I am at the conference now so voice recorded this and used an LLM to clean up prose. N.B. wrote this in <30 min so it is very rough--but I thought speed was more relevant than polish given timeliness of the conference happening now
Lewis Bollard opened EAG London with the first social reform movement: abolitionism. He touched on others too, including Henry Spira bringing McDonald's to the table instead of demanding they stop serving meat. His argument: the through-line from them to EA is three virtues (altruism, radical empathy, and ambitious pragmatism). On the latter, the original abolitionists choosing the slave trade over slavery because it was winnable. Compromise as the price of victory.
I've spent my career in movements and systems change: getting community health workers salaried and accredited as national policy, the unglamorous machinery of #HealthforAll. This looks like approaching a problem by asking what conditions are holding the problem in place and changing them. i.e. Playing general manager for a whole problem, the way Lewis does for ending factory farming. I really, really dug his talk.
It also confused me. Because what he celebrated is not what I understood EA to be. From the outside (this is my first EAG), EA comes off as cost-per-outcome driven, and (particularly in GHD) a this manifests as huge emphasis on direct delivery: i.e. the bednet, not the movement that makes governments guarantee the right to health. Lewis spent his talk on scale, system change, and narrative reform, whereas my impression was that the question EA is obsessed with is the measurable marginal dollar, not the big system bet.
Relatedly, what struck me most in my first 12 hours @ EAG is EA clearly does believe in movement-building in the sense thatit spends an enormous amount on its own! EAG, 80,000 Hours, the fellowships, the community-building grants, this whole beautifully run machine (thank you, swapcard). So the conviction is there but it just seems to (largely) run inward, toward growing EA, rather than outward, toward funding the kind of movements Lewis spent 25 minutes admiring.
I may be wrong about my sense of mismatch and I'd like to be. Who here works on scale, systems change, narrative reform (esp in health delivery)? Am curious about those who have made the “GM for an entire problem” approach legible within EA rather than optimizing one slice of it.
Hi Madeleine, PauseAI might be a useful example for your list, @Toby Tremlett🔹 already flagged it above.
On whether it's an "EA organisation" (to @david_reinstein's question on Toby's thread): it isn't, but it shares most of the values and collaborates closely with both the EA and broader AI safety ecosystems, and that collaboration is central to our theory of change. We're squarely a movement-building / constituency-building org of the kind you're describing: our bet is on building an organised public constituency that creates the political conditions for governance of frontier AI.
On your two axes, we sit at the far upstream, low-measurability corner. I'll be candid that converting movement activity into attributable political outcomes is hard to measure, but it's something we're actively trying to get better at. We've got a longer write-up of our ToC coming soon and I'm happy to share it when it's up.