On the one hand, job testing is cool and I personally would enjoy having more opportunities to occasionally try out very different kinds of work, just to see if I would like it. This kind of experience seems under-supplied in normal life.
On the other hand, taking a multi-week course of basically different homework tasks, with no big payoff at the end, seems like a big commitment with correspondingly big opportunity cost. Personally, I wouldn’t sign up for a course like this even though I like the idea—maybe I’d get around to it after I finally get around to going through that textbook on C++ coding to brush up my software development skills, or maybe after I do that deep-dive of learning about the finer points of index investing that I’ve been putting off...
I can imagine a couple of variations on this idea:
Take it in a less-intensive direction, by dropping most of the “work trial” focus and just having virtual programs that cover 80K’s material. Maybe we just take the existing EA virtual programs (which are mostly focused on learning about EA cause areas), and nudge them in a more explicitly career-choice focused direction instead of just an intellectual / study-group structure. The more advanced programs seem like they already do this a little, but I could imagine tilting even the introduction-to-EA materials towards 80K style career planning. This might be an especially good fit since a lot of students (who are right in the middle of deciding what to do with their lives) sign up for EA virtual programs. It would be hard to really do serious work tasks on such a broad range of areas, though, so this would look more like “covering 80K material during a virtual program”. But maybe you could still come up with some helpful trials, and then at the end you could help connect people with mentors and potential employers in their area of interest.
I would say something like “maybe instead of work trials pitched at individuals we should be exhorting EA organizations to step up their work-trial game”. But actually, I recently applied for a job with an EA organization, and their work-trial game was top-notch! I was seriously impressed at how much I felt like I got a detailed feel for what it would be like to take on the role. So fortunately I think that (at least some) EA organizations have this base covered!
On the one hand, job testing is cool and I personally would enjoy having more opportunities to occasionally try out very different kinds of work, just to see if I would like it. This kind of experience seems under-supplied in normal life.
On the other hand, taking a multi-week course of basically different homework tasks, with no big payoff at the end, seems like a big commitment with correspondingly big opportunity cost. Personally, I wouldn’t sign up for a course like this even though I like the idea—maybe I’d get around to it after I finally get around to going through that textbook on C++ coding to brush up my software development skills, or maybe after I do that deep-dive of learning about the finer points of index investing that I’ve been putting off...
I can imagine a couple of variations on this idea:
Take the idea in a more-intensive direction, as a coaching/accelerator program for key fields: Especially for highly-talented technical people thinking about AI safety research, their time is valuable! How are you going to make it worth their while to take an unpaid, structured class working a bunch of example problems and tests? Maybe by giving them lots of 1-to-1 guidance with experts in the field, and helping them refine research ideas for their own unique project, and then having everyone present their projects at a Ycombinator-style “demo day” where they get matched with interested employers or awarded big grant-money prizes. (See that linked google doc for more details.)
Take it in a less-intensive direction, by dropping most of the “work trial” focus and just having virtual programs that cover 80K’s material. Maybe we just take the existing EA virtual programs (which are mostly focused on learning about EA cause areas), and nudge them in a more explicitly career-choice focused direction instead of just an intellectual / study-group structure. The more advanced programs seem like they already do this a little, but I could imagine tilting even the introduction-to-EA materials towards 80K style career planning. This might be an especially good fit since a lot of students (who are right in the middle of deciding what to do with their lives) sign up for EA virtual programs. It would be hard to really do serious work tasks on such a broad range of areas, though, so this would look more like “covering 80K material during a virtual program”. But maybe you could still come up with some helpful trials, and then at the end you could help connect people with mentors and potential employers in their area of interest.
I would say something like “maybe instead of work trials pitched at individuals we should be exhorting EA organizations to step up their work-trial game”. But actually, I recently applied for a job with an EA organization, and their work-trial game was top-notch! I was seriously impressed at how much I felt like I got a detailed feel for what it would be like to take on the role. So fortunately I think that (at least some) EA organizations have this base covered!