?
I think the key is the way that Palworld uses monster collecting to relieve the survival crafting tedium. The pals you collect are not just used in combat, you put them to work in your base to gather resources and craft items. The tedium is a motivator to find new and better Pals. There is a virtuous cycle at work here; you collect Pals to build a better base and craft better items that you use to find better Pals that you use to build a better base and so on.
This ever-building cycle is exactly the hook that pulls me so deep into automation games, the initial tedium is a motivator to build more complicated factories that you use to build yet more complicated factories and so on.
I'm only ten or so hours into Palworld so I'm curious to see if its hook can sustain itself to an endgame. I suspect not, automation games have another endgame hook that Palworld seems to be missing: player expression. As you build out more complicated factories in a game like Satisfactory the player expresses their design sensibilities in their builds (how they prioritize aesthetics, compactness, modularity, etc.). At least so far, I don't see an equivalent space for player expression in Palworld (I'm not counting base building because it's not tied back into the core mechanics, bases are mostly cosmetic).
In any case, Palworld is very enjoyable and I'm interested to see how it unfolds both as I play more of it and as the developers evolve it during the early access period.
I had a neat little interaction with an LLM this week where I had a markdown note in Obsidian I wanted to convert some of its formatting. The note had a bunch of external markdown links:
- [Note1](app://obsidian.md/path/to/Note1)
- [Note2](app://obsidian.md/path/to/Note2)
- ...
that I wanted to convert to internal links:
- [[Note1]]
- [[Note2]]
- ...
Previously to do this I'd either reach for something like a vim macro or write a little regex. Realistically for most cases the activation energy would be too much and I'd just not do the reformatting at all.
But with a LLM (in this case I used the TextGen plugin) you can just show an example and tell the LLM to do it and it just works. The prompt was something like
- [Note1](app://obsidian.md/path/to/Note1)
- [Note2](app://obsidian.md/path/to/Note2)
- ...
---
convert the list above from `[Note](path/to/Note)` to `[[Note]]`
It's pretty cool that we basically get a kind of programming by example for "free" with LLMs.
This is a writeup of how I use Obsidian to organize my personal knowledge base (PKB).
My goal in building out a PKB is to capture and reflect on the most relevant thoughts, ideas, and events in my life. I've tried a bunch of different tools over the years and have found that it is critical to decide on the right level of structure to impose. Too much structure is overwhelming and nothing gets written down. Too little structure leads to sprawling write-once read-never snake nests.
As a disclaimer, PKBs are very, well, personal so this approach might not work for you.
At a high level, the kinds of notes I use are:
On any given day I use a daily note to capture whatever comes up. Daily notes serve as a kind of scratchpad to capture random thoughts, drafts of other documents, and any other bits of ephemera.
The only structure I impose in a daily note is for capturing meeting notes where I'll use a heading and link to the related project or person. For example, a 1/1 meeting will have a heading like # [[Person Name]] 1/1. The links in the heading to [[Person Name]] provide backlinks so that in the future I can quickly see all the topics we've covered.
For a while I tried to get by with just the daily note level of structure but it wasn't enough. It was too easy for something I wrote in a daily note to get lost. I needed a structured way to reflect on a longer timeline than a day.
So, I use the periodic notes plugin to create both weekly and quarterly notes.
The weekly note is more structured than a daily note and I use a template to fill in headings that I expand on throughout the week.
The first section is for reflection on how the week went. It uses a retro format I've used for years at work with simple "happy" and "sad" sections.
# Reflect
## Happy
- good progress on [[Project X]]
- three day weekend!
## Sad
- [[Twitter]] is at it again
- stalled discussion on [[Project Y]]
- need to follow up with [[John Smith]]
The next section is to track the media I consumed that week. Any movies, shows, or games.
# Media Consumed
- [[Silo]]
- [[Factorio]]
I don't bother rating or writing down much else since the point is just to remember what and when I encountered various bits of media.
The final section is the Readwise highlights from the week. I use Readwise Reader for almost all my reading so when I make a highlight in Reader it gets automatically synced to Obsidian via the Readwise plugin. Having highlights synced is great for preservation but realistically I never looked at them again (outside of the Readwise spaced repetition emails). What I wanted was a way to see my recent highlights and have a chance to reflect on them again.
Enter the dataview plugin:
list
from "Readwise"
where
file.ctime > date("{{date:YYYY-MM-DD}}")
and file.ctime < date("{{date:YYYY-MM-DD}}") + dur(7 day)
sort file.ctime
That query pulls back a list of all the Readwise notes created during the week. At the end of the week I look over them and create any evergreen notes that might make sense.
The quarterly notes are similar to weekly notes but with a few more sections to encourage me to reflect a bit deeper.
Quarter notes start with a Happy/Sad reflection section just like weekly notes however this isn't meant to be inclusive of all the reflections from every week, just the most salient items from the quarter.
The other sections are aspects of my life that I want to spend time reflecting on and should be fairly self-explanatory:
I try to spend a bit more time on each of these sections than I would in a weekly note, writing full sentences and paragraphs rather than just listing out bullet points.
So the way I use my PKB is essentially jotting down whatever is happening in the moment in a daily note, on the weekend filling out the weekly note and reflecting on highlights from that week, and then at the end of the quarter spending a bit more time reflecting on what happened in a number of different areas over the past few months.
I've struggled over the years trying to find the right cadence and tools to effectively capture and reflect on my "personal knowledge", but this combination of daily, weekly, and quarterly notes has been working pretty well for me so far.