Thanks so much Sofia for building on my post! So good to see a movement leader weighing in. Great observations. To build on yours in turn:
- Love the idea of “context-absorption infrastructure”. Interestingly this is something I recommend that organisations do for the opposite reason as well. I think that organisations should be trying to document as much of their context as possible because giving that context to AI allows AI to do better work. I like that this could also have the side benefit of creating a very good set of learning materials to allow new hires to rapidly come up to speed.
- I also think it’s an excellent point that the movement might collapse to a smaller number of very experienced people. Something that AI can’t really replicate is the tacit experience that comes from having done something every day. That kind of thing is also quite hard to teach. There are just so many things that you build up unconsciously day after day of doing something.
- When it comes to mentoring and management being more important. Mentoring: yes, management: maybe not! Truly great employees need a lot less managing by definition, and so we may just shift to prioritising autonomy and need for less management in our hiring. However coming back to your comment about professionals from outside the movement needing to rapidly get context, I do think the mentoring aspect does become very important.
- I think your point about “Engagement without employment may become core infrastructure” Might be the most important point. I don’t think it’s inherently bad that we shift towards this. However we must set expectations of the people doing these engagements. For example, hackathons in Tech are a great way of doing this: everyone gets together for a day, a weekend, form teams and attempt to solve a problem. After the hackathon there is no expectation that they will do a part II, or get a job working with any of the organisations they help. It’s just seen as a one-off project. On the other hand, I could easily imagine that many people start fellowships and cohorts with the main priority of using it as a stepping stone to get a job in the movement. We should be very clear if our engagements are not going to help people get jobs in the movement, otherwise we might end up with lots of people becoming disillusioned.
- On your tradeoff: “Explicitly weighing internal training against external recruitment”: I strongly external recruitment in all but one case. Internal training costs money we don’t have, so we should be very careful about doing it. I think the essential thing to bring it back to is the idea that the tacit knowledge built up over years is the valuable thing we want. External hiring is in effect paying money for the crystallisation of someone’s life experience. The only time when I think internal training makes sense is when you have activists with very deep context and experience and you train to upskill them in AI tooling. I don’t think the movement should be devoting time or resources to training people who don’t have useful experience and context either inside the movement or outside.
I think that your points 2 (Separating learning roles from output roles.) and 3 (Treating coordination and mentorship as core infrastructure) are important, and we’ve kind of beeing doing this in EA for a while as there are many orgs dedicated to fellowships and cohorts. However coming back to my point about setting expectations: many of these organisations do emphasise the job prospects of doing their cohorts. These organisations should probably be thinking about whether continuing to communicate this is a good idea or not.
Hey Richie, many thanks for reading and your comment! Appreciate it. I basically agree with you on most of the points.
Really like the additional detail on how engagement should be clearly for the engagement’s sake and not advertised as leading to employment, if it’s increasingly rarely the case, or in the future, when we use it as a tool to keep people in the movement without employment.
Re: management and mentorship: I think I mostly agree with you, but I may be using “management” in a slightly broader sense.
I agree that high-performing, motivated people usually need much less direction and oversight, and that hiring for autonomy matters a lot. In that sense, great people absolutely reduce the need for micromanagement.
Where I still think management remains important, even with very strong people and very advanced AI, is around coordination, prioritisation, decision-making under uncertainty, and looking after the human side of work. Once you have more than a handful of people, someone still needs to hold the whole picture, notice misalignment early, make trade-offs explicit, and ensure people aren’t burning out or duplicating effort.
I also think mentoring and management can blur a bit. Helping people understand context, giving feedback, supporting growth, and creating the conditions for good work are often framed as mentoring, but they’re also core parts of good management.
So I’m less convinced that we’ll “need less management” so much as we’ll need a different kind of management: lighter-touch, more relational, more focused on sense-making and coordination rather than task assignment. AI may reduce some parts of the job, but I’m not sure it replaces the human judgment and care elements that show up as soon as teams scale beyond a few people.
Very much enjoying thinking through this topic and grateful to you for starting the conversation!
Cross posting my comments on your blog:
Thanks so much Sofia for building on my post! So good to see a movement leader weighing in. Great observations. To build on yours in turn:
- Love the idea of “context-absorption infrastructure”. Interestingly this is something I recommend that organisations do for the opposite reason as well. I think that organisations should be trying to document as much of their context as possible because giving that context to AI allows AI to do better work. I like that this could also have the side benefit of creating a very good set of learning materials to allow new hires to rapidly come up to speed.
- I also think it’s an excellent point that the movement might collapse to a smaller number of very experienced people. Something that AI can’t really replicate is the tacit experience that comes from having done something every day. That kind of thing is also quite hard to teach. There are just so many things that you build up unconsciously day after day of doing something.
- When it comes to mentoring and management being more important. Mentoring: yes, management: maybe not! Truly great employees need a lot less managing by definition, and so we may just shift to prioritising autonomy and need for less management in our hiring. However coming back to your comment about professionals from outside the movement needing to rapidly get context, I do think the mentoring aspect does become very important.
- I think your point about “Engagement without employment may become core infrastructure” Might be the most important point. I don’t think it’s inherently bad that we shift towards this. However we must set expectations of the people doing these engagements. For example, hackathons in Tech are a great way of doing this: everyone gets together for a day, a weekend, form teams and attempt to solve a problem. After the hackathon there is no expectation that they will do a part II, or get a job working with any of the organisations they help. It’s just seen as a one-off project. On the other hand, I could easily imagine that many people start fellowships and cohorts with the main priority of using it as a stepping stone to get a job in the movement. We should be very clear if our engagements are not going to help people get jobs in the movement, otherwise we might end up with lots of people becoming disillusioned.
- On your tradeoff: “Explicitly weighing internal training against external recruitment”: I strongly external recruitment in all but one case. Internal training costs money we don’t have, so we should be very careful about doing it. I think the essential thing to bring it back to is the idea that the tacit knowledge built up over years is the valuable thing we want. External hiring is in effect paying money for the crystallisation of someone’s life experience. The only time when I think internal training makes sense is when you have activists with very deep context and experience and you train to upskill them in AI tooling. I don’t think the movement should be devoting time or resources to training people who don’t have useful experience and context either inside the movement or outside.
I think that your points 2 (Separating learning roles from output roles.) and 3 (Treating coordination and mentorship as core infrastructure) are important, and we’ve kind of beeing doing this in EA for a while as there are many orgs dedicated to fellowships and cohorts. However coming back to my point about setting expectations: many of these organisations do emphasise the job prospects of doing their cohorts. These organisations should probably be thinking about whether continuing to communicate this is a good idea or not.
Hey Richie, many thanks for reading and your comment! Appreciate it. I basically agree with you on most of the points.
Really like the additional detail on how engagement should be clearly for the engagement’s sake and not advertised as leading to employment, if it’s increasingly rarely the case, or in the future, when we use it as a tool to keep people in the movement without employment.
Re: management and mentorship: I think I mostly agree with you, but I may be using “management” in a slightly broader sense.
I agree that high-performing, motivated people usually need much less direction and oversight, and that hiring for autonomy matters a lot. In that sense, great people absolutely reduce the need for micromanagement.
Where I still think management remains important, even with very strong people and very advanced AI, is around coordination, prioritisation, decision-making under uncertainty, and looking after the human side of work. Once you have more than a handful of people, someone still needs to hold the whole picture, notice misalignment early, make trade-offs explicit, and ensure people aren’t burning out or duplicating effort.
I also think mentoring and management can blur a bit. Helping people understand context, giving feedback, supporting growth, and creating the conditions for good work are often framed as mentoring, but they’re also core parts of good management.
So I’m less convinced that we’ll “need less management” so much as we’ll need a different kind of management: lighter-touch, more relational, more focused on sense-making and coordination rather than task assignment. AI may reduce some parts of the job, but I’m not sure it replaces the human judgment and care elements that show up as soon as teams scale beyond a few people.
Very much enjoying thinking through this topic and grateful to you for starting the conversation!