In this post, I summarize a tough career decision I have had to take over the last few weeks.
Setting the stage
I am a few months from finishing my Ph.D. in quantum algorithms. During these 4 years, I have become quite involved in Effective Altruism: I attended two EA Globals, facilitated a couple of virtual intro fellowships, and helped organize the EA group in Madrid. Given my background, I have also felt closer to AI Safety than any other cause area. As such, I have also been involved in the AI Safety community, by participating in two AI Safety Camps, and as a facilitator in some intro fellowships. I even did a summer internship in AI Safety with José Hernández Orallo last summer which led to a rather lucky AAAI publication.
My Ph.D. has also gone well. While it started a bit dubitative and was unable to get anything published for the first two years, at that point I got my first two publications and over the last two, I have been able to do well. It is not a superstar Ph.D. but I believe I have learned enough to make contributions to the field that actually get used, which is harder than it looks. In fact, I feel happy that thanks to my last article, one rather serious quantum startup contacted me to collaborate, and this led to another quite high-quality paper.
Since I am finishing my Ph.D., I had to plan my next step. The first obvious choice was to apply for funding to do AI Safety. I cold emailed Victor Veitch, who I found through the super useful Future of Life AI Safety Community, and he was happy to take me as long as I could work more or less independently.
The reason why I opted for applying with Victor was that my research style is more about knowing well the tools I am using, not to the level of a pure mathematician, but to the level where techniques are known and can be used. Additionally, I think causality is cool, and being able to apply it in large language models is rather remarkable. I am also a big fan of Ryan Carey, who works in causality and is one of the people in the community who has helped me the most. I am really grateful to him. Apart from the Future of Life postdoc program, I also applied to the EA Long Term Future fund, and Open Philanthropy, with this proposal. Out of this, the EA Long Term Future fund accepted funding me, most likely on the basis of a career change, while FLI declined based on the proposal and probably an interview that while I prepared was not able to use to explain well why I think this perspective could be useful. This came a bit of a disappointing result, to be honest. Open Philanthropy, on the other hand, is behind schedule, and I don't know their answer yet.
The alternative was to pursue, at least for now, a more standard research topic. I applied to IBM, Google, Amazon, and three startups: Zapata, PsiQuantum, and Xanadu. With the latter, I have been working, so it was not really an application. I never heard back from any of the big companies, but got offers from the three startups. To be fair I also applied to a couple of non-EA AI postdocs with the hope of getting them, but they were a very long shot. For a bit of context though, PsiQuantum is a very serious contender in building a photonic quantum computer and error correction and has really deep pockets, while Xanadu is probably a bit behind them, but it's also quite good and has a bit more focus on ML.
The situation and conditioning factors.
Perhaps the main conditioning factor on all of this was the fact that my girlfriend is really a very important part of my life, and responsible for my happiness. A friend of mine called this the unsolvable two-body problem 😛. In particular, I think that we both would like to have a smooth life. She is currently not able to move with me because she wants to be a teacher and that requires that she get some substitutions beforehand in Spain. Finally, there is the issue of the cost of changing research areas, and of an academic life, which is rather stressful and not very well paying.
What I did to take this decision.
To make this decision I have talked to pretty much everyone I could think of. I posted a couple of posts in the EA forum. Some people argued that it may be helpful to have someone in the community with expertise in Quantum Computing, although I am not convinced for two reasons: a) the impact of QC in AI Safety seems very unlikely, and b) even if QC were to become important at some point, I think I could still pick it up fairly quickly. However, it is true that I would marginally be one of the very few people in this position.
I obviously also talked a lot about this with my girlfriend and my family. They do not really understand why we cannot leverage the academic ML community for this instead of asking for career changes, and I am not totally sure they are wrong on this. Additionally, I talked with Shay, the certified health coach who collaborates with AI Safety Support. I also tried to talk to Habiba, from 80k hours, with whom I had previously done coaching, but I think she's been quite busy lately. The biggest favorable point for the quantum computing startup offers are the capability to work from home, and the salary (which is much higher than anyone in my family has ever earned). However, it also allows me to learn some more general ML and software engineering skills, which may be helpful down the line, without having to stop publishing.
The decision I have (reluctantly) taken.
While I believe that it would be a great opportunity to work on AI Safety, I have often felt that this is not only my decision but one that affects my couple too. It is perhaps for this reason that I have been more hesitant to decide on what I like best independently. To be fair, it is also the case that I am much happier being with her than going abroad, so overall I have (a bit sadly) felt that it might be a good idea to work remotely for Xanadu. They value me, and I think that I could work well with them.
However, I have not really renounced working in AI Safety, I just believe this will make things a bit more flexible down the line. In particular, I have proposed myself to become a distiller of academic papers in the field of causality for AI Safety. It's the same niche as my Open Philanthropy proposal, and I expect to dedicate 1 day/week (weekends, as I should then be more or less free) to writing beautiful posts explaining complicated papers. The objective is again similar to the postdoc, albeit much less impactful: to learn well a topic. Let this then be a preannouncement of a sequence on Causality, Causal Representation Learning, and Causal incentives, all applied to a certain degree to AI Safety.
I also expect to retry doing something a bit more serious in the future in AI Safety, and my girlfriend has agreed that I should do it, as it is something I value. And she will try to move around with me then, or by then I might be able to work remotely or in Valencia, Spain with Jose Hernández Orallo. In any case, it is true that I am a bit sad about not taking this opportunity right now.
I am happy to hear casual comments that I am mistaken and I should take the postdoc: I expect to make the final decision by this Sunday, but consider it more or less taken in any case.