I think many of the differences between how traditional GiveWell makes funding decisions and how YCNP makes them is that GiveWell is aiming to fund larger, more established organisations, while YC is aiming at startups. If GiveWell switched to funding startups, then they’d use similar criteria to YC, except (i) put more emphasis on funding evidence-backed interventions (ii) not target organisations that use technology (which is just a pragmatic decision by YC given their expertise in and enthusiasm for that area).
Outside of international development (i.e. in Open Phil), then difference (i) would likely be reduced too, since in most causes it’s much harder to fund evidence-backed interventions that in international development. Though, Open Phil is still putting significant emphasis on cause selection, which YCNP puts much less emphasis on. (Though YC recently started doing something a bit like cause selection: https://www.ycombinator.com/rfs/).
In short, if you take Open Phil + focus on startups—cause selection + tech focus, then you get pretty similar criteria to YCNP.
This is non-obvious to me; GiveWell seems to recommend startups as well as more established organizations in my opinion. Even the more established ones are quite new in relative terms. One of GiveWell’s newer issues seems to be that these startups aren’t experienced at scaling and thus (in GiveWell’s opinion) don’t have enough room for more funding.
AMF is a two-man show operating in a limited area, there’s three full-timers on Deworm the World and they operate in just one geography, DMI is quite new and in one geography, etc.
With regards to YC, I’d say ‘(i) put more emphasis on funding evidence-backed interventions’ is a very significant difference. I’d also say another major one is considerations of cost-effectiveness, which don’t seem to factor in for YC nearly to the extent of GW.
It’s a good point the existing GiveWell nonprofits are very small in terms of staff. They don’t seem like startups at this point though because they’ve already had millions of dollars of funding (well over $10m for the top three) - I think revenue is more relevant than staff number—and they already have a clearly established model. Being focused on an already proven intervention is also a significant difference to most startups.
What I meant is that if the Open Phil team started funding brand new projects doing new interventions, then I guess they’d use pretty similar criteria; except Open Phil would put more emphasis on cause selection; and YC would put more emphasis on tech.
‘(i) put more emphasis on funding evidence-backed interventions’ is a very significant difference. I’d also say another major one is considerations of cost-effectiveness, which don’t seem to factor in for YC nearly to the extent of GW.
I’m lumping these together.
It is a significant difference, I agree. Though like I say, it’s most relevant in international development, which Open Phil is expanding out of.
Also I should be clear that two groups could have similar criteria but apply them in very different ways e.g. GiveWell seem more sceptical, so would need more evidence that a program works.
I think many of the differences between how traditional GiveWell makes funding decisions and how YCNP makes them is that GiveWell is aiming to fund larger, more established organisations, while YC is aiming at startups. If GiveWell switched to funding startups, then they’d use similar criteria to YC, except (i) put more emphasis on funding evidence-backed interventions (ii) not target organisations that use technology (which is just a pragmatic decision by YC given their expertise in and enthusiasm for that area).
Outside of international development (i.e. in Open Phil), then difference (i) would likely be reduced too, since in most causes it’s much harder to fund evidence-backed interventions that in international development. Though, Open Phil is still putting significant emphasis on cause selection, which YCNP puts much less emphasis on. (Though YC recently started doing something a bit like cause selection: https://www.ycombinator.com/rfs/).
In short, if you take Open Phil + focus on startups—cause selection + tech focus, then you get pretty similar criteria to YCNP.
This is non-obvious to me; GiveWell seems to recommend startups as well as more established organizations in my opinion. Even the more established ones are quite new in relative terms. One of GiveWell’s newer issues seems to be that these startups aren’t experienced at scaling and thus (in GiveWell’s opinion) don’t have enough room for more funding.
AMF is a two-man show operating in a limited area, there’s three full-timers on Deworm the World and they operate in just one geography, DMI is quite new and in one geography, etc.
With regards to YC, I’d say ‘(i) put more emphasis on funding evidence-backed interventions’ is a very significant difference. I’d also say another major one is considerations of cost-effectiveness, which don’t seem to factor in for YC nearly to the extent of GW.
It’s a good point the existing GiveWell nonprofits are very small in terms of staff. They don’t seem like startups at this point though because they’ve already had millions of dollars of funding (well over $10m for the top three) - I think revenue is more relevant than staff number—and they already have a clearly established model. Being focused on an already proven intervention is also a significant difference to most startups.
What I meant is that if the Open Phil team started funding brand new projects doing new interventions, then I guess they’d use pretty similar criteria; except Open Phil would put more emphasis on cause selection; and YC would put more emphasis on tech.
I’m lumping these together.
It is a significant difference, I agree. Though like I say, it’s most relevant in international development, which Open Phil is expanding out of.
Also I should be clear that two groups could have similar criteria but apply them in very different ways e.g. GiveWell seem more sceptical, so would need more evidence that a program works.