I suspect there is a confusion of terminology here, and also perhaps some loss of institutional knowledge. Givewell did post-hoc analyses starting in 2011 of their 2009 and 2010 recommendations to donate to VillageReach, but these were not technically "grants", but rather "charity recommendations", so I guess wouldn't be considered a "grant lookback".

In recent years GiveWell shifted from a charity recommendation model to a more direct grantmaking model, so this could be the first reviews of grants under that new model.

I would guess that it's based on the marginal grant, but of course someone at GiveWell should be able to confirm.

Not a coincidence! I remembered reading this but couldn't find the link.

Indeed, I think a contrast with existing services like Paypal Giving Fund would be helpful.

Not addressed by this article is that growing the human population isn't that expensive, if you target your intervention. Like, we could probably incentivize births in (for example) Uzbekistan for a tiny fraction of the cost of doing so in (for example) South Korea.

I think those in favor of higher population should say what they think the ideal human population should be. Most of the hand wringing about population seems to be about whether it should be higher or lower than it is now, but that question seems to be anchored by the current population? If indeed "More Good is Better", does that imply that an Earth with a trillion people would be better than our Earth today? Because personally, I don't really agree with that.

Have you read Holden's classic on this topic? It sounds like you are describing what he calls "Influence matching".

It seems like even if workers are indifferent between the sweatshop jobs and the informal jobs, it could still be the case that the introduction of sweatshop jobs into the labor marketplace results in improvements for workers' conditions. For example, it could drive up the wages of informal labor.

One would have to believe the workers are overestimating the quality of their working conditions

It seems to me very possible that workers might not have complete information about working conditions. It’s not like they necessarily get to go on a factory tour beforehand. In fact, from my understanding, broken promises are the norm rather than the exception with this sort of thing.

I also think your argument implicitly assumes that the work is freely taken up, and freely withdrawn, which isn’t necessarily the case either. More egregious things like the forced labor situation in the shrimp and cocoa industries exist, of course, but we have decent evidence there is forced labor in Cambodia also.

To be clear, I agree with the main argument of the piece, I just think we need to acknowledge that it rests on these other assumptions which are not necessarily guaranteed (and should anyway at least be made explicit).

One thing that it seems to me often gets lost in this conversation is the assumption that the labor market is indeed a free market. Sweatshops seem fine to me if the labor is freely given; but when workers are locked in, chained to their machines, beat down when they try to organize, etc., it’s not clear to me that these free market conditions actually apply. The article mentions Cambodia, but Cambodia does not have a great reputation when it comes to corruption and governance. So, I think it’s worth asking, is the labor market free? If so then the rest of it seems sound to me.