[FBT]

It is true: right at this very moment, billions of people have enough food do eat. It is a worthy task to try to extend that to the hundreds of millions who are not there yet, but it's hardly something that has never been done before, unlike driverless cars.

That doesn't mean it will be any easier (or harder), but it does mean that it is less of a "moonshot".

[estsauver]

Yah, it might also be different connotations on moonshot. If you believe a moonshot has to be something new + unlikely/hard then it's definitely not a moonshot. If you think a moonshot is something that's just incredibly unlikely but potentially super valuable, it definitely is.

Empirically, at least two investors who had put money into successful early-stage driverless cars and one who put money early into Planet Labs thought what we were doing was going to be harder. For some VCs, Africa might as well be the moon.

[Firadeoclus]

To me one implication of a moonshot is that you're aiming to hit a small target, and a miss is a miss no matter how close you get.

You made "end[ing] human hunger for a few hundred million people" your goal, but it isn't the case that anything short of that equals complete failure. Even if you only got a fraction of the way that'd be a massive success in my view.

[estsauver]

Totally, and it that sense not sure if we would count. There probably isn't much of a sustainable space in below a few hundred k though.

[whb07]

Here’s a quick question, what makes you think it’s the lack of technology that is the cause for poor farming yields? What makes you think it’s not due to property rights and rule of law?

If you went 50 years in to the future and came back with X widget for farming and just dropped it in random place in South Sudan or Niger, I don’t think they would just become farming power houses.

[estsauver]

When we talk about basic ag tech, we're not talking about drones or robots, we're not even talking about mechanization, it's mostly fertilizer and seed breeding.

- The accumulated knowledge of crop science is pretty strong, it's not a secret what the limiting factors in cereals crops are.

- Empirical results (see groups like The One Acre Fund, extensive research farm results in the region, high confidence RCTs leave basically no doubt)

- For US farmers prior to ~1930, they had the same average yields as Kenya does currently. In fact, fertilizer usage and yield have a very strong correlation and Kenya has a very similar curve.

Are there factors besides inputs that can impact yield? Absolutely. You could see from satellite imagery the places where conflict reached in Syria by harvest time as that conflict was developing. Market access and infrastructure in Kenya are super hard also, fertilizer is more expensive in rural Kenya than it is in the United States even though the consumers are much poorer.

Basically, there are certainly places that we're not a good fit for. I don't think we could operate in places with active conflicts like South Sudan, but Africa's a real big place. Kenya is quite stable and has effective if not formalized property rights for most smallholders. The thing I always try to remember is that America had very similar problems. The area that is now the great bread basket of the world also used to be referred to as "The Wild West."

[whb07]

That doesn’t address the core issue of rule of law and property rights though. Pick any number of countries in Africa and air drop them ag-tech that’s 50 years into the future. I don’t for a second believe that it would cause much of anything to change or improve.

Your US example appears to have been constrained by the current tech and knowledge of the day. In Kenya like you bring up, they aren’t limited by those things but rather lack of rule of law and property rights. So how can you compare them apples to apples?

This isn’t a chicken or egg thing. Rule of law and property rights are the necessary condition to allow people and society to grow and flourish. You can’t skip those necessary steps.

I mean look at the agricultural output for Zimbabwe for reference.