Someone on the forum said there were ballpark 70 AI safety roles in 2023
Just to note that the UK AI Security Institute employs more than 50 technical staff by itself and I forget how many non-technical staff, so this number may be due an update.
I think that is a reasonable path; SWE/ML will give you a good foundation anyways in early career if you want to switch to AI safety later as you build experience. Additionally, something in security is a good idea as well.
Why security?
I am a university student in computer science. My plan is to earn to give via SWE/ML because I don’t think there is enough opportunity in AI safety. Someone on the forum said there were ballpark 70 AI safety roles in 2023. Meanwhile, SPAR got like 600 applications last round. MATs has a 1% acceptance rate. AI safety is very competitive. However, I think most AIS researchers work under grant funding which surely cannot pay as well as an earning to give tech job. Surely I could do more good earning to give by funding multiple AIS researchers who would have otherwise chosen a different research field.
Rob Wiblin and a few others said that in the ideal world, only 15% of EAs earn to give. But I can’t understand the reasoning
I am looking for opinions, please comment if you have any.
I think this roughly makes sense?
A crux here would be whether you think there are enough people and funding going to the most important opportunities/directions of work.
OP has a lot of money, but they are not funding everything. If you think they cover almost everything important then that is there is a stronger case for direct work (e.g. try to get a grant from them)
You should also think about how confident you are that you can spend your own money effectively. Are you planning to choose projects to fund yourself? Will you defer to LTFF? How much do you trust the judgement of their evaluators compared to yours in knowing what's important?
You should likely explore high-absorbancy career paths (e.g., fieldbuilding, comms) as well, as the ones you refer to in the post seem to be low-absorbancy.
Why is SWE/ML low absorbency?
sorry for the late reply. SWE/ML is low absorbency in a sense that there are a limited number of AIS orgs that can employ people. They are high absorbency if you want to use these skills to do earning to give. Hope that clarifies what I meant!