Launching the FRI blog

I’m happy to announce the launch of FRI’s new blog.

We were moved by the many good reasons to make conversations public. At the same time, we felt the content we wanted to publish differed from the articles on our main site in the following ways:

  • While most of our main site content is aimed at audiences that share our mission and requires the ability to engage with an unusual breadth of considerations drawing from a variety of disciplines, it usually doesn’t require extensive prior knowledge. By contrast, on the blog we may publish research notes that assume a greater familiarity with some field such as AI safety or decision theory.
  • While not every researcher agrees with every statement in all our main site articles, it is usually safe to assume that our team’s views are clustered around the general picture painted by the content we’ve published. By contrast, on the blog individual researchers will be free to publish content that hasn’t been vetted by other team members, or about which they themselves retain considerable uncertainty.
  • Blog posts may sometimes contain content which is only tangentially related to our core research areas but which we still think is interesting to our audience.
  • Blog posts may be shorter than main site articles.
  • We’ve used the main site strictly to publish research output. By contrast, the blog will feature both research-related content and organizational updates.
  • Main site articles are usually edited before publication, but we don’t think it’s worth investing this extra effort for every blog post.

Upcoming blog posts

Content we plan to publish on this blog before the end of July includes:

  • A strategic update, outlining FRI’s vision and mission, through which inputs, outputs, outcomes and impacts we plan to put our mission into practice, our strategic goals for the next year or so, and how we plan to measure whether we’re still on track.
  • The written notes and slides of a talk about s-risks which I gave at this year’s EAG Boston.
  • A summary of an internal survey that polled our researchers for their credences in assumptions underlying FRI’s strategy and statements that play a central role in our research. We did this survey mostly to experiment with ways to track our epistemic progress, and we feel the results shed more light on some challenges associated with this than they constitute evidence other people should account for when making decisions.
  • Reposts of relevant pieces that have been published by our researchers on their personal blogs (see below).

How the FRI blog relates to our researchers’ personal blogs

Several of our current and former researchers publish content on their personal websites or blogs:

Researchers will maintain their personal websites as they anticipate that not all of the content they’ll publish there will be relevant to FRI. Content that concerns the research of current staff members will, however, be crossposted to the FRI blog. This way, readers who are primarily interested in FRI-related content need only follow one blog.

Similarly, we anticipate that not all FRI blog posts would qualify as a good EA forum post, but we plan to publish the ones that do there. We’ve also (cross-)posted to LessWrong and the Intelligent Agent Foundations Forum and plan to continue doing so.


We’ll explore enabling comments on this blog as this may encourage additional readers to engage with our work and to provide feedback. I expect to reply to questions being asked in the comments.

Leave a Reply

Your email address will not be published. Required fields are marked *