> I’m writing a new guide to careers to help AGI go well, in collaboration with 80,000 Hours
> Yet only a few thousand people are working full time on navigating the risks
In the final post, I implore you to say something to the average reader about their very low chances of landing one of these jobs. EA jobs are exceptionally competitive, and AI jobs much much moreso.
I say this from experience; articles like these gave me the false impression that such jobs weren't getting enough qualified applicants, when in fact most had more than they could handle.
Given the current constraint, it follows that 80K should focus on increasing the job pool instead of making a small job pool even more competitive than it already is. But, even if you disagree with me on that, I earnestly request candor about the competitive nature of these jobs.
I also want to do more to emphasise e.g. the vast numbers of government positions that are relevant but not at 'AI risk' organisations, and how to contribute from outside of your job.
Benjamin, this is a great primer. I think it applies quite well for those curious about AGI. I will save it to share it with those in my network who are starting to become more interested about the space.
I have started my journey towards pivoting from my current career in finance to a new career in AI Policy & Governance. I have begun to take action following the advice given to me by an 80,000 hours career advisor. I have however struggled to see how I can fit into the space given my education & experience.
I almost feel like I need to get credentialed in order to maximize my chances of entering the space. I have an undergraduate degree in business, a globally recognized professional designation, and 10 years of work experience in finance. However I feel like I might need to get a graduate degree in Policy just to get in. I am conflicted because this would not only cost time and money, but also how relevant is a graduate degree in Policy in such a nascent space? I feel like I would only be doing it for the piece of paper.
Do you have any thoughts on my situation? I will continue to advance in journey as much as I can, and even if I have to stay in my career for a little longer, I will continue to learn about the space and help in any way I can during my free time.
I'd reading a bunch about AI policy (e.g. the material in the Bluedot course) and meeting people if the field if you can.
And then while you're doing that, try to applying to a bunch of organisations (or speak to people at those organisations if they could use someone with your skills).
Best case, you get a job right away.
Otherwise, ask them what you could do over the next year that would put you in the best position to get a job with them in the future, then try to do that.
If you already have 10 years' experience in finance, a masters is probably not necessary. You could enter government positions directly, and there are probably quicker ways to get into think tanks & non-profits too.
Finally keep in mind you could have a significant impact via earning to give. (And probably other paths too.)
Thank you for the detailed response. I have applied to the AI Governance course at Bluedot, waiting to hear back for next steps.
"Otherwise, ask them what you could do over the next year that would put you in the best position to get a job with them in the future, then try to do that." - This is fantastic advice, I will start inserting a variation of this question in my conversations.
Grateful for these excellent writings on AI! As a senior in uni, I’m eagerly awaiting the new guide
Thanks Benjamin! Your writings on AI are concise and actionable which I love!
May the techno-heroism be with you! Follow my example - postpone your job, align AIs and save worlds instead!
I have a lot of time as I don't work. Have all the credentials for this work. (elite tier credentials too, mathematics, AI, etc, big names (yuge))
If you shoot me an idea of something i can do i may do it. Not like looking for a career. more something I can do as an unemployed person by myself.
You could take a look at these lists of research project ideas and see if any grab your interest (then explore the further reading):
https://www.openphilanthropy.org/request-for-proposals-technical-ai-safety-research/
https://www.openphilanthropy.org/request-for-proposals-improving-capability-evaluations/
https://docs.google.com/document/d/1gi32-HZozxVimNg5Mhvk4CvW4zq8J12rGmK_j2zxNEg/edit?tab=t.0#heading=h.q14pjbvzx1x
You could also try reading some of the recent research and seeing if you can improve on any of it:
https://www.lesswrong.com/posts/fAW6RXLKTLHC3WXkS/shallow-review-of-technical-ai-safety-2024
There's also stuff to do in forecasting and data analysis (e.g. more stuff like Epoch).
i'll take a look
the world basically has the amount of people that fit in one google doc working on this doesn't it haha
> I’m writing a new guide to careers to help AGI go well, in collaboration with 80,000 Hours
> Yet only a few thousand people are working full time on navigating the risks
In the final post, I implore you to say something to the average reader about their very low chances of landing one of these jobs. EA jobs are exceptionally competitive, and AI jobs much much moreso.
I say this from experience; articles like these gave me the false impression that such jobs weren't getting enough qualified applicants, when in fact most had more than they could handle.
Given the current constraint, it follows that 80K should focus on increasing the job pool instead of making a small job pool even more competitive than it already is. But, even if you disagree with me on that, I earnestly request candor about the competitive nature of these jobs.
Yes I'll try to do better on this.
I also want to do more to emphasise e.g. the vast numbers of government positions that are relevant but not at 'AI risk' organisations, and how to contribute from outside of your job.
Benjamin, this is a great primer. I think it applies quite well for those curious about AGI. I will save it to share it with those in my network who are starting to become more interested about the space.
I have started my journey towards pivoting from my current career in finance to a new career in AI Policy & Governance. I have begun to take action following the advice given to me by an 80,000 hours career advisor. I have however struggled to see how I can fit into the space given my education & experience.
I almost feel like I need to get credentialed in order to maximize my chances of entering the space. I have an undergraduate degree in business, a globally recognized professional designation, and 10 years of work experience in finance. However I feel like I might need to get a graduate degree in Policy just to get in. I am conflicted because this would not only cost time and money, but also how relevant is a graduate degree in Policy in such a nascent space? I feel like I would only be doing it for the piece of paper.
Do you have any thoughts on my situation? I will continue to advance in journey as much as I can, and even if I have to stay in my career for a little longer, I will continue to learn about the space and help in any way I can during my free time.
Hi Jorge,
I'd reading a bunch about AI policy (e.g. the material in the Bluedot course) and meeting people if the field if you can.
And then while you're doing that, try to applying to a bunch of organisations (or speak to people at those organisations if they could use someone with your skills).
Best case, you get a job right away.
Otherwise, ask them what you could do over the next year that would put you in the best position to get a job with them in the future, then try to do that.
If you already have 10 years' experience in finance, a masters is probably not necessary. You could enter government positions directly, and there are probably quicker ways to get into think tanks & non-profits too.
Finally keep in mind you could have a significant impact via earning to give. (And probably other paths too.)
No worries! Also note you can see the blue dot curriculum and start reading, even without being on the course.
Thank you for the detailed response. I have applied to the AI Governance course at Bluedot, waiting to hear back for next steps.
"Otherwise, ask them what you could do over the next year that would put you in the best position to get a job with them in the future, then try to do that." - This is fantastic advice, I will start inserting a variation of this question in my conversations.
Cheers and I look forward to your future posts.
By PhD level questions, I mean discrete questions that PhDs in the relevant discipline can answer, but other people can't, even with access to Google.
As you're correct to point out, that's a long way from being able to do a whole PhD.
I'm clearer about this in the full article - this is just a summary. I might edit to clarify.