Time spent: Roughly 3 hours.

Epistemic status: My opinions and observations, all my own

tldr It might be easier than you think to do direct work in EA if you put just a little time in looking for good projects. EA Public Interest Technologists recently ran an event trying to find such opportunities, and found quite a lot with little effort.


I help run several EA communities, particularly EA Public Interest Technologists and EA Melbourne. One of the very common ideas people have is that it's difficult to get into doing direct work for EA. Probably born from posts such as this one, which are very commonly cited, as well as general ideas that EA is very elitist, or that good things like direct work for EA can't come without sacrifice.

I noticed this problem with my perspective on AI safety. I’ve had the belief that "It's really difficult to find concrete problems in AI Safety" without even trying to find concrete problems in AI safety. However, when I take maybe half an hour or so to see what's actually there, I have much more than I could ever hope for.

It reminds me of Eliezer’s concept of 5 minutes by the clock. Where people often don’t spend the time to identify alternatives to their current beliefs, and very quickly jump to the conclusion that “Why yes, I’ve searched for alternatives, but there weren't any”. The steps where we often fail is not within the process of searching, but just simply choosing to search.

I believe this is definitely the case with other EA direct work as well, there's more than enough interesting direct work to do if you just spend some time looking for it. As of such, EA Public Interest Technologists ran a working bee to create a list of technical EA projects.


To start our working bee off, we used the columns and data from Edo Arad's collection of research questions, and whittled them down until only the technical projects were left.

The working bee went for an hour, and the four participants were given the following resources to start off in finding projects:

In the end, we created a spreadsheet of interesting technical projects. The spreadsheet is quite small (only 15 entries currently). All participants found EA direct work to follow up on. Edits and additions are welcome.

If you want an extra push or don’t really know where to start, I’d be happy to point you in the right direction. If you want technical work: Join the EA Public Interest Technologists Slack.


There are two takeaways from this. The first is that if you think that direct EA work is not for you, think again and have a look through the different projects that the community is having a go at! I recommend looking through our spreadsheet as well as the links above.

The second is that if you are interested in direct technical work, please do join us at EA Public Interest Technologists. We are having chats about both (very) low hanging fruit for technical projects as well as others that are more technically interesting. We are on Slack. We'd love to see you there.

23

New Comment
2 comments, sorted by Click to highlight new comments since: Today at 2:13 PM

Great post! I think this kind of list is really valuable. Just a quick note that the link to Michael Aird's central directory of research questions points to the wrong place, here's the right link.

Thanks for pointing that out! I just fixed it up.