You're smart, want to help the world and are willing to work hard. You have no serious ties such as children or a marriage that would prevent you from making serious changes to your life, and you are willing to place others needs ahead of your own hedonistic desires. Given this, what should you do?
Should you aim to get involved personally with causes you feel passionately about? You can have greater control over your contribution if you do this, but can you achieve the most good in this way? Should you operate at a meta-level, such as by trying to convince other people to change their charitable giving, attempting to influence government policy, or by raising awareness of existential risks, or should you try and directly tackle the problems facing the world -- such as by donating money yourself, or by tackling open problems in friendly AI?
Once you've figured out what to do, you still have to find a way to support yourself, and fund any organizations or projects you wish to support. You could work for an existing organization active in the area that you are interested in - bearing in mind that ones contribution will only be the benefit of hiring you rather than the next-best guy. Or you could work in a completely unrelated job, and work part-time on the cause you are interested in; this is a route followed by many open source developers, e.g. the prolific Fabrice Bellard. Alternatively, you could aim to earn as much money as possible, and use this money to fund causes or projects you are interested in; this is the route followed by Jeff Hawkins, who founded Palm, Inc. in order to fund AI and neuroscience research, as well as notable philantropists such as Bill Gates and Warren Buffet.
The problem is a simple one: how should one lead ones life in order to maximize the positive impact it has on others? There is an ample amount of data to draw from, such as charity rankings by GiveWell, salary data and personal experience. If rationality has any real-world benefits, then a discussion amongst rationalists should make it possible for substantially better decisions to be made than would otherwise be the case.
References
Existential Risk Reduction Careers Network
I arrived at the belief primarily instinctively, and am not particularly confident in it; I'd be happy to revise it on the basis of any more data I receive.
My rational is, roughly, that most adequately funded philanthropic organisations have no difficulties attracting talent, and sot the number of "doers" is determined primarily by demand-side factors. Therefore, by becoming a doer, I would be preventing another would-be doer from attaining a job. Whilst it's possible that this person would go on and contribute in other ways, such as by being a donor, I think it's very likely that they would not (most people seem to be strongly attracted to personally making a difference.)
Therefore, were my becoming a "doer" to have a positive impact, I would have to do such a better job than the person who has been displaced by me, that it would outweigh the loss of donations that I would otherwise have made. Whilst it would probably would be true that I would be more capable than the person who has lost out on getting a job because of me (supply growing should result in the least-capable losing out most), I don't believe that I have any major advantage over other people.
Whilst there seems to be no shortage of talented people looking for non-profit jobs, there is always a shortage of money, and my donating would be unlikely decrease anyone else's donations. So, I feel this is the more effective option.
I think I'm partly also influenced by a heuristic I sometimes use, that of avoiding what seems to be the easy option out. Most people seem to want to be intimately involved with the causes they help, yet there seems to be little justification for this, so I feel compelled to do the opposite. However, it occurs to me that I might in fact be attracted to the status and other benefits of high-paying jobs, so this may be just a rationalization.
I'm much less confident of this conclusion now than when I began writing this comment, which I think is probably a good thing. I'd be interested to hear arguments from people who've come to opposing conclusions to me.
That's probably a safe assumption to make in most charitable fields, but not in existential risk reduction. For example, there are not enough qualified candidates for the position of FAI programmer, or to lead the project that Nesov recently suggested.
Compare the following hypothetical scenarios: