Hacker News new | past | comments | ask | show | jobs | submit login
The FarmBot Genesis Brings Precision Agriculture to Your Own Backyard (smithsonianmag.com)
109 points by Osiris30 on July 24, 2016 | hide | past | favorite | 37 comments

"All the food you need" is a bit of a stretch for me. Maybe I'm mistaken. A human needs around 2000 kcals per day, or a little less than 100 joules/second. Easy to remember: we use as much energy as a 100 watt incandescent light bulb.

Solar energy has a maximum output of 1120 W/m2 at the surface of the earth. So theoretically, a square meter could support 10 people. But there are one or two efficiency losing steps that throw this all off.

There's only light for half the day. Plants only are about 3-6% efficient at absorbing solar energy. The plants also are only taking up a fraction of the ground they're using, let's be liberal and say half. Then there's the whole 'how much of the energy the plant used was actually converted to edible calories' question, which is beyond my ability.

Multiplying that together, the best I can imagine is around 12 m^2 per human.

I'd love someone to show me a mistake in that logic, because I am super excited about things like this. I know I'm going to own one of these someday in the future.

I do some work on crop modeling. A few years ago at a project meeting, in the evening in a bar, we did some rough calculations; there were agricultural engineers and agricultural economists in our group, so it wasn't completely amateur hour (but not scientifically robust either, of course). Anyway, the result was that in Western Europe you need about 100m2 per person to grow enough food to sustain oneself for a year. This took into account that you need to grow crops that keep a long time to last you through winter etc. It didn't particularly consider variety in crops (neither for taste not for soil management purposes), but it wasn't potatoes year round either. You need to be a competent farmer as well, but it didn't consider really high tech closed-greenhouse-like setups either. Still, that was a sobering number for many in the crowd - it completely undermines the whole 'self-sufficiency = sustainable' mantra.

100m^2 doesn't seem like so much to me at all, certainly not sobering.

This would mean one acre of land would sustain about 40 people (or 100 per hectare).

Taking wikipedia's number of 13,958,000 km^2 of arable land in the world, this would mean a maximum world population of 137,963,876,580, which is about 20x the current world population.

Of course, this assumes equal quality of land to the land in your rough calculations, and I am guessing your number is assuming far more intensive cultivation than is typical (typical numbers I've seen are 1-2 acres / person).

The full calculation would also need to take in to consideration wastage / spoilage; crop failure; loss to pests; nutrient density (probably can't thrive off potatoes and corn only); animal feed (pets as well as animals for food).

Having said that, I'll now go on to say: with one inhabited planet there are limits to resources, but the limits to human resourcefulness are undefined.

Don't eat animals. One problem solved. Might want to have some cows for milk and using their waste as fertilizer

Edit: the not eating animals is not in a vegatarian way. I love meat. But if you want to fully sustain yourself on the smallest area possible, animals do not aid your cause.

Hens could probably be useful, as they can turn pests into eggs.

Did you drop a zero?

On an all-potato diet, assuming intensive cultivation (40,000 lbs/acre yield) you could feed about 20 people on one acre. That's 200m^2 per person. For a varied diet that might contain chicken and eggs, 1000m^2 (a quarter acre) is more reasonable.

It's possible. I dumped a lot of variables into Wolfram Alpha in a big equation and it came out with an answer in "m^2 per person" units.

Also, it's not a practical minimum but a theoretical one, like supposing we had so many plants absorbing light that it's dark underneath them, etc.

In reality, potatoes can do about 17m cals per acre per year, and a human needs nearly 1m per year. In short, Matt Damon definitely wasn't growing enough potatoes (in the film version at least).

Are you sure you didn't mean 100 by 100 I.e. 10k m2? 100 m2 is practically nothing.

Has something changed? I don't see the string "All the food you need" in the article (or anything like it... possibly I missed it).

Anyway, in Japan the amount of rice that one person historically has eaten for one year is 1 koku (which is about 150 kg). The yield for rice is about 0.41 kg per square meter. So you need well over 300 square meters to grow enough rice for one person for one year. That doesn't include other vegetables or fish (or whatever).

In Walden, Thoreau describes self sustenance and living off the land. I believe he talked about farming a quarter of an acre, which is about 1000 square meters. Many people doubt he was able to rely completely on that quarter acre.

You can get some extra efficiency if you build vertical grow systems, but even your 12 m^2 is a pipe dream. Their 3.5 m^2 system will provide snacks at best. So I think you are right to be skeptical.

"All the food you need" is in the video. The 12m^2 is a theoretical limit, like the minimum one could do without violating the law of physics. It's not practical.

For reference, I had a very productive set of raised beds for a while in sunny coastal California that represented about 7.5 square meters. I could produce more than enough vegetables for my wife and I for about 1/4 of the year. That did not include meat, and various carbohydrates like rice, pasta and bread. I felt very confident though, that a with a total of 8 8x4ft raised beds that I could feed us all the fruits and vegetables we needed. We ate quite a few vegetarian meals so when the garden was doing well it reduced our food budget enormously.

There is also all the molecules from the local environment that play a part as well. Sunlight energy is used by the plant to convert the available local resources into a different form. The total energy available isn't limited to sunlight energy, but includes plants' local environment.

Precision agriculture is all about controlling variability. When you have a small plot of all the same soil, under the same weather conditions, your really not going to have much, if any variability, and probably won't actually need any precision agriculture. As opposed to a 400ha paddock where you can have 3 different soil types and different weather at either end of the field.

As for this technology being used on a larger scale, I'd be really interested to see the cost for a similar setup over a few thousand hectares, and what sort of efficiencies might be gained compared to our current system. However I'm not convinced it'll replace our machinery any time soon.

The other interesting use for this would be, if it gained a large enough user base, the sharing of different planting setups and rotations for nutrient replacement. i.e planting crop x and y in rotation, or interrow with crops a and b.

> i.e planting crop x and y in rotation, or interrow with crops a and b.

I think this will be the real win of this technology eventually. Most if not all first world farmers understand that over the lifetime of their farm doing proper crop rotation increases yields but there is still a very reasonable instinctual reluctance to give up the short term profit of a high yield crop in order to plant a rotation crop of less profitable plants to keep the soil up correctly. Evening out that profit per field would probably be a great boon to operations that are too small to rotate entire fields.

It would also probably be an excellent way to get non farmers involved in the food production chain, especially in semi-rural areas. I grew up just in front of a community pasture which served as a grazing ground for three or four different farmer's herds and a couple of people who kept a couple of milk cows for local cheese making and stuff like that. Extending that concept to plants would be great, a giant community garden where people could plant crops that they are interested in and receive the yields at the end of the growing period. Larger farming outfits could plant around the regular people and ultimately keep costs down because of subsidy. Round out five or six gardens with some cash crops like wheat, soy or corn while locals use some of the field to grow their own grocery vegetables.

> Extending that concept to plants would be great, a giant community garden where people could plant crops that they are interested in and receive the yields at the end of the growing period.

> Round out five or six gardens with some cash crops like wheat, soy or corn while locals use some of the field to grow their own grocery vegetables.

Like allotments? https://en.wikipedia.org/wiki/Allotment_(gardening)

Sort of except instead of one person being responsible for each of their plots all of the plots are tended by a small number of farmers with electronic assistance.

Like 3d printing, I think this only solves problems for a small subset of the people who will buy this. In the case of 3d printing, industrial engineers could prototype things much faster, before ultimately sending it to a factory for large scale production. In the case of FarmBot, I can see some agriculture scientists working in greenhouses benefiting from automating the planting, testing and controlling of certain crops. Especially if precise timing and its documentation are needed.

But for the average consumer, this doesn't seem particularly economical. Automation of this kind of thing only makes sense when you have a significantly larger field. I still think this is a great idea, but I think its reach is a bit overstated.

The design looks reasonably scalable.

But for larger scale operations, does it scale any better than the existing technology, which is also pretty much fully automated these days?

Take soybeans. Typically you want to plant around 130,000 seeds per acre. I estimate by the video that it takes about three seconds to plant a single seed. That's four days just to plant a single acre of land. The John Deere DB120 planter claims to be able to plant 75 acres per hour.

Naturally, you can add more of these machines, but with that goes up the cost and complexity. I'm not so certain the economics would favour such a device.

Wouldn't arm design be much simpler though?

An arm would be simpler, but might get awkward for larger plots. Some kind of self-driving vehicle with an arm might be the most scaleable.

They didn't really explain how it pulls weeds.

The nerd in me has my credit card out and is ready to overnight one of these things to my backyard.

The suburban dad in me scoffs because one of the best parts of having a garden is planting stuff with my kids and helping them water and care for the garden.

Robots vs kids....

Hmmm. That gives me another idea...

They punch it underneath the soil so it can't receive sunlight. They show it happening in the video.

See also : Marcin Jakubowski on Open Source Blueprints for civilization : https://www.ted.com/talks/marcin_jakubowski?language=en

..and the construction set: http://opensourceecology.org/gvcs/

How has that been going by the way? I've been following OSE since 2010 and I've lately been given the impression that it has been dying down.

Seems like mostly what they've produced is TED talks.

That's a bit snarky. I was pretty pumped with the project at first, but it really seemed poorly run, and also seemed to ignore a lot of prior art when designing hardware. Their compressed block machine does seem fully thought out and functional at least, but their first tractor snapped in half on its maiden voyage, and the replacement looks like a post-apocalyptic bobcat. There's a reason most farmers aren't using bobcats in the field.

Anyways, neat idea, and I hope it inspired others to focus on open source hardware for agriculture and small-scale civilization.

Marcin was not able to lead the project coherently. Groups of contributors joined for fairly short periods of time and then left, often disappointed and even angry. No idea, if Catarina has been able to influence him for the better. Let's hope so.

I have no idea -- I just dug it out of my bookmarks.

When I looked at the progress of the machines I had a sneaky feeling it might have slowed down - I was interested in the "aluminum extractor from clay" a few years back :)

A bit much for a the simple one time task of planting. Maybe if you need to plant reproducibly a large variety of seeds? But this use case is rather uncommon.

For a more novel approach to farming, see the combination of fish tanks and planting, called "Hydroponics":


It also waters and weeds with interchangeable heads. Weeding is done by pushing anything unexpected back underground. Watering is done by dropping water on the plants.

This isn't right. You water the soil, and the soil moisture is uptaken by the plants. Watering the area immediately around the stem doesn't work, except for growing seedlings.

Ignore this article (which I think was on YC a few weeks ago), and follow the link in it to Modern Farmer. Their article links to the AgBot Challenge.[1] That's both more interesting and more useful.

Autonomous weeding robots which roam big fields and take out weeds one by one are being demoed by several manufacturers. Here's the most practical-looking one, out in a huge field of beets, quietly killing weeds one at a time.[2]

[1] http://modernfarmer.com/2016/05/2016-agbot-challenge/ [2] https://vimeo.com/140036163

Regarding weeding, by virtue of the precision overhead watering the weeds are likely to only grow at the base of the planted crops - how would it differentiate between the two (or more)?

There are sensors designed for broadacre cropping that will pick weeds out of crop.[0]

As well as sensors that will pickup weeds from dirt.[1]

This would be in the middle of those two problems. There isn't a dense crop to pick the weeds out of, but it's probably slightly too dense to just pick them up from the dirt.



Once I had idea about growing plants as service. You would not need to have your own piece of land but you could rent space and check your plants over the internet with mini game where you control watering choosing the fertilizer and other things (Farmeville but with real plants). You could look at your plants through ip-camera or receive pictures by mail. When plant would grow then you could have it picked up and shipped for lunch.

With this farmbot I can imagine it would be doable.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact