It seems that by default, most people want to do good in their career. Doing useful work for society is preferable to not doing something good for others.
It also seems that we have a natural preference for doing more good than less. Doing very useful work for society is preferable to doing minimally useful work, all other things equal.
Naturally you have to ask,
Asking this question is the heart of the philosophy and community of Effective Altruism (EA). The philosophy is about more than career choice, but that’s what I’ll focus on here.
I don’t like to write new explanations for things when great ones exist, so watch this instead:
The rewards from most inputs follow a pattern of marginal returns: the 100th person working on something typically has a much larger marginal impact than the 10,000th, for example. It seems in general that there are potentially massive differences, 100x or more, between the altruistic-impact of one career versus another.
For example, here’s the difference in effectiveness for five different strategies of reducing HIV/AIDS spread:
(Source) (DALYs are a measurement of marginal healthy years of life added by the intervention.)
The least cost-effective intervention, too small to show on the graph, is 1,400 times less effective than the most effective intervention. This isn’t particularly unique to HIV/AIDS spread anyway (per the video embedded earlier, for example).
It seems in general that some options might simply be hundreds, if not thousands of times more effective than others.
Thus career choices may closely determine whether our impact is x, or 100x—or perhaps 10,000x.
First of all, I know I would be most productive in science/ tech/ engineering/ math areas. I also want to do research—as far as I can tell with my inclinations, this is exactly my thing. (Note: I want to do research in industry, and not in academia.)
For now, I’m aiming at working on brain-computer interfaces in my career.
I’m particularly interested in working on neural signal processing. (Translating raw neural firing data into intention.) This is the most mathy part— or at least as far as I was told by a Neuralink employee.
I’ve picked this somewhat arbitrarily out of a few other potentially high-impact research options. This one just feels the most viscerally interesting for whatever reason, so I’m going with it.
Optimistically: develop safe BCIs → enhance human cognitive capability and improve our ability to solve problems → solve other hard problems. BCIs as a ‘meta-problem’ solution.
If you have ideas for or contributions to my philosophy here, please reach out. Also, if this led you to rethink your path, let me know! I’m interested.
And if you think this summary could be better, let me know— I like feedback.
Also, six of the six people I’ve shared this article with never read it. I imagine that they saw its length, thought to themselves ”I‘ll do this later’, and thus never read actually it. To avoid this fate yourself, I recommend starting with this podcast (also mentioned in the Key Ideas article), as it’s less overwhelming.
Here is another general area that has extremely high expected value for the future of humanity.
2020 July (posted) - 2020 November 2 (last updated)