I’ve said a lot of things on this blog over the nearly ten years since I started it. During that time, I’ve written thousands of articles [1]. Most of them directly or indirectly express my beliefs about the world. I’ve talked about psychology, learning, philosophy and which techniques I think work and which don’t.
Unfortunately, a lot of my beliefs are probably wrong. This is true of anyone, but perhaps more so of someone who isn’t a specialist researcher on any of the topics he writes about. Sometimes I get new evidence that a previous belief I had was wrong, and I make an effort to update it in a new blog post, as I did when I switched my opinion from advocating speed reading to seriously doubting its efficacy [2].
But most of the time I change my beliefs the old posts remain unchanged. Carefully combing over a thousand articles for updates in belief that may or may not their change their thesis is completely impractical. Also, often my beliefs change by a matter of degree, not a 180 flip, so it’s not possible to expressed new shades of belief without completely rewriting all the articles.
Despite the impracticality of ensuring that every article I’ve written faithfully represents my current best estimate of how the world works, people still read many of my archived posts. So there’s a tendency to think that something I’ve written eight years ago is exactly how I see the world today.
A good compromise to reduce this problem, is to simply post an article every so often which includes beliefs I’ve changed my mind about, or at least started to doubt. That way you can see the directions of my thinking over time.
Things I’ve Changed My Mind About
Here’s a few things, where I’ve changed my beliefs. These are situations where I’ve either flipped my opinion about a subject, or at least changed it considerably enough that I would have written things differently if I had to redo them.
- Speed reading doesn’t work. I used to think it helped, based on personal experience. Deeper research and longer-term personal experience now tell me it has pretty limited use.
- IQ does matter. I’ve argued in the past against IQ [3], and in particular, the lay interpretation of it being up to 80% heritable. I’ve now been persuaded that it’s one of the success stories of psychometrics [4], it predicts an enormous amount of things and there is good reason to believe much of it can be reduced to a common g-factor.
- Repetition is important for learning. I used to be against mindless repetition [5] as a way of not understanding a concept. I still think there are problems with rote memorization of concepts, but repetition is good for learning even when you are trying to understand things.
- The key to learning well is not connections but practice. I still believe holistic learning matters, but I see it more as a side-player than as the main phenomenon. Practice, in the form of active recall and problem solving, was the dominant method I used in both the MIT Challenge [6] and the year without English [7].
- Innate, immutable ability does matter for success. I’ve previously taken positions arguing against focusing on talent [8] or innate ability that we don’t have the power to change. Now I have a more moderate opinion where I see both effort and talent playing roles.
- Long-term perfectionism has dangerous side effects. I’ve previously argued in favor of being a long-term perfectionist [9]. While I still think it’s important to strive for continual growth, I’ve since become better informed on the research on perfectionism and how it often has effects which go far beyond not getting your work out there.
- I don’t believe free, online education will replace universities. I think challenges like the MIT Challenge will probably stay niche. Computer science is a possible exception, but for the purposes of employment, coding bootcamps are more efficient than a comp sci degree.
- I believe much of formal education is signalling. Signalling [10] is the theory that people get useless educations to signal their ability and conformity. It’s contrasted to human capital theories where people gain useful skills from school. Having done one real degree and one simulated degree now, I believe there are many interesting things to learn, but they probably don’t improve labor productivity in most cases.
- I believe it takes much longer than 30 days to form a habit. Thirty day trials are still probably useful for experiments, but if you are certain you want a sustained habit, the real interval may be more like six months.
- Vegetarianism alone is probably not the healthiest diet, or necessarily the simplest constraint for optimizing health. Simple carbs and sugars appear to be greater enemies than animal products in the quest for health. This doesn’t mean a vegetarian diet isn’t healthy, just that if you were picking constraints on eating it might not be the first choice. It also doesn’t mean the ecological and ethical arguments for vegetarianism are invalid. I’m currently pescetarian [11], and have no plans to change this.
- Asian languages are much harder to learn than European ones. I never held the belief that they were equally difficult, but setting up for the year without English I underestimated how big the gap was. I think 9-12 months would have been acceptable for Vat and I to reach the same level in Korean or Chinese as we did in Spanish.
Things I’ve Been Questioning Lately
In addition to those beliefs that I’ve gathered enough evidence on to change my mind, I’ve also had a lot more open questions—areas I hadn’t thought about deeply or beliefs I have that I’m currently feeling some doubt. These are areas where I see myself easily being persuaded in either direction, so I have less certainty about them.
- Is anxiety necessary for success? I’ve written about this recently [12]. I unfortunately chose to use the word stress in that article (rather than anxiety), but the argument is otherwise the same.
- Is willpower or energy a depletable resource? I’ve long been a fan of energy management [13]. Although the concept, is more self-help than hard science, I saw a connection to Baumeister’s work on ego depletion [14]. Now it looks like that theory is under some serious scrutiny [15].
- Is a growth mindset actually helpful for success? This one also seemed obviously true to me, but Scott Alexander raises some good critiques [16]. With the dismal replication performance [17] of psychology experiments in the news lately and a lot of the experiments coming from one group of researchers, I’m less confident of this than I was before.
- Are a lot of “good” habits just signalling attempts? Maybe I’ve just been reading too much Robin Hanson [18], but he does provide a convincing explanation of why we often don’t things we think we should. The reason being, that we don’t actually want to do those things, but want to seem like we do, so we make striving to do those things part of our conscious self-presentation but subconsciously sabotage ourselves in actually doing them. Some possible examples: reading more books, not watching television, giving to charity.
- Does free will exist? Does the self? This is one I’ve gone back and forth on. I began straightforwardly believing that free will doesn’t exist, since the universe is either deterministic or random. Then believing in a compatabilist definition [19] which redefines free will to be a useful concept if not having any cosmic significance. Now, rereading Buddhist and Taoist ideas, I’m wondering whether these ideas are even necessary, given their seemingly many self-contradictions.
The Perils of Intellectual Consistency
I’m sure some of these ideas I’ve switched on, including ones I advocated strongly for, have probably surprised a few long-time readers. People like to be able to group the people they know in a consistent fashion, saying I’m the person who believes in X or writes about Y. If I suddenly express a view saying I believe in not-X, that’s understandably confusing and upsetting.
The other reason I believe we value consistency is that we equate consistency with competence. People whose beliefs flip back and forth are those who don’t have much experience in the matter. In theory at least, gaining more experience and research should drive you closer to the truth of the matter. In this way, those who are inconsistent are also usually ignorant and incompetent.
That may be the case, but I suspect that trying to defend consistency invites a worse sin, where you ignore evidence that doesn’t fit your preexisting worldview. I’d rather be ignorant in appearance than dishonest in actuality.
I hope that this public update of my beliefs will be a regular occurrence on the blog, something I do every year or two, capturing the changes in my best understanding of the subjects I write about that don’t make there way into full articles. I hope it also shows that learning is a messy business, reinforcing some beliefs while casting doubt on others. The path to the truth is rarely a straight line.