What is Going On With CFAR?

CFAR is not defunct.  I am at a workshop right now with Jack and Vaniver (who also work with CFAR), and 7 other people, who are mostly adjunct CFAR instructors (aka, people who are skilled enough to run CFAR classes if they want, and who've trained with us some, but who usually do only very few hours of CFAR-esque work in a typical year), trying things out.  Dan Keys is the fourth person who is also on CFAR's "core staff" (though we are all part-time hourly); there's also a number of other adjunct instructors.

Like some at MIRI, I've been taking something of a sabbatical year, loosely speaking.  That is, I've been looking at the word trying to understand its joints.  For me this sometimes involves running small experimental workshops, e.g. to try to see a thing with a group of people helping and geeking out about it together.  It doesn't so far involve trying to do anything at scale.

There are no super-secret projects at CFAR.  I suppose I might not say if there were, but I at least wouldn't say this.

We haven't run mainlines in awhile.  Irena (who is an adjunct CFAR instructor, and also upstairs in the venue asleep right now) keeps saying she may run one in Prague; if so, I and some others may fly in to help.  Davis keeps saying he may co-organize one in Berkeley, in which case I'll probably help also, but it's not quite clear.  If someone wants one and wants to co-organize, I may be down in general, but I don't quite have the fire to generate those on my own right now -- there may be people who would like to attend a CFAR workshop (okay, I know there are at least some), but running one isn't quite the thing that feels like it'll help me unravel the things I'm trying to understand, and also I don't quite have the stomach for it in some ways, although I'm glad that the people who attended got to attend so this is a bit of a muddy thing to express well.  It's possible the workshop we're currently experimenting in upstairs may lead to a revised mainline at some point, and one my heart could be more solidly in, but that "scale this up to a mainline" outcome is not a thing we're driving at especially hard.

CFAR's internal structure lately involves a "telos committee" who authorizes the allocation of funds internally by checking that the individual (from within CFAR's "core team" or CFAR's adjunct instructors / outside collaborators) has "telos" for the thing they want to do (and tries to get out of the way and let people do things they have "telos" for, without them needing to persuade others much).  I like how this has been going.  It is pretty low-key, though.  It is plausible to me that we had to wind down before we can build freshly.  We wound things down by ceasing to do anything that nobody had "telos" for (even things that were traditional, such as mainlines).  Sort of a theory that things would be easier to see, or at minimum we'd have more slack with which to see, if there wasn't any of that sort of clutter around.

I would not advise anyone wishing to solve human rationality, or to do anything else awesome, to refrain from attempting said awesome thing on the theory that we or anyone else has that covered.  Such thinking always seemed insane to me; if it is more transparently insane now, that seems good.  In hindsight, I wish we had chosen a more local name than the "Center for Applied Rationality" (such as "A Center for Applied Rationality" or "some random group in Berkeley who are taking a go at some applied rationality stuff while also having some more local goals about supporting MIRI's staffing needs, and should totally not be occupying the entire namespace here").  We do not have a super secret rationality sauce such that people attempting such a thing from outside CFAR are at a bad disadvantage, or anything like that.  If you want to try and want to make sure we don't know something you're missing first, I'm probably happy to talk.  Others might be too but I can't speak for them.

In terms of whether there is some interesting thing we discovered that caused us to abandon e.g. the mainline: I can't speak for more than myself here either.  But for my own take, I think we ran to some extent into the same problem that something-like-every self-help / hippy / human potential movement since the 60's or so has run into, which e.g. the documentary (read: 4-hour somewhat intense propaganda film) Century of the Self is a pretty good introduction to.  I separately or also think the old mainline workshops provided a pretty good amount of real value to a lot of people, both directly (via the way folks encountered the workshop) and via networks (by introducing a bunch of people to each other who then hit it off and had a good time and good collaborations later).  But there's a thing near "self-help" that I'll be trying to dodge in later iterations of mainline-esque workshops, if there are later iterations.  I think.  If you like, you can think with some accuracy of the small workshop we're running this week, and its predecessor workshop a couple months ago, as experiments toward having a workshop where people stay outward-directed (stay focused on inquiring into outside things, or building stuff, or otherwise staring at the world outside their own heads) rather than focusing on e.g. acquiring "rationality habits" that involve a conforming of one's own habits/internal mental states with some premade plan.

The above is somewhat scattered; feel free to ask more questions.

Don't know if I have all the info, but I believe Anna Salamon continues to run small experimental workshops. For instance there's one happening right now.

But both CFAR and MIRI exist much less than they used to, having wound down and let go lots of staff and not done much to wind back up again.

I'll be excited for either of them to start new projects. That said I'm not betting on CFAR to solve rationality or MIRI to solve alignment, and I'd encourage anyone to solve those problems directly if you would like to see those problems solved.

Duncan Sabien here, worked at CFAR 2015 to 2018 (currently work at MIRI).

My understanding of the state of CFAR:

"Defunct" is a reasonable description, relative to the org that ran 5-15 workshops per year and popularized TAPs and Goal Factoring and Double Crux and so forth.

Currently, it is not-quite-accurate but closer to true than false to say "CFAR is currently just two people."  The org still has the venue, as far as I know, and it's still occasionally being used by the EA/rationalist/longtermist communities.  The org also still has some amount of funding, which it is using for various individual projects (e.g. I don't know if former CFAR employees like Dan Keys or Elizabeth Garrett are getting CFAR grants to run their own individual investigations and experiments, but I would not be surprised and also to be clear this would be a pretty appropriate and good thing, according to me).

There are some smaller, quieter, workshop-esque things happening from time to time, but they are more geared toward specific audiences or accomplishing narrow goals, and they are not the generalized "develop the art of human rationality and deliver that art to high-impact people" goal that CFAR used to somewhat fill.  As far as I can tell, there's a decent chance that new ambitious projects might rise from the ashes, so to speak, but they'll likely be AI-xrisk oriented.

I personally have been wishing for more clarity on all of this, for precisely the reason that I remain interested in people furthering human rationality and would like people to not be thinking that "CFAR is on it" when afaik it has not been on it since some time in 2019.  

I'm part of a small group of people who might plausibly launch new projects in that direction, and I myself am currently running one-off workshops-for-hire (am typing this from Day 3 of a workshop in the Bay Area, as it happens) for groups that are capable of pulling together participants, venue, and ops and just need the content/leadership.

Additionally, it would be interesting to hear why the endeavour was abandoned in the end, to avoid going on wild goose-chases oneself (or, in the very boring case, to discover that they ran out of funding (though that appears unlikely to me)).

I certainly cannot speak with any authority or finality for CFAR, having not been there since late 2018/early 2019.  But my sense is more like "it was always more about the AI fight than about general rationality, and the org in its evolved state was not serving the AI fight goals particularly well, so it just kinda fizzled and each of its individual members struck out on their own new path."