What the "Rationalist Community" means to me

By Arthur Milchior, . Permalink Rationalism

For years, I wanted to write about what the aspiring rationalist community meant to me. Seeing a lot of people criticizing it, sometime with argument I agree with, often with ones which does not represents the reality I've seen, I made a twitter thread about it. Threads are helpful because I find it more acceptable to write whatever comes through my mind randomly, which makes it easier to write than a blog post. I'll translate and post it here too.

The thread

What do I call "Rationalism"

I fear it might be necessary to state that I do not necessarily agree with everything aspiring rationalist wrote. I am far from having read everything. I'll add that this term seems to be used by various unrelated communities. I have sometime seen people stating "I use rationalist reasoning, so I'm right". I have not seen it in the circles I've been in and can not imagine it ever occurring.

I honestly don't know which ideas originally comes from the rationalist movement, which of those ideas are now known outside of it, what ideas were already very well known, classical, and straighforward for most people. There are of course a few exceptions, such as the paper-clip maximiser, the word of Em, many IA-risk related ideas, but I've never been able to understand why this very idea is so important, so I'm not going to consider it here.

I do love rational(ist)[1] fiction. Harry Potter and the Method of Rationality being the best known, Metropolitan Man, Instrument of Destructions, Sword of Good, I'm beginning Pokemon: the origin of species[2]... Simply, I consider it to be a genre of fiction. Not better than romance, polars, or genre fiction. It emphasizes a notion I really love, considering as true an idea, and trying to figure the consequences, however surprising they may be. It seems Science-Fiction did this a lot, but now that SF became Star Wars, 2001, A space Oddissey, is not what I'm looking for anymore. Maybe the tag "rationalist" let me find what I'd have loved as SF decades ago.

Some concepts

Let's come back to non-fiction, blog post, the "sequences", etc... I owe them, at the very least, to explain concepts in a language I can understand. If only for this, I owe a lot to Eliezer Yudkowski's writting.

Make mistake

As a first example of something important that changed for me, I know accept to consider positive that I see huge mistakes I made in the past.

I know for a fact that this advice is as classic as an advice can be. I have heard I don't know how many artists explaining that you should practice, that you should find things badly done in your past artworks. I heard stand-uppers testing their acts during a full year. I never understood. It seemed to me that it is a lack of respect for the audience. I read Letters to a young poet, which states that poems written in your first years should never been shown to anybody. I disagree because I believe it's okay for people to see mistakes. I wonder why I never understood that artists must take time to progress, since I do love seeing young artists growing, improving their notoriety and their skills. I did stand-up comedy during a decade, I've seen artists that I love becoming famous. There are artists I discovered in a cave with 10 people in the audience and then I saw them in a 1700-seat theater some years later. Some other excellent artists are still not famous at all and I'm kind of sad for them, especially for the one who left. Some other artists are still beginning and I hope to see them having success in a few years.

Let's ignore the pleasure of seeing someone try and evolve and let's move on. I heard about gradient descent for the first time in 2008. In plain English, it just means that you try something, you make a lot of little twist, keep the one improving your goal maximally, and do it again and again until you don't find a way to improve. As an example, at bowling, you can try to change the weight of the ball, whether you throw directly in front of you or try to make it rotate, and you can see what changes. If you realize that you get better with a heavier ball you take the heavy ball and then try all other tweaks again: rotation, diagonal or straight, heavier... And you automatize all of this so that the computer find the ideal choice for you, without you being here at each step to give feedback. That is obvious for the computer scientist I am that you can't know in advance the effect of all parameters of the problems. That it is mathematically interesting to consider that each parameter affect a probabilistic distribution that you want to discover, and to consider the number of try that you need to make to calibrate your model.

I would also add that I discovered a caveat that lower the importance of the error. You must consider whether failing would have huge consequence. Trying to do stand-up comedy and failing cost some money, but I was limited by the amount I had at the time, so my loss was limited. It was safer to calibrate myself and fail if I know I can't lose a lot. That's the point of fail-safe: an elevator can be in an empty corridor because at the slightest problem, there are brakes that appear. It would be almost impossibly expensive to create an elevator that never have the slightest problem, but making them fail safely allow for affordable and safe elevator.

Nowadays, I find it incredible that all of those thought were revelations. That I never applied the theory I know and the real world where I live. That it is normal and expected that my creation and ideas were less good in the past, that it is part of the research process. It may be a huge problem with me that I had to wait until someone mention the mathematical formalism in order for me to understand something that seems natural to so many artists, so many activist (outside of twitter?). I'm really thankful to this community to find the words that I could understand. As a teenager, I told that my Fursona would be a robot[3]. I knew nothing about error correcting code at the time. My ideal was Jarod, from The Pretender; I love this notion of being optimal that he shared with Poe's C. Auguste Dupin. I've heard that in SF, the "rational" character is Spoke; I don't know him, I never was into Star Trek. Since then, I learned about error correcting code, I know that robot must calibrate themselves and keep correcting errors, that the thought I had about their perfection were entirely false... all of this was probably a part of the problem. So, in my state of mind, it was unimaginable, because it was shameful, to even consider that some actions such as relaxing my jaw, changing the light or color of a lamp, taking a shower, could have an impact on my thought process. Once again, the idea is absolutely not new, it's in all self-help book, in all new-age clichés. I often read that rejecting feelings is a cause of toxic masculinity. Nothing new, except that for the first time I understood it.

Meaning of words

I can also add some tools I find useful in rationalist writings. For example, "does the tree that falls in the forest makes any sound" can be reduced to the question "does 'sound' mean 'what someone perceive' or 'a wave in air'", which, strangely, help on a lot of debate. For example, in "The thief doctor: he gave vaccine before they expired to people who do not belong in a high-priority category" seems to create a lot of heated debate. It's useful to see that people debating are actually debating on the meaning of "thief" and not on what the doctor actually did. It does not mean that arguing about words is useless, words are political tools. It's even of utmost importance to decide what a court should decide. What I have read as advise and sometime used with a partner is "rationalist taboo", and in those case, it would simply forbid me to use the words "sound" or "thief".

Another advice helped me to blog. If I can't find how to answer a question, I can consider why this question is important to me. If I don't know how to explain why something is true, I can at least explain why I believe it or why I want to write about it.

Science

I also read texts related to knowledge. But honestly, it's really rare that I use them. Probably "learn how to say 'oops' and move on" helps accepting that I failed and that I don't have to make excuses, try to panic and make things worse with damage control. Just admitting that I was wrong save time. Also, Ockham razor's "It occurred because the witch did it" is a very simple explanation and can explain everything, while the 10^80 particules state the universe should not exists. It was fun to discover that what should minimized is the number of concept, e.g. explaining how the witch act is extremely long, far longer than physics rules.

Related: in third year of undergraduate, in my first mathematics logic lecture, the prof explained to us that we are still allowed to know all maths we know, that we are not going to try to create everything from scratch (or from the empty set). I assume that it was a false idea of what Logic is that a lot of mathematical fans have. Bourbaki tried to create everything from minimal basis, that's a really fun game, but useless most of the time. However, since it's fun, most popularization texts mention it, and since we only knew logic from popularization book, we thought that it was what logician actually do. In reality, we can use all normal maths, in order to get interesting concepts, and use them to study formal concepts of model theory and proof theory. We are allowed to assume that current maths works correctly and go on doing things with them that are more interesting than painfully rebuilding everything.

Decrease the number of errors

The sequences also help me to understand that, even if it's okay to do mistakes, trying to reduce the number of mistake is cool too. That means that I should behave, taking into account who I am today, and not only who I would be ideally. That is far more practical than having a theoretical model of thoughts. And it's really hard for me to remember that it was not always obvious. A quote I saw often in 2020 told: "You're personally responsible for becoming more ethical than the society you grew up in". As an example, I realized how counter intuitive it is when I heard former French President Nicolas Sarkozy, who was asked about the ecological impact of Christmas Tree, answering that, as a child, he never felt like a monster for liking those trees. I also never considered Christmas trees as a bad thing, and that the start of the thought path, it should not have been the end.

Random list

Another cool thing I did discover, that is not created by this community, is that people lists low-hanging fruits- simple trick that helped them, that they would have loved to learn earlier. It seems to me that live discussion are better for this, since you can ask question and really understand why something is cool. Here are 2019 and 2020 European LW community lists.

Far more personally, I applied to MIRI, they offered to pay for a trip from France to California to go to a conference. I was not hired, but I spent almost three months there, in rationalist shared housing. I spent a month in decision tree, that Scott Alexander mentionned in a post. I am slightly sad I never met him nor any blogger I like (however, I met a journalist I sometime read, from Vox's future perfect). I was able to do Anki's lectures at Reach, meet people who knew my work on Anki. During those time I also interviewed at Google. I thought I failed the hiring and went back to France, while in fact, they told me a few weeks later that I was hired.

Related: I discovered Anki and AnkiDroid thanks to the rationalist community. At least thanks to Gwern post[4]. That what told me how to become a developer, read an existing codebase, contributing to it, becoming a mainteneur of AnkiDroid.

It was during Europe's LessWrong Community Week-end de Berlin (plenty of plushies) that an exercice make me question my career. I was sure I'd want to be a searcher at university and never questioned that before, never thought that I had other options. After all, after undergraduate studies, there are graduates studies, PhD, post-doc, and so on. This community week-end is also the place where I discovered cuddle-fort, with plenty of people on an ocean of mattress, with a sheet 1m50 above us. I really suggest everyone to try, that's fantastic!

Also, while learning about double crux is nice(for me at least), having a place to practice is better. Sadly, I was never able to practice since, as I don't know anyone that use it.

And that's it. I believe I mentioned the various ways why this community was important to me. That's not all of the cool stuff that I discovered here, but that show the variety of cool stuff here. I hope it shows why I love this community and feels good in it.

Notes

[1] The difference between rational and rationalist don't matter in this post.

[2] but the podcast seems to have stopped years ago, and I listen more than I read

[3] actually, I've got an hamster fursona, given by Fur Piled's author, because I was his webhamster

[4] I may need to do a contra to this post. It's very often quoted, and I disagree with multiple points here. And the user base is not 100 000 users anymore, Ankidroid itself has 2 millions users!

Add a comment

HTML code is displayed as text and web addresses are automatically converted.

Add ping

Trackback URL : http://www.milchior.fr/blog_en/index.php/trackback/766

Page top