Can confirm, I did not appreciate Austin's reply! I would have preferred something less warm that actually responded to what I said. Of course I can't know what Austin intended, but the response made me more doubtful of his positive intent: was talking past me to love bomb AIM was a good faith response to me or a performance good intent to onlookers? (it being the blanket approach used for multiple critics didn't help my doubts).
A note for the more AI-focused readers, which I've left out of the post because it's a different conversation. If your views on AI are such that you think ~everyone should be working on it, broad timelines be damned, then you won't find this post persuasive. But we'd still like to help you! There are plenty of ways for AIM to help with short-timelines work, that you may not be aware of:
Lots of people are trying to start incubators right now, in causes we work on and in ones we don't, like AI safety and catastrophic biorisks. Within EA, we have the most incubation experience(?), so many of them come to us for advice. We plan to help them a lot more through:
Publishing learnings: Taking more time to distill and publicly share our institutional knowledge (see our shiny new substack).
Advising: Bespoke support for new incubators throughout their journey, with our exec team carving out time for this from now on. For example, we're starting to work closely with Catalyze, an AI safety incubator staffed by AIM alumni and working out of our London office.
Incubating incubators: The founder of a South American incubator is currently shadowing our program to see how it works from the inside. We're in talks with others about doing the same, and might pilot a version of the program built specifically for people starting incubators.
Partnering: We're open to partnering with others, as long as it doesn't pull us away from our core work of launching charities.
We've heard it said: "If AIM took AI seriously, it would pivot to incubating AI safety organizations." I don't think that's right. Incubating AI safety organizations isn't our comparative advantage: we have close to zero institutional knowledge about AI safety (how to research it, how new organizations in the space should navigate it, and so on), and pivoting would mean starting again on everything we've learned about incubating the kinds of charities we already launch. But we can and should help the people who do have that advantage. Those causes need different things from ours, and we'll have more impact helping other incubators do it well than doing it ourselves, especially because incubation can very easily go badly.
If you're starting or running an incubator, in AI safety or anywhere else, and think we could be useful, we'd love to hear from you here.
I enjoyed this post, thank you for writing it. You point to examples of fights that showed little/no progress for a long time before being won, as an argument against dismissing diet change / moral persuasion advocacy for animals.
But how do we make this argument falsifiable? What would you need to see to decide it’s not a fight worth having?
And what about base rates: if you consider not just the long fought fights that eventually won, but also those that failed, how sensible does it look to bet on a given fight?
[These aren’t supposed to be gotchas or rhetorical questions, they’re genuine questions I wonder about (and I feel the need to clarify that I, a human wrote this, despite using the word “genuine” in 2026)]
I agree it’s super unintuitive. I was very skeptical myself. We had the confidence to enact this policy because (a) our salaries had been uniformly low since we began in 2018, and motivation, talent density and retention were strong, and (b) our staff salary preference survey and intell from potential hires we lost due to salary suggested that this policy would result in modest increases in costs but substantive improvements in staff retention, hiring and wellbeing. The policy is live and (c) so far so good! Maybe things will change, but so far, seemingly the decades of research and experience about motivation in general aren’t the best model of motivation amongst AIM’s current and target staff.
Hi Jeff, thanks for sharing. I agree that senior talent can well be worth paying more for, and that where someone's financial needs are higher now because of what they gave up earlier, that's a legitimate need like any other. I can't speak for the conversation you had as I'm new to AIM's leadership team, but It would happen differently under our new policies :)
Yes! Annually, staff have the option to (1) accept an inflation adjustment, (2) decline it if their current salary is working for them, (3) propose an adjustment to their salary based on a change in their circumstances
Good question! I think this change will be net positive for founder talent.
Firstly I think that in practice there is less competition for the same talent than one might expect:
(1) The talent pools are actually pretty different. Many of the people who are drawn to our incubation program specifically want to be founders, not staff. Conversely, often when a founder comes to work at AIM it’s because they don’t want to be founding at that time.
(2) AIM can’t actually absorb many founders. We’re increasingly specialized, meaning there are fewer and fewer roles a generalist founder could fill.
A relevant stat: Currently 4 out of 20+ staff are former CEIP founders (and 3 out of 4 are in the Programs Team -- where founding experience matters most)
(3) Hopefully the new salary policy will attract more talent to the AIM ecosystem, such that (a) AIM would have less need to hire people who would otherwise found, and (b) more new founders are unlocked than AIM is capable of absorbing.
Can confirm, I did not appreciate Austin's reply! I would have preferred something less warm that actually responded to what I said. Of course I can't know what Austin intended, but the response made me more doubtful of his positive intent: was talking past me to love bomb AIM was a good faith response to me or a performance good intent to onlookers? (it being the blanket approach used for multiple critics didn't help my doubts).
A note for the more AI-focused readers, which I've left out of the post because it's a different conversation. If your views on AI are such that you think ~everyone should be working on it, broad timelines be damned, then you won't find this post persuasive. But we'd still like to help you! There are plenty of ways for AIM to help with short-timelines work, that you may not be aware of:
Lots of people are trying to start incubators right now, in causes we work on and in ones we don't, like AI safety and catastrophic biorisks. Within EA, we have the most incubation experience(?), so many of them come to us for advice. We plan to help them a lot more through:
We've heard it said: "If AIM took AI seriously, it would pivot to incubating AI safety organizations." I don't think that's right. Incubating AI safety organizations isn't our comparative advantage: we have close to zero institutional knowledge about AI safety (how to research it, how new organizations in the space should navigate it, and so on), and pivoting would mean starting again on everything we've learned about incubating the kinds of charities we already launch. But we can and should help the people who do have that advantage. Those causes need different things from ours, and we'll have more impact helping other incubators do it well than doing it ourselves, especially because incubation can very easily go badly.
If you're starting or running an incubator, in AI safety or anywhere else, and think we could be useful, we'd love to hear from you here.
I enjoyed this post, thank you for writing it. You point to examples of fights that showed little/no progress for a long time before being won, as an argument against dismissing diet change / moral persuasion advocacy for animals.
But how do we make this argument falsifiable? What would you need to see to decide it’s not a fight worth having?
And what about base rates: if you consider not just the long fought fights that eventually won, but also those that failed, how sensible does it look to bet on a given fight?
[These aren’t supposed to be gotchas or rhetorical questions, they’re genuine questions I wonder about (and I feel the need to clarify that I, a human wrote this, despite using the word “genuine” in 2026)]
Ambitious Impact will also not be participating in or collaborating with Manifund projects for the foreseeable future.
I agree it’s super unintuitive. I was very skeptical myself. We had the confidence to enact this policy because (a) our salaries had been uniformly low since we began in 2018, and motivation, talent density and retention were strong, and (b) our staff salary preference survey and intell from potential hires we lost due to salary suggested that this policy would result in modest increases in costs but substantive improvements in staff retention, hiring and wellbeing. The policy is live and (c) so far so good! Maybe things will change, but so far, seemingly the decades of research and experience about motivation in general aren’t the best model of motivation amongst AIM’s current and target staff.
DM'd you on Substack :)
Hi Jeff, thanks for sharing. I agree that senior talent can well be worth paying more for, and that where someone's financial needs are higher now because of what they gave up earlier, that's a legitimate need like any other. I can't speak for the conversation you had as I'm new to AIM's leadership team, but It would happen differently under our new policies :)
Yes! Annually, staff have the option to (1) accept an inflation adjustment, (2) decline it if their current salary is working for them, (3) propose an adjustment to their salary based on a change in their circumstances
Good question! I think this change will be net positive for founder talent.
Firstly I think that in practice there is less competition for the same talent than one might expect:
(1) The talent pools are actually pretty different. Many of the people who are drawn to our incubation program specifically want to be founders, not staff. Conversely, often when a founder comes to work at AIM it’s because they don’t want to be founding at that time.
(2) AIM can’t actually absorb many founders. We’re increasingly specialized, meaning there are fewer and fewer roles a generalist founder could fill.
A relevant stat: Currently 4 out of 20+ staff are former CEIP founders (and 3 out of 4 are in the Programs Team -- where founding experience matters most)
(3) Hopefully the new salary policy will attract more talent to the AIM ecosystem, such that (a) AIM would have less need to hire people who would otherwise found, and (b) more new founders are unlocked than AIM is capable of absorbing.
I really like this framework! Thank you!