Just a short update here on the outcomes from this process! The updated Handbook content is now live, with the end product mostly consisting of small changes to both structure and content. Below are some brief notes on the process that we followed, and the changes this culminated in.
We took the following steps to assess what edits to make:
Interviews with previous EA Handbook authors & CEA staff
Open-input survey for the broader EA community, via the survey linked above
Review of reading time data for each section of the Handbook
Review of qualitative data: Feedback that participants gave on the Handbook
Interviews with highly rated facilitators
Expert consultation on what are better up-to-date readings for a few topics
This led to a smaller number of edits overall than we were expecting, which we think was driven by the following factors:
People were overall pretty happy with the content
Some “outdated” things were hard to replace without having something that needs continuous updating (e.g. recent content on the Executive Order on AI, the EU AI Act etc. which will probably become outdated as the regulatory landscape evolves)
Sometimes we got conflicting feedback – so while from the outside you may look at it and think “people are unhappy, we need to make major changes,” which changes were appropriate in these cases were often non-obvious. That said, we’re not sure we made the right decision in all of these cases, and we think that a different editor might have made a higher number of edits. We’re considering coming back to this sometime later this year in light of this.
Some notes on the edits that we did make:
We redistributed some topics to different weeks, mostly to better standardise average reading times, since this was something that we got feedback on.
We moved some of the content around to improve the coherence of the learning objectives for the different sections, for example, by moving the content on the ITN framework and marginal impact to Section 2.
We made some updates to the readings in the Handbook to provide a more up-to-date overview of the topics covered, such as adding Our World in Data’s page on Global health to Section 2, and Managing AI Risks in an Era of Rapid Progress to Section 6. We also swapped some of the critiques in Section 7 to better reflect the current ecosystem. We traded out some older critiques on free-spending with some readings on the risks of naiveconsequentialism to highlight reflections since the FTX collapse.
Just a short update here on the outcomes from this process! The updated Handbook content is now live, with the end product mostly consisting of small changes to both structure and content. Below are some brief notes on the process that we followed, and the changes this culminated in.
We took the following steps to assess what edits to make:
Interviews with previous EA Handbook authors & CEA staff
Open-input survey for the broader EA community, via the survey linked above
Review of reading time data for each section of the Handbook
Review of qualitative data: Feedback that participants gave on the Handbook
Interviews with highly rated facilitators
Expert consultation on what are better up-to-date readings for a few topics
This led to a smaller number of edits overall than we were expecting, which we think was driven by the following factors:
People were overall pretty happy with the content
Some “outdated” things were hard to replace without having something that needs continuous updating (e.g. recent content on the Executive Order on AI, the EU AI Act etc. which will probably become outdated as the regulatory landscape evolves)
Sometimes we got conflicting feedback – so while from the outside you may look at it and think “people are unhappy, we need to make major changes,” which changes were appropriate in these cases were often non-obvious. That said, we’re not sure we made the right decision in all of these cases, and we think that a different editor might have made a higher number of edits. We’re considering coming back to this sometime later this year in light of this.
Some notes on the edits that we did make:
We redistributed some topics to different weeks, mostly to better standardise average reading times, since this was something that we got feedback on.
We moved some of the content around to improve the coherence of the learning objectives for the different sections, for example, by moving the content on the ITN framework and marginal impact to Section 2.
We made some updates to the readings in the Handbook to provide a more up-to-date overview of the topics covered, such as adding Our World in Data’s page on Global health to Section 2, and Managing AI Risks in an Era of Rapid Progress to Section 6. We also swapped some of the critiques in Section 7 to better reflect the current ecosystem. We traded out some older critiques on free-spending with some readings on the risks of naive consequentialism to highlight reflections since the FTX collapse.