Hello from CFAR! It’s springtime in Berkeley and we’re gearing up for a busy summer. Here’s a look at what we did during the past few months and what we’re planning for the next few.
What We Did
Alumni Double Crux Workshop
In January, a few staff members joined Eli Tyre in Bodega Bay as he led an experimental alumni workshop on the Double Crux technique. We invited a dozen mentors to spend the weekend testing out the curriculum by generating disagreements, searching for cruxes, and practicing sub-skills like operationalization and paraphrasing. Based on the results, we may do a second iteration of this experiment in the future.
We ran our first mainline workshop of the year in February. It was Luke Raskopf’s first time as workshop lead; he’ll be back in that role for our mainline workshops this summer.
Alumni rEUnion in the Czech Republic
In March, about 80 CFAR alumni and plus-ones made their way to the Jizera Mountains of the Czech Republic for the first-ever alumni reunion in Europe. Irena Kotíková and a team from the Czech Association for Effective Altruism organized the event. It created an excellent opportunity for alumni from around the world—together with most of CFAR’s staff—to swap knowledge and share experiences. Attendees gave talks on models of cognition, self-development processes, and the AI safety landscape. Other activities ranged from debugging and singing to saunas and snow hikes—and a few Hamming Circles, of course. We’re encouraged by the success of this event, the growth of the European community, and the likelihood of future collaboration among international alumni!
AI Risk for Computer Scientists Workshops (AIRCS)
We’ve continued collaborating with MIRI to run regular AI Risk for Computer Scientists (AIRCS) workshops. The workshops take place at our Bodega Bay venue and combine selected CFAR-style rationality content with a variety of topics related to AI risk.
We are excited about the potential for AIRCS workshops to boost recruitment efforts for AI safety programs, and to contribute towards building a broader community of skilled computer scientists who understand the AI alignment problem. We’ve also seen some of the curriculum development happening there make its way into CFAR’s Mainline Workshop content, and we expect this valuable cross-pollination to continue.
If you’re a computer scientist who might enjoy the workshop (or if you know such a person), you can find the interest form here; previous experience with AI risk and/or rationality is not required—though it is welcome!
What We’re Planning
We’re excited to announce we’re training a new cohort of instructors this year! It’s been two years since our last instructor training, and our curriculum is due for a reforging. Staff and instructor candidates will be diving into original questions of applied rationality, developing fluency in a particular area, and getting their feet wet assisting at our workshops this summer and fall. We kicked off the training series with a retreat in Bodega Bay last month and were excited by the amount of talent that was in the room.
Mainline Workshop in Bodega Bay (July 17-21)
We’ve set the dates for our next mainline workshops. We’re conducting interviews now through June, so complete the application at this link if you’re interested in attending. Due to high demand, we won’t be interviewing everyone who applies—but we’ll get to as many as we can.
SPARC (July 24 - August 2)
The Summer Program on Applied Rationality and Cognition will once again take place in the Bay Area this year. CFAR will be sending a few staff members to join the instructor squad as they welcome some of the most technically skilled high school students in the country for two weeks of thinking about thinking.
ESPR (August 13-23)
We’re pleased to announce that CFAR will be spinning off the European Summer Program on Rationality into a separate organization run by Jan Kulveit. Jan is currently a Senior Research Scholar at FHI working on AI alignment and macrostrategy. He previously worked as Strategy Director for the Czech EA Association, co-founded two social businesses and one for-profit organization, and did research in physics. Jan also brings some experience organizing events: he has worked on the Human Aligned AI Summer School and several experiential education programs. We’re excited that he’ll be taking ownership of ESPR going forward, as we believe he’s well-equipped to cause the program to flourish and grow as an independent entity. ESPR applications are now open at this link, if you or a talented high school-aged aspiring rationalist you know are interested in attending.
MIRI Summer Fellows Program (August 9-24)
We’ll be co-running the MIRI Summer Fellows Program again in Bodega Bay this August. MIRI reports that last year’s program was very successful, resulting in participants producing interesting write-ups on ideas they developed. We’re hopeful that mixing a bit of rationality with the talented mathematicians and programmers MIRI is bringing together can help them make serious technical progress on the AI alignment problem.
Mainline Workshops in Prague (September 16-20, September 27 - October 1)
We’re headed back to Europe this fall for another pair of European mainline workshops. The Czech EA team will be handling logistics again, and we’ll be sending staff—including some of our instructors in training—to provide the content. If you’re interested in attending, check out the application page.
That’s all for now!
All the best,
The CFAR team