“He was born with a gift of laughter and a sense that the world was mad.” —Raphael Sabatini, opening line of Scaramouche (1921)
I have nothing against laughter. However, the thoughts this blog will be based around have evolved from the gradual finding that, somewhat unfortunately, the world is not mad at all.
The world is not simple. The world does not always work in the ways we want it to. The world frequently looks mad at first impression. Yet, the world has reasons for what it does, oftentimes surprisingly compact rules. In discovering the secrets of this sane world, a primary instrumental goal will be to compactify the rules down as much as I can. We want not, in a full post, to find and memorize that 20% of advertising has 80% of the effect. We want to discover and remember Pareto’s Principle. We want to learn what power laws govern and how to discern them.
Further, I claim that viewing the world as not mad, but in fact, quite sane, will allow us a significant amount of leeway toward discovering the underlying rules that make the world go round. I will examine many cases, spanning transhumanism, economics, math, social justice, birding, movie heroes, philosophy, physics, politics, and more, to ensure a wide range of examples we can draw from (the two special foci will be on human interaction and on mathematical structure). The tool we use to discover these rules will be “technical rationality”.
Before you run away, this is not Spock’s version of rationality so often parodied, but the beautiful flower blooming from the seed of a single commandment and its corollary:
End: achieve your goals.
Means: have your beliefs about the world reflect, as closely as possible, the world’s actual state.
Brief acknowledgment: this collection of writings comes down from a long line of rationality-based blogs, of which I think I can say that Less Wrong is the centerpiece. The genealogy of this rationality blogosphere is complicated, but luckily a map was made by the venerable Scott Alexander of Slate Star Codex, who will be either the most or second-most cited author in all of my writings. I cannot do them justice for the seminal influence they’ve had upon my life. But as in all cases, I stand on the shoulders of giants.
Less Wrong was a pioneer of instilling rational ways of thinking and developing the methods of rationality as a whole (many of the best blogs posts are now published as an e-book, and come out in book form in a few short months; as a warning, I do note that Less Wrong was integral to forming the methods of thinking that generated this blog, and you will have a significant advantage in understanding here if you first read through some of the Sequences there, or at least post-surf a little bit by following the links around). Slate Star Codex is one of the best blogs for applying rationality, discerning well-hidden truths and beautifully tying things together with general principles that are used often.
I hope to build on top of these a compactification of rationality applied to the world, and further a compactification about the rules of the world.
(This is not to say that making things simpler is always good. In most cases, it is not. I refer instead to the idea that finding patterns within things allows us to express more nuance in simpler terms. Take my word now that I will not be a strawman about this, and in the Structure post I will expand and clarify).
The title was chosen to reflect the fight against this twofold enemy to knowledge. First, we hack, slash, and fire at the figurehead villain, that Joker: falsity, the spreading of misinformation and painful instances where the map has malevolently omitted a chasm reaching before us. Where current arguments may be sometimes right, usually right, almost always right… I hope to promote a more formalized way of thinking, in many circumstances, that precludes a wrong conclusion. Not less wrong, but never wrong. I want surety in our conclusions, and that requires them to be probabilistic when we are considering the world. That requires them to admit the potential for error, as nothing but formal logic and mathematics can truly be built on only tautologies. To have surety that we haven’t missed anything, we need to formalize our arguments and identify points at which our reasoning or data could be flawed. Case in point: Nick Bostrom’s simulation paper, which we’ll discuss in detail in a later post. A common rationalist technique is steelmanning:
“Steelmanning is the opposite of strawmanning. Strawmanning takes a strong opposing argument and converts it into a weaker version to avoid having to engage with the main points. Steelmanning takes a weak opposing argument and converts it into a stronger version to ensure you’re engaging with the underlying ideas as seriously as possible.”- Scott Alexander
Steelmanning does countless wonders for discerning the truth about a claim. But too many treatises, essays, posts, books, and messages only treat one claim, one side of the argument. In determining truths about the world, we want to go above: steelmanning both sides, creating a fair arena, and allowing the knights to do battle. Woe to the vanquished… except that surety requires we always are realistic about the chance that one side might have gotten lucky.
Second, we slowly whittle away that shadowy figure in the backdrop, the Two-Face who may appear as a friend to knowledge at many times: the extraneous information, the unstructured dump of facts, the vast troves of encyclopedic knowledge. While the internet has allowed us access to a great array of information, the vast majority of it is extraneous, redundant, or disorganized. Even the plants in a garden must be thinned, for it is hard to cultivate a mound of seed. The goal here is to structure our knowledge so that it admits a compactification, a ZIP compression of our understandings, a way to reduce the barrage of facts into a consistent, general, powerful view of the world. In the spirit of effective altruism, in which the goal is to do the most good per dollar… effective information processing, to do the most good per minute spent by the reader, seems similarly important in certain contexts (do not recoil in horror at a perceived slight against art or humor).
At the meta level, this is the first step of a structure which will reoccur in many cases across this blog: abstract-level claim (just made), object-level analysis, abstract-level generalization, and object-level applications. Too few object<->abstract jumps are the reason that many sources of information fail to impart a full understanding. Many of you may have taken a math course…
Some of you may already see why these two principles of structure and surety are motivated, and worthy of ascendance to guiding precepts. Others, not. Then, in the next few posts, I will delve into object-level situations in which we extract some benefit when proceeding by their torch.
