[ Question ]

Is learning about EA concepts in detail useful to the typical EA?

by Linch 6d16th Jan 20201 min read9 comments

33


Say someone already spent 10-20 hours acquiring basic EA knowledge. Maybe they read Doing Good Better and the 80000 Hours career guide, maybe listened to a few 80,000 Hours podcasts. Is learning more about EA (eg, reading this forum) helpful to them?

Here are some guesses for various roles I brainstormed:

  • Safety engineer at OpenAI/DeepMind. You should probably know a few high-level things about AI Safety (maybe some about longtermism as a whole), but beyond that, ML, software engineering, and general productivity skills seem to matter much more.
  • Earning to Give. Unless someone really likes and is really good at thinking through and applying these concepts, there doesn't seem to be much value in learning more about EA, since donating to the EA Funds of your chosen cause area (maybe the donor lottery) is probably higher expected value than trying to pick out donation opportunities yourself.
  • EA Career coach. Having broad knowledge of EA seems valuable.
  • Animal rights activist. You should probably have some broad knowledge of expected value and what interventions work in the effective animal activism literature, but presumably most of your learning time is better spent networking/learning from other activists.
  • Developmental econ researcher. Maybe EA can help you prioritize research questions, but mostly I just don't see the added value of learning about EA relative to normal dev econ tools?
  • Cause prioritization researcher. Having broad and deep knowledge of EA seems very valuable.
  • Community builder. Probably a good idea to have a broad knowledge of both community building models and for individual cause areas (so you can help advise members).
  • American Politics/Policy practitioner. Doesn't seem like EA at the moment adds much beyond normal skills in the policy toolkit. Might be helpful to network with EAs so things in the grapevine can eventually reach you however.
  • AI Policy researcher. EA tools seem valuable (low confidence).
  • Grantmaker in EA-heavy field. Having broad and deep knowledge of EA seems very valuable.
  • Journalist. Having a broad knowledge of EA seems valuable.

Naively, it looks like most roles that individual EAs could be in does not, at this moment, benefit from substantial EA knowledge. So for most of us, the main benefit of learning more about EA is something more nebulous like entertainment, "unknown unknowns", or "feeling more connected to the community." Am I missing something major?


33

New Answer
Ask Related Question
New Comment
Write here. Select text for formatting options.
We support LaTeX: Cmd-4 for inline, Cmd-M for block-level (Ctrl on Windows).
You can switch between rich text and markdown in your user settings.

5 Answers

Yeah, I think you're missing the flow-through effects of contributing to the EA hivemind. There's substantial value in having a large number of reasonably well-informed EAs thinking about and discussing EA ideas. It's possible this is insignificant compared to the contributions of a handful of prominent full-time thinkers, but seems like an important and nontrivial question in its own right.

To answer my own question, I suspect a lot of this comes from EA still being young and relatively pre-paradigmatic. A lot of valuable careers and projects seem to be in the timescale of not decades but rather years or even months, so keeping very up-to-date with what the hivemind is thinking, plus interfacing with your existing plans/career capital/network, allows you to spot new opportunities for valuable projects to do that you otherwise may not have even considered.

I suspect that as EA matures and formalizes, the value of following current EA thinking becomes less and less fruitful for the typical EA, and engagement (after some possibly significant initial investment) will look more like "read a newsletter once in a while", and "have a few deep conversations a year", even for very serious and dedicated EAs with some free time.

I think the important questions are "what type of information?" and "what would they be doing otherwise?"

If you're interested in development economics and the choice is between reading everything posted on the Forum and taking a micromasters course in data and development*, I'd suggest the latter. But if it's a choice between scrolling Facebook and learning about EA, learning about EA is probably more useful.

*https://micromasters.mit.edu/dedp/

If you knew more about EA, this would allow you to make decisions that you believe to be better informed or aligned with your values. How do you decide where to donate and what to do with your career? Not just which charities within a cause (you could trust the grantmakers for various funds for that, but you might disagree with them, too, if you knew more), but which causes, too?

You might disagree with the values or arguments that lead to certain recommendations, and find others more compelling, but you need to be aware of them first.

My initial thoughts are similar to those of Adam and Linch, but I'll post them here anyway, despite possible redundancy:

  • If someone learns about an EA concept in detail, they might be able to use that to generate unique insights or follow-up, even if many other people already knew a lot about that concept. Reasons this might happen:
    • They are more up-to-date on the concept than others who learned about it a while ago (e.g. they've read very recent studies related to the concept)
    • They bring unique "outside knowledge" to the concept (e.g. they have a background in sociology, which none of the other people who knew about the concept had)
    • They have the time/skill to write out what they know in a way that a much wider audience can understand without much effort, which wasn't the case for the others
    • They have access to funding/other resources that the others didn't have (e.g. they can pay for a massive survey to gather direct data that furthers their knowledge, or they're in touch with experts in the relevant field who can help to deepen their knowledge)
  • Learning about concepts in detail seems important if someone might pursue a career in one of several fields. I'd want them to make such a decision based on a lot of firsthand information about each field, in addition to whatever shared resources the EA community can contribute (as personal characteristics and background play a big role in how well someone's career will go in a given field)
  • Someone who learns about a concept in detail might wind up explaining it (in less detail) to people they know personally who might otherwise never learn about it. Even if the community provides many "authoritative" sources of written knowledge, it's still hard to replace the benefit of having someone you know explain something to you and answer your questions in real time.
    • Similarly, they might be able to persuade other people with their combination of knowledge + personal connection. Many members of this community are only here because someone they knew persuaded them to get involved, and I'm guessing that "go read this article" is less persuasive than "let me tell you about this exciting thing that I'm clearly well-informed about"
  • Learning about things in detail builds good epistemic habits related to understanding how research works and so on. The more people in EA have good epistemic habits, the less likely it is that we'll miss something obvious or be vulnerable to bad arguments, scams, etc.