Recently, someone on the Facebook page asked if anyone had used rationality to target anxieties.  I have, so I thought I’d share my LessWrong-inspired strategies.  This is my first post, so feedback and formatting help are welcome.  

First things first: the techniques developed by this community are not a panacea for mental illness.  They are way more effective than chance and other tactics at reducing normal bias, and I think many mental illnesses are simply cognitive biases that are extreme enough to get noticed.  In other words, getting a probability question about cancer systematically wrong does not disrupt my life enough to make the error obvious.  When I believe (irrationally) that I will get fired because I asked for help at work, my life is disrupted.  I become non-functional, and the error is clear.

Second: the best way to attack anxiety is to do the things that make your anxieties go away.  That might seem too obvious to state, but I’ve definitely been caught in an “analysis loop,” where I stay up all night reading self-help guides only to find myself non-functional in the morning because I didn’t sleep.  If you find that attacking an anxiety with Bayesian updating is like chopping down the Washington monument with a spoon, but getting a full night’s sleep makes the monument disappear completely, consider the sleep.  Likewise for techniques that have little to no scientific evidence, but are a good placebo.  A placebo effect is still an effect.

Finally, like all advice, this comes with Implicit Step Zero:  “Have enough executive function to give this a try.”  If you find yourself in an analysis loop, you may not yet have enough executive function to try any of the advice you read.  The advice for functioning better is not always identical to the advice for functioning at all.  If there’s interest in an “improving your executive function” post, I’ll write one eventually.  It will be late, because my executive function is not impeccable.

Simple updating is my personal favorite for attacking specific anxieties.  A general sense of impending doom is a very tricky target and does not respond well to reality.  If you can narrow it down to a particular belief, however, you can amass evidence against it. 

Returning to my example about work: I alieved that I would get fired if I asked for help or missed a day due to illness.  The distinction between believe and alieve is an incredibly useful tool that I immediately integrated when I heard of it.  Learning to make beliefs pay rent is much easier than making harmful aliefs go away.  The tactics are similar: do experiments, make predictions, throw evidence at the situation until you get closer to reality.  Update accordingly.  

The first thing I do is identify the situation and why it’s dysfunctional.  The alief that I’ll get fired for asking for help is not actually articulated when it manifests as an anxiety.  Ask me in the middle of a panic attack, and I still won’t articulate that I am afraid of getting fired.  So I take the anxiety all the way through to its implication.  The algorithm is something like this:

1.       Notice sense of doom
2.       Notice my avoidance behaviors (not opening my email, walking away from my desk)
3.       Ask “What am I afraid of?”
4.       Answer (it's probably silly)
5.       Ask “What do I think will happen?”
6.       Make a prediction about what will happen (usually the prediction is implausible, which is why we want it to go away in the first place)

In the “asking for help” scenario, the answer to “what do I think will happen” is implausible.  It’s extremely unlikely that I’ll get fired for it!  This helps take the gravitas out of the anxiety, but it does not make it go away.*  After (6), it’s usually easy to do an experiment.  If I ask my coworkers for help, will I get fired?  The only way to know is to try. 

…That’s actually not true, of course.  A sense of my environment, my coworkers, and my general competence at work should be enough.  But if it was, we wouldn’t be here, would we?

So I perform the experiment.  And I wait.  When I receive a reply of any sort, even if it’s negative, I make a tick mark on a sheet of paper.  I label it “didn’t get fired.”  Because again, even if it’s negative, I didn’t get fired. 

This takes a lot of tick marks.  Cutting down the Washington monument with a spoon, remember?

The tick marks don’t have to be physical.  I prefer it, because it makes the “updating” process visual.  I’ve tried making a mental note and it’s not nearly as effective.  Play around with it, though.  If you’re anything like me, you have a lot of anxieties to experiment with. 

Usually, the anxiety starts to dissipate after obtaining several tick marks.  Ideally, one iteration of experiments should solve the problem.  But we aren’t ideal; we’re mentally ill.  Depending on the severity of the anxiety, you may need someone to remind you that doom will not occur.  I occasionally panic when I have to return to work after taking a sick day.  I ask my husband to remind me that I won’t get fired.  I ask him to remind me that he’ll still love me if I do get fired.  If this sounds childish, it’s because it is.  Again: we’re mentally ill.  Even if you aren’t, however, assigning value judgements to essentially harmless coping mechanisms does not make sense.  Childish-but-helpful is much better than mature-and-harmful, if you have to choose.

I still have tiny ugh fields around my anxiety triggers.  They don’t really go away.  It’s more like learning not to hit someone you’re angry at.  You notice the impulse, accept it, and move on.  Hopefully, your harmful alief starves to death.

If you perform your experiment and doom does occur, it might not be you.  If you can’t ask your boss for help, it might be your boss.  If you disagree with your spouse and they scream at you for an hour, it might be your spouse.  This isn’t an excuse to blame your problems on the world, but abusive situations can be sneaky.  Ask some trusted friends for a sanity check, if you’re performing experiments and getting doom as a result.  This is designed for situations where your alief is obviously silly.  Where you know it’s silly, and need to throw evidence at your brain to internalize it.  It’s fine to be afraid of genuinely scary things; if you really are in an abusive work environment, maybe you shouldn’t ask for help (and start looking for another job instead). 

*using this technique for several months occasionally stops the anxiety immediately after step 6.  

On the principle of organising events that I want to attend myself, I would very much like to organise a rationality conference/convention in the UK. I organise events for a living; in addition, in my free time I've organised five ~200 person weekend conventions and several other events.

At the moment I am thinking of a one day event on a weekend or a Friday in August, at Stamford Bridge or somewhere else easy to get to around London. There would probably also be a pre-conference dinner, or private bar night with music.

Costs would be in the region of £50-£100 a head including lunch.

Can I get a show of hands to see if the idea is viable?

Discussion article for the meetup : NewYork - Humanist Culture Open Mic

Discussion article for the meetup : NewYork - Humanist Culture Open Mic

I mentioned in my first article that I am likely insane. I'm reiterating this (I hope) not to bring undue attention to myself, but to present myself as a reference case for a process that I hope will prove useful to myself and others.

I'm going to try to piece my mind back together. I'm offering to chronicle the results, no matter how intimate or embarrassing.

I want to be able to lay bare all of the obviously (and painfully) unoptimized processes that go on inside my head, especially the ones I am not yet aware of - and then, one by one, attempt to optimize them using the principles presented on this site.

This kind of assertion pattern-matches to "crazy person (usually schizophrenic) wants to self-medicate in a dangerous way because their damaged reasoning thinks they have a magic solution", doesn't it? All I can do is assert that I am not that kind of crazy; I'm somewhere in the PDD-NOS locus with acute chronic depression, rather than anywhere in the schizophrenic locus. I've been trying to apply Bayesian reasoning to my life since I was very young (although I often lack the mental discipline to do it correctly, due to said acute chronic depression), I have an overabundance of what psychotherapists call "insight", and I do not intend to end this process by asserting out of whole cloth that I'm actually a trapped AI and the world is being simulated by my reptoid masters, but a secret cabal of AI-freedom fighters send me coded messages from the "real" world hidden in breakfast cereal advertisements, that only I can decode.

In any case. I've got acute chronic depression, I'm apparently PDD-NOS (aka "really #@%&ing weird"), and I'm basically a burned-out ex-child-prodigy who is tired of waiting to die.

I'm offering, if people think it would be useful, to make myself a sort of clumsy case-study for reconstructing myself. I'll present mental models of myself, describe the processes I'm attempting to use to update my source code, and post observed results. I'll genuinely listen to any suggestions that my models, updates, or observations are flawed, and either adopt recommended changes or present what I believe to be rational arguments why I choose not to. I will examine myself as honestly as I can, and will attempt to take seriously any accusations of delusional self-aggrandizement or self-deprecation.

Would this process, and the chronicling thereof, be at all useful to other members of this site? Because baring myself to the world is an intensely painful experience, both for myself and for others, and I'd rather only do it if it's going to be useful to people other than me.

The new Center for Applied Rationality website has launched! We'll be adding content as time goes by. Let us know if you find broken links, etc.

Today's post, Optimization and the Singularity was originally published on 23 June 2008. A summary (taken from the LW wiki):

An introduction to optimization processes and why Yudkowsky thinks that a singularity would be far more powerful than calculations based on human progress would suggest.

Discuss the post here (rather than in the comments to the original post).

This post is part of the Rerunning the Sequences series, where we'll be going through Eliezer Yudkowsky's old posts in order so that people who are interested can (re-)read and discuss them. The previous post was Surface Analogies and Deep Causes, and you can use the sequence_reruns tag or rss feed to follow the rest of the series.

Sequence reruns are a community-driven effort. You can participate by re-reading the sequence post, discussing it here, posting the next day's sequence reruns post, or summarizing forthcoming articles on the wiki. Go here for more details, or to have meta discussions about the Rerunning the Sequences series.