I already have a rough idea of the kinds of content that I plan to share on this blog. I’m expecting different content to appeal to different people so I want to categorize things as I publish them to help make it clear up front what a given post is. I’ve summarized the planned post categories below, and posts that belong to a given category will begin with “Category Name:” so you can even set up email filters to make sure you only see what you care about.
I’ll keep this page updated over time as a sort of Table of Contents to help readers find topics of interest once there’s a healthy history of posts. I’m sure I’ll also end up creating new categories over time, or writing posts that aren’t categorized.
Research
Summaries of my learnings from different (preferably primary) sources on a particular topic. I’ll share all the claims I find likely to be both credible (>50% chance to be true) and relevant (based on intuition) to improving forecasting of Global Catastrophic Risks (GCRs). These posts won’t make arguments, they’re just intended as an overview of the most important relevant existing literature that I’ve found, in a format that hopefully makes it easy for you to dig deeper into anything that catches your attention. If I’m citing something in one of these it means I’ve read the source in full, and discovered it through some combination of keyword searching and following trails of interesting citations. Wherever possible I link to DOI’s to avoid links breaking over time.
I plan on continually updating these over time and they’ll have a change log to try and make it easier to keep up with new information. Initially, I’m targeting broad but somewhat shallow overviews across the history of a topic in the context of applying it to Global Catastrophic Risks (GCRs), that then gets deeper over time as I dig into more and more specialized sources. These posts will also give me a place to collect learnings from new sources as they’re released.
Research Posts:
Theory
Ideas I’ve had that are relevant to this blog. I haven’t tested them and they could be completely wrong. These posts are an invitation for collaborative brainstorming, and I expect the best ones to be necessary precursors to some of the more useful things I end up doing. I’m planning on them mostly being about exploration (and documenting that exploration to revisit later), but where I think I’m making specific arguments I will try to attach confidence levels.
Theory Posts:
Application
These posts document attempts to apply findings from Research posts or ideas from Theory posts to something that could actually be useful in applying forecasting techniques to GCRs. In the best cases I will be able to test these out myself and report back, but I expect that often I’ll be floating blueprints out onto the internet hoping to connect them to someone in a position to put them to use, or to spark future iterations on the design. This latter type will be inherently hard to validate without collaboration, and is intended more to gesture in the direction of something I’m confident would be useful, vs. to detail a highly specific implementation with associated highly specific expected outcomes. Regardless, I’ll do my best to capture not just what the application is but why I expect it to be useful along with where this expectation came from.
Application Posts:
Post-Mortem
These posts are retrospectives on projects I do or events I participate in, like forecasting tournaments. Expect these to be rare, but I will write them up whenever I do something relevant. These will be mostly focused on my experiences and how they inform my intuitions, surfacing them for others to potentially learn from, rather than me making particular source-based arguments.
Post-Mortem Posts:
Forecast
Anytime I do a forecast related to GCRs, I'll share it here. I want to primarily focus this blog on the meta question of how to generate the most useful forecasts vs. making actual forecasts, so expect these to be rare as well. That being said, I do have a backlog of forecasts covering the most commonly studied GCRs and I’ll be posting it here soon. These are intended to be so thoroughly explained such that another forecaster could readily take and adapt my model to match different sources or intuitions.
Forecast Posts: