Calibrating your humility

Last year, I did a six-month work placement at a large consulting firm, and in my first few weeks there, I messed up many tasks in a wide variety of ways.

In my interview, my manager had warned me that the biggest problem they have with interns is that they pretend to understand things and don’t ask enough questions. I made a mental note of this, resolved to not be that kind of intern, and proceeded to do it anyway except in a way where I felt extra guilty about it. This was obviously a horrible move, but one I found surprisingly difficult to avoid.

One piece of this is clearly impostor syndrome—the fear that your manager will realize you are less competent than they thought, and regret hiring you, or maybe even fire you on the spot (the latter is almost always unfounded—firing people is a Big Deal™). But another is that it’s genuinely hard to be well-calibrated about when to use your own judgement when you have no experience and are the junior-est junior on your team.

For context, my team was way too overworked for anyone to have the time to walk me through all the different terms and calculations they use, so my training was basically reading a 50-page how-to PDF and then giving actuarial analysis a go.

After submitting my first report, my manager noted that I had left in some things from the report template that weren’t applicable to the specific company I was working on. What had happened was, I didn’t understand what those things meant and just deferred—they must have made the template say that for a reason, right?—but she said that was bad and I shouldn’t write anything in my reports that I couldn’t justify. Which, like, fair enough.

Then, in the second report I submitted, I got the opposite feedback—I deleted some stuff I didn’t understand, and my manager wrote “this isn’t the wording from the template” in big red letters next to where it was missing. And it wasn’t just reports—there were other projects where I, for instance, struggled with a task for hours and, when my manager found out, was told I needed to ask more questions; then asking a question, and my manager saying she knows I can figure it out on my own.

This was frustrating and embarrassing, but at least common, which I knew because of my manager’s comment during my interview. And the only strategy I had to get better was basically gradient descent. This was a fairly painful process, but fortunately kind of worked, given the pretty immediate feedback loops (and to any potential hiring managers reading this: I was slaying by the end).

The implications

The problem is, these kinds of feedback loops are less clear or nonexistent outside of structured environments like work. That is when this phenomenon is probably at its worst—for me and other young people trying to “do projects” and “have ideas”, it is really hard to be well-calibrated about when to trust our judgement versus defer, when to ask for help versus try harder, when we need mentorship versus are ready to venture out on our own.

We have all read the occasional post by a cocky young person presenting a known idea or criticism as if they were the first to think of it—we have also all heard the criticism that EAs defer too much to “EA elites”, and in our personal lives, we’ve all probably made a mistake related to being miscalibrated about something along these lines.

So, I’ve been trying to come up with rules of thumb for this. Here are the ones I’ve thought of so far:

Some Rules Of Thumb for Humility

1. If you think of:

   1. A research question

   2. A possible solution to a problem

   3. A criticism of a popular idea

   4. An idea for an app/website/startup/blog/institute/project

   5. Or some thing else along these lines,

   …stop and do a literature review first. There’s a high chance you’re not the first to think of it, especially if it took you 3 seconds to come up with. This is true roughly in proportion to how young you are or how new you are to a field. (Also, there is unironically ~so much alpha~ in actually reading papers and becoming deeply familiar with the academic literature on a subject. And also in googling your quandaries in general).

2. It is common for people in a given field to overestimate how much people typically know about their field, or to otherwise underestimate the inferential distance between them and the average person. If they assume you know some term or Excel shortcut that you don’t, there is a high chance this assumption isn’t reasonable, and they might realize that right after you ask, so don’t feel stupid for asking.

3. When you are assigned a task at work, ask how long the person giving it to you expects it to take. If it starts looking like it will take a wildly different amount of time, reach out.

4. How much you should defer to experts/more knowledgeable people is roughly proportional to how much consensus there is among them. If they all agree on something and you don’t, there is a high chance that it’s because you’re missing something. If there is a lot of controversy, it makes more sense to go with your own judgment.

5. For long-term independent projects, set regular feedback checkpoints, such as:

   1. Every 50 hours of work

   2. After completing 30% of the project

   3. If you’ve changed more than 20% of the content

See also: Chesterton’s fence, cluster thinking, the LessWrong tag for “principles”, this thread of rules of thumb.

---

Postscripts

This blog is one week late. I think I have a good justification for this—I have second-degree burns on both of my hands that have made it very difficult to type or really do anything for the past week. However, this happened right before this blog post was due so I do think I am at least partially at fault. Therefore, I am still paying the financial penalty I agreed with Max:

---

Links to stuff I found interesting recently

• The Aesthetics Wiki. What it says on the tin. If you like Twitter user @coldhealing, you will like this one. Even if you don’t, it’s a good bit of modern anthropology.

• IsItNormal.com, a website where people post things they do and others vote on if it’s normal. Really says a lot about society.

• Buddhist economics. Did you know that’s a thing?

• Video update on those houses in Italy that were selling for €1, if you were wondering how that played out.

• The 80,000 Hours episode on moral weights. Honestly the 80k podcast has been putting out extra interesting episodes lately, and this is an especially good one.

• The 2023 “Mental State of the World” report. I had never heard about this before and am not really sure what to make of it, but it has different findings from the World Happiness Report, so it’s interesting.