Thank you to you and the 80,000 Hours team for the excellent content. One issue that I’ve noticed is that a relatively large number of pages state that they are out of date (includingseveralimportantones). This makes me wonder why it is that 80,000 Hours does not have substantially more employees. I’m aware that there are issues with hiring too quickly, but GiveWell was able to expand from 18 full-time staff (8 in research roles) in April 2017 to 37 staff today (13 in research roles and 5 in content roles). Is the reason that 80,000 Hours cannot grow as rapidly that its research is more subjective in nature, making good judgment more important, and that judgment is quite difficult to assess?
Hi there, I think how quickly to hire is a really complex question. It would be best to read the notes on how quickly we think we should expand each of our programmes in our annual review as well as some of the comments in the summary.
Just quickly on the comparison with GiveWell, I think we’re on a fairly similar trajectory to them, except that GiveWell started 4-5 years earlier, so it might be more accurate to compare us to GiveWell in 2015. We are planning to reach ~25 staff, though it will take several more years. Another difference is that we allocate across a wider range of programmes (headhunting, advising, job board etc.), so even if we were the same size as GiveWell, we wouldn’t be doing as much research and content.
The out-of-date content is a problem that bugs me, though. One improvement we’ve made recently is that all the bottom lines are now kept up-to-date on the key ideas page.
I second this. I imagine that updating the AI problem profile must be a top priority for 80K because AI safety is a popular topic in the EA community, and it’s important to have a central source for the community’s current understanding of the problem.
Hi Ben,
Thank you to you and the 80,000 Hours team for the excellent content. One issue that I’ve noticed is that a relatively large number of pages state that they are out of date (including several important ones). This makes me wonder why it is that 80,000 Hours does not have substantially more employees. I’m aware that there are issues with hiring too quickly, but GiveWell was able to expand from 18 full-time staff (8 in research roles) in April 2017 to 37 staff today (13 in research roles and 5 in content roles). Is the reason that 80,000 Hours cannot grow as rapidly that its research is more subjective in nature, making good judgment more important, and that judgment is quite difficult to assess?
Hi there, I think how quickly to hire is a really complex question. It would be best to read the notes on how quickly we think we should expand each of our programmes in our annual review as well as some of the comments in the summary.
Just quickly on the comparison with GiveWell, I think we’re on a fairly similar trajectory to them, except that GiveWell started 4-5 years earlier, so it might be more accurate to compare us to GiveWell in 2015. We are planning to reach ~25 staff, though it will take several more years. Another difference is that we allocate across a wider range of programmes (headhunting, advising, job board etc.), so even if we were the same size as GiveWell, we wouldn’t be doing as much research and content.
The out-of-date content is a problem that bugs me, though. One improvement we’ve made recently is that all the bottom lines are now kept up-to-date on the key ideas page.
For those who are curious,
in April 2015, GiveWell had 18 full-time staff, while
80,000 Hours currently has a CEO, a president, 11 core team members, and two freelancers and works with four CEA staff.
We have 12.7 FTE of full-time staff, and 1.4 FTE of freelancers.
FTE = full-time-equivalent.
I second this. I imagine that updating the AI problem profile must be a top priority for 80K because AI safety is a popular topic in the EA community, and it’s important to have a central source for the community’s current understanding of the problem.
It it a top priority, though we only have one full-time writer at the minute, so it may still take a while.