Donation history

My donation history begins in 2015 because that is the year during which I first started working/earning money that could be considered “my own”.

2015

I earned around $2,000 doing mostly content creation work under Vipul Naik. I did not donate in 2015 because I am still building a financial buffer.

2016

I earned around $9,700 (not counting deferred payments) working under Vipul Naik. I did not donate in 2016 because I am still building a financial buffer.

2017

I earned around $34,000 [ETA: this ended up being more like $37,000 due to tasks from 2017 being added to Vipul’s Contract Work website in January–March 2018]. This amount includes the $9,360 deferred payment from last year. My main source of income was contract work for Vipul Naik. I also did some contract work for Peter Hurford (recorded on Vipul Naik’s contract work site) and sold leftover bitcoin that I had (worth around $250).

In addition, Vipul Naik offered me $500 to donate to a charity of my choice (kind of a bonus for working for him all year).

I decided not to donate in 2017 (both with my own money and Vipul’s offer, which I deferred to a later time) due to a combination of (1) uncertainty about my understanding of the world; and (2) continuing to build a financial buffer (which is a smaller priority now that I’ve crossed the $30k threshold).

2018

I earned around $18,000. Like last year, my main source of income was contract work for Vipul Naik. I also wrote a timeline of the Future of Humanity Institute, which was funded by Alex Flint.

In addition, like last year, Vipul Naik offered me an additional $500 to donate to a charity of my choice.

I decided to take up Vipul’s offer this year, using both last year’s and this year’s amount (so a total of $1,000). I chose the following allocation:

• $500 to Machine Intelligence Research Institute (with an additional $500 from REG’s Double Up Drive). This was mostly intended as a retrospective donation, as thanks for producing useful content and ideas, and for doing work that I consider useful in the AI safety space. As of late 2018, for prospective funding, I do share concerns about MIRI’s new nondisclosure-by-default policy and also think that its “room for more funding” may not be as great as in the past (the latter is sort of hard to assess due to the former).
• $500 to the $500k block donor lottery, hosted by Centre for Effective Altruism. After watching the donor lottery for the first two years, I felt comfortable about giving to it. The donor lottery seemed like a good choice given my general uncertainties about the world.

Some other things I considered but didn’t end up choosing:

• EA Funds (especially the Long-Term Future Fund). Since the fund managers changed only recently, I decided I want to continue watching their activity.
• Donating to certain “woke individuals”. There are maybe around four individuals at the intersection of (1) people I consider especially smart/“woke”/“have good world models”, and (2) people who don’t seem to have access to a bunch of resources with which to implement their “vision” (sorry if this language is too vague). I decided to skip this and continue thinking about it for the following reasons: (1) These people don’t generally advertise a need for funding; and (2) I don’t want to make my future interactions with them awkward by offering them money.
• Upcoming AI safety people. In principle I would like to fund some early-stage AI safety people, but in practice I found that (1) I don’t know any of them well; (2) some of them are pseudonymous; (3) they are generally not anywhere near as transparent as I would like them to be. These factors made assessment difficult, so I decided to skip this and keep watching the space.

I decided to not donate any of my own money as I am now reducing my contract work hours in order to do something about AI safety (so I might actually need to use some of my savings for something).

Besides the donations listed above, Vipul also gave another donation on the same day (that he chose himself); you can read about that donation in the post he wrote about his 2018 donations.

2019

I earned around $5,000. Like the previous years, my main source of income was contract work for Vipul Naik. I also did some contract work for BERI. (There were also some smaller sources of income, but I won’t bother to list them all.)

I decided not to donate in 2019. This is mainly for two reasons:

Personal finances

I worked very little this year, so my income is small. My future income is also uncertain at this point, so I might need to use my savings in the future.

Low value of small marginal donations to targets I am most interested in funding

I am currently most excited about funding independent AI safety researchers and small projects (in AI safety and some other areas). My best guess is that funding these individuals or projects to a point where there are significant counterfactual differences requires at least around $50,000 (amount calculated based on paying one researcher who doesn’t require a lot of money for one year). As far as I know, there isn’t a simple way to coordinate with other people interested in funding independent researchers, so I can’t just say something like “I pledge to give $1,000 to person X if enough other people also pledge to do the same” and have all the logistics taken care of. The best alternatives to do this sort of coordination to fund independent researchers at the moment seem to be EA Funds’ Long-Term Future Fund and the EA Hotel; I think these are decent options but I prefer having more control over what gets funded.

2020

I earned around $10,000.

• I started giving $5 per month to Andy Matuschak’s Patreon account ($35 total over the course of the year, as I started partway through the year). While I do think this work is valuable, I started giving mostly for selfish reasons of wanting to see more of Andy’s output, so I don’t consider this a donation calculated to maximize impact on the world.
• I gave $10 total to MIRI as bounty payment (i.e. someone completed a bounty I put out, and this person requested I donate to MIRI rather than paying them).

Otherwise I did not donate, for basically the same reasons as 2019 (though the details have changed slightly).

2021

I earned around $15,000 (or more precisely, I received around $15,000 during the year, which is my income for tax purposes, but not all of that amount has been “earned”).

• I continued giving $5 per month to Andy Matuschak’s Patreon account.

Otherwise I did not donate, for basically the same reasons as in 2019.