[thomas_howland]

The amount of labor involved in grain crops like barley per dollar of output is minimal already - the guy running the tractor/combine/etc is mostly making sure nothing breaks and keeping an eye on things (which is important when you're operating $N00,000 worth of heavy equipment on open ground). You do need guys, however, for when things do go wrong, and for all the other farm tasks (equipment maintenance, fencing, maintaining irrigation, and so on).

[joncp]

Imagine, though, a time when the automation reaches a stage where all weeding and most insect killing can be done mechanically with a robot instead of chemically with pesticides. That will be a huge boon for the environment and for farmers' bottom lines.

[robotresearcher]

Peter Corke and his huge team at QUT are doing great stuff in this area. They have spot-spraying robots, that zap only weeds and fertilize only crops, and also a great device that just mechanically smashes up weeds one-by-one.

Peter has an admirable ability to choose projects and deliver the goods on them. Robotics needs more like him.

Interview with Corke on this subject:

http://robohub.org/robots-zero-tillage-robotics/

Homepage with free stuff including books and MOOC:

http://petercorke.com/wordpress/

edit: I'm not affiliated with QUT and have no conflict of interest. I'm an academic that does related work but not in agriculture.

[reachtarunhere]

Peter Corke is an amazing guy. Loved his MOOC. Highly recommended if you are just curious about robotics. MATLAB is a sad aspect of it though.

[RoboticsAU]

Peter Corke launched a new site recently, which is based on his MOOC: https://robotacademy.net.au/

He outlines the reasons why they shifted away from MOOCs in the Robots Podcast episode #239.

I recently made a site too, for anybody interested in robotics/AI: http://roboticsaustralia.org

[randomdata]

The environment perhaps, but would it really be better for the bottom line? Mechanical weed control already lost out to chemicals because the chemicals were cheaper.

On top of that, chemical applicators are relatively simple and proven technology with few moving parts. Farm equipment isn't like your car that can go 100,000 miles without problem. Moving parts on farm equipment will break – constantly. On top of that, computer technology does not come cheap when it is low volume specialized systems, built to withstand the harsh environments of agriculture.

I have no doubt the technology is possible, but I do wonder if it is possible in a price range that would actually improve the bottom line.

[uuilly]

Yup. A few days ago we (60 people) got bought by John Deere for $305M b/c our machine reduced herbicides by 90% using CV/ML. Our only moving part was a solenoid. Would not have survived if it had been more complicated. You clearly get it.

[wikibob]

Sounds really cool. What company? Do you have any stories you can share here or on a blog?

[uuilly]

https://agfundernews.com/blue-river-clear-leader-john-deeres...

Many stories. Will blog soon :)

[maxerickson]

It only has to be cheap enough to make banning various pesticides reasonable.

[randomdata]

You bring up a good point that I forgot about. Some of these chemicals, like Roundup, are not just for pest control, but serve other purposes such as desiccation. Since you are going to need a chemical applicator anyway, a second machine to duplicate functionality further increases the cost ineffectiveness.

Not only that, but what I am sure most would consider the most devastating type of chemical in recent times, neonicotinoids, are applied as a seed treatment, planted with the planter. Even if you could devise an advanced machine that continually keeps pests from eating the seed while it is buried beneath the ground, that requires trips over the field that are currently not required at all. It is tough to compete with no additional trips over the field. Energy and wear is expensive and potentially devastating to the environment, including issues like carbon emissions and topsoil depletion.

(Neonics are already banned where I farm, but it is a good illustration of how chemicals can sometimes eliminate the need for machinery altogether and how the robo-pest terminator has to be able to compete with that)

[kuschku]

The robots don’t have to necessarily compete with traditional farming.

They just have to make Naturland/Bioland/Demeter style organic farming (which except for very rare cases not even allow any chemical pesticides or herbicides) cheap enough that it becomes competitive.

[zaroth]

I think it's a myth that Roundup is used as a desiccant.

However the robots are not trying to replace chemicals. As you point out, that's a losing proposition. But they can definitely reduce overall man-hours, and so provide constant value as long as they operate.

[randomdata]

> I think it's a myth that Roundup is used as a desiccant.

I can say from personal experience (I am a farmer) that it is not. My white beans get the roundup treatment. They would never dry down in time for a successful harvest otherwise.

In the olden days before roundup they used to pull them and leave them to dry before harvest, but that brought its own issues, including requiring many more trips over the field. That is a costly endeavour, including needing multiples more greenhouse gas emitting fuel, which has its own fair share of environmental impacts.

> But they can definitely reduce overall man-hours, and so provide constant value as long as they operate.

Even a simple autosteer system, which does not even replace the operator, can cost more than many farmers would spend on labour in their lifetime. Low-volume specialized farming technology is insanely expensive.

What I believe is primarily pushing the technology is difficulty in finding labour, especially skilled labour. Farmers are seeking more automation to simplify the tasks enough that you can throw any random person on the machine and get a quality result out of their work.

Operator-less equipment is inevitable, but there is no way the manufactures are going to let the farmer capture any potential gains. It is going to be priced to meet the costs of labour. The farmers will still choose it because they are having trouble finding labour in the first place.

[grogenaut]

And what price point is that. Pesticides are pretty cheap

[maxerickson]

The impact on the final price of the produce matters more than the comparison to pesticides.

Like if pesticides represent 0.1 percent of the retail price of some vegetable, I personally would consider it reasonable to spend 10x avoiding some pesticides. Because it would only have a modest impact at retail.

[mikekchar]

The price of wheat in the US is currently ~$170 per tonne. Retail price of flour in the US is currently a bit over $1100 per tonne. Fertiliser and pesticide cost seems to be about 23% percent of revenue, which would work out to about $40. So fertiliser and pesticide cost is currently about 3.6% of the final retail price. Figures come from random google searches, but I think they are about right.

There are a few things you have to understand about these figures. The first is that only 15% of the retail price is the price of wheat. The rest is shipping, processing, packaging, waste, advertising, and profit. You can actually work out how much each bit actually adds to the price by looking at the price at each step. For example, if you look at wholesale wheat flour prices, you are looking at about $3-400 per tonne.

The reality is that distribution is the vast majority of the cost and each entity that touches the product along the way wants a big markup. Money costs and each party wants a return on investment, not a fixed return.

So, if we replaced fertiliser and pesticide with something that costs $400 instead of $40 per tonne of wheat, then the price of the wheat will go to $530. Ideally that will result in a retail price of ~$1500 (about a 25% increase), but in reality it's likely to be at least double.

Personally, I think this is worthwhile -- especially if the extra $360 per tonne would allow smaller farms and provide more jobs in the farming industry. I don't think it is realistic for 2 main reasons. First I have no idea if even $360 would be enough. Second, large corporations and rich people have a very vested interest in continuing the current trend of moving all of the price into distribution. These distribution channels can be controlled relatively easily and provide rent for them. It also keeps the world food price below the level that traditional farming can support. This means that rich countries can dismantle the agriculture industry of poor countries and then control them. I encourage anybody interested in this kind of thing to look at the conditions attached to government loans/subsidies to poor countries that are used to buy food. When you start looking at this stuff, you are playing with big players who have lots to gain/lose and who aren't going to give up anything easily.

[losteric]

I am interested to see how the pest ecosystem evolves in response to robotic "predators"... Can insects evolve effective camouflage? Will weeds evolve to mimic cash crops? How will fungi/bacteria/viruses respond to a lack of other pests?

[agavegoose]

Something very similar to that has already happened - in (very often hand-weeded) rice, shattering appears to have evolved to de-domesticate the crop - making it essentially impossible for 'weedy' rice to be distinguished from the real thing - https://www.sciencedaily.com/releases/2013/07/130717132418.h...

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2988683/

[Animats]

It's here now. Blue River makes LettuceBot, which is a machine-vision system for killing weeds in lettuce fields. The machine is towed behind a tractor, looks down at the lettuce and weeds, and zaps weeds and too-close-together lettuce plants with targeted sprays of concentrated fertilizer.[1] It's a full neural net vision system, with many cameras and software borrowed from face recognition. They train the system with pictures of weeds and pictures of good plants, and as it rolls over the field, it targets the bad stuff and sprays it.

This isn't experimental. It's now owned by John Deere and is currently doing 10% of the US lettuce crop. Next, cotton. There's no problem with plants becoming resistant to weed-killers. Unless pigweed learns to look like lettuce, it's going to get zapped.

It's sold as a service. For $220 an acre, Blue River comes and zaps your weeds. Now available in the Salinas Valley and Yuma, AZ.

[1] https://www.youtube.com/watch?v=-YCa8RntsRE

[dejv]

Insects are never gonna be killed mechanically, those are too small and too many in count. Modern trend is to use biological controll: you "plant" organism that eat them or eat their eggs.

[dejv]

Just to expand on my comment, biological control is quite varied, but the main theme is selectivity: you don't want to destroy all the insects or butterflies, just those who do harm to your produce.

So you basically have couple of forms, either some microbes, viruses or some form of insects that eat on just one type of insect or of insect class. Another main form is pheromone confusion thingies: you put them in your orchard/vineyard/market garden and it make hard for males of your target pest to find female and procreate.

[PeterisP]

Certain pests (mainly large beetles or their larvae) used to be controlled by removing them manually - it's more expensive than pesticides, but it was feasible (and worth it!) to do it simply by bunch of people going over the crops. This means that they could be controlled mechanically in the future.

[njarboe]

If you had a system that could target such small things, lasers might work.

[gcb0]

if you simply build a decent greenhouse it's already bug and weed free. plus your crop can survive unexpected frost.

adding high maintenance robots is just silly.

[ams6110]

Farming crops like wheat and barley in greenhouses is not feasible at the scale of the market today.

[gbrown]

I've daydreamed about trying exactly that on a small scale. Industrially, that would be amazing for the environment.

[ModernMech]

Especially if said robots can run on the weeds and bugs they are removing!

[QAPereo]

So basically ladybugs, but with a profit margin and TOS.

[losteric]

Not invented here, not relevant! /s

[danblick]

Have you seen farmbot.io?

[gbrown]

No, looks interesting, thanks.

[shanev]

We are destroying topsoil at an alarming rate due to industrial farming that we may have no choice but to go this route.

[thomas_howland]

We are absolutely not. Productive farmland in the USA is fairly expensive and farmers are extremely interested in preserving its value. Most "destroyed" topsoil is due to eg suburban developments spreading into former agricultural areas, but it's still minimal in the grand scheme of things.

[Dylanfm]

Development of arable land is a problem, but our agricultural practices are a major one. I think shanev is speaking about the effects of modern agriculture on top soil. We should prioritise the _creation_ of soil. Instead, top soil is being degraded and then lost though erosion. There are other issues, such as the planting of monocultures, overuse of nitrogen fertilizers and so on.

As climate change brings us longer droughts and larger downpours, we should be farming in ways that build top soil for more resilient farmland.

(edit - adding link to http://www.fao.org/docrep/t0389e/t0389e02.htm)

[shanev]

Yes this is exactly what I was getting at. Thanks for putting it more eloquently than I did. I’m just used to getting downvoted anytime I talk about monoculture farming, soil, and climate change so I don’t put effort into it anymore.

I think the pro-GMO crowd thinks talking about sustainable farming is anti-science when it actually involves more science at a macro level to understand the effects of synthetic monoculture farming on soil, and it’s effects on climate change.

[Dylanfm]

There's science and technology all the way along the sustainability spectrum... but it isn't evenly distributed.

I'm new to growing and enjoying the steep but rewarding learning curve. Soil is an amazing thing and I think humans are only just getting a better idea of what's going on down there. This recent conference talk helps shed some light: "Building Soil Health for Healthy Plants by soil scientist Dr. Elaine Ingham" https://youtu.be/xzthQyMaQaQ

[jamiegreen]

"One-third of world’s soils are degraded: FAO"

http://www.livemint.com/Politics/NVc4UYKzFxn8zeQqkydyPM/Onet...

[Gustomaximus]

I think this is the key vs the planting side. Maybe soil monitoring and watering/fertilising at a micro level.

If someone can make a drone to fly around and take out pests like fruitfly that would be a massive step forward vs spraying.

[lightedman]

"Imagine, though, a time when the automation reaches a stage where all weeding and most insect killing can be done mechanically with a robot instead of chemically with pesticides. "

Why bother with that when we already have controlled facilities where pests simply can't gain a foot hold in the first place, making that labor totally unnecessary?

[njarboe]

Greenhouse the entire mid-west? With big enough structures to run tractors? Not sure the ROI is going to work out.

[lightedman]

I don't think you understand how indoor vertical farming works. In 1/8 of an acre you can produce an entire acre of crops, utilizing upwards of 99% less resources.

To boot, you don't even need tractors to harvest. You can just pull the entire tray yourself once the NFT solution's drained from the channel.

I can take 1/8th of an acre of land, and produce a full acre of wheatgrass in two weeks and harvest it myself with no mechanical methods required in less than a couple of hours.

https://www.youtube.com/watch?v=9ZTikdxj8AI

I was doing this 20 years ago in high school horticultural science class.

[rasz]

What farmers? there wont be any farmers left, just mega corporations.

[burger_moon]

Some of my family members drive truck at farms during harvest and I've rode along a lot. The guy driving the combine can mostly be automated except the part of having to stop and clear out metal or other jams that stop them. The machines have sensors which shut them down instantly when they detect metal because the cutter is expensive as hell. Not sure there's really a way to automate that. There's a lot of other less common things that cause the chopper to stop but foreign debris is the most common.

As far as field trucks go, well they're mostly old 15+ years with new-ish trailers because well trucking doesn't pay shit so it'll be long time before self driving trucks replace the current fleet (you have to figure the new trucks on the road now will because part time field trucks 20 years from now). If can get GPS accurate enough you could automate a lot of it, but trucks get stuck all the time in the field which mean the guy driving the chopper has to stop and pull the truck out.

It'll be tricky to automate all the weird oddities you come across out in the field. In the end these guys aren't making very much so the cost benefit doesn't really seem to be there but we'll see.

[dejv]

I had one vineyard planted this year using fully automatic planter using gps and laser mavigation. There was driver inside, but he was just a controller: once he moved few cm out of the path whole thing just shut down and he had to drive back to spot.

[Houshalter]

Yeah this isn't terribly exciting. People have been demoing self driving tractors and farm machinery for decades. Ford had one in the 1950s. Most of the hard work like actually picking the crop has already been solved, and making the tractor self driving just saves a few hours of work at most.

What would be really cool is automating crops that still require lots of manual labor. Like vegetables. That's the reason they are still so expensive. An automated greenhouse would be enormous news.

Whats also cool is stuff like this (https://spectrum.ieee.org/automaton/robotics/industrial-robo...) which eliminates the need for pesticides.

[bamboozled]

That's the reason they are still so expensive.

My observations have been that vegetables are so inexpensive they're often thrown away?

Obviously these cheap crops are often grown in a monetarily cheap but environmentally destructive way but I've no idea where you're seeing these expensive vegetables for sale?

[Houshalter]

I meant expensive in a relative sense. A pound of corn is a few cents, a pound of vegetables is a few dollars (it's far worse when you compare price per calorie.) Maybe expensive is the wrong word to use, but the economic impact of automating this would be quite enormous. And I have seen people complain that eating healthier is more expensive.

[chiefofgxbxl]

> What would be really cool is automating crops that still require lots of manual labor. Like vegetables. That's the reason they are still so expensive.

Same deal with coffee beans, which contributes to their higher price. The coffee beans do not ripen at the same time, so they must be hand-picked so the worker can identify which ones are ready. Lots of manual labor.

And with coffee being so widely consumed, savings here would easily be felt by a lot of people, although I doubt Starbucks would like it...

[setr]

Why would starbucks be against lowered supply prices? Either the whole market drops in price and starbucks keeps the same profit margin but lowers actual price, or starbucks just sells the same price but now with higher margins than before

[kuschku]

Because Starbucks has found a carefully balanced point between being affordable enough that it has mainstream appeal, and expensive enough that it seems premium.

Any change, absolute or relative, would reduce their margin.

[philh]

Do you claim that the current price of coffee beans is the price that happens to maximise Starbucks' profits? (This seems implausible to me, unless Starbucks has fairly strong control over the price of coffee beans.)

Or is your claim more narrow than that?

[kuschku]

My claim is that the current price of coffee beans happens by coincidence to be close to the optimum for Starbucks’ profits, and any major change would be a disadvantage.

[Alex3917]

> The amount of labor involved in grain crops like barley per dollar of output is minimal already

What crops have the highest labor costs, in both absolute dollars and as a percentage of what the consumer pays?

[thomas_howland]

Hand-harvestable-only crops like lettuce and tomatoes. Basically the CA Central Valley.

[Alex3917]

Interesting. Between farm labor automation and self-driving cars, do you foresee a point where it's cost effective to give certain foods away for free and monetize via advertising? Even if it's slightly offbeat stuff like dandelion greens and seaweed or whatever?

[prawn]

If anyone wants to eat thistles, I'll give those away... Oxalis too.

[Alex3917]

> If anyone wants to eat thistles, I'll give those away.

How do they actually taste? I know they're in the artichoke family, but I've never actually dug one up or whatever.

[prawn]

Never tried them, but they are a painful weed here, grow quickly and seed very easily. We used to give the seed heads to our canary once upon a time, but currently have hundreds of thistles to zero canaries.

Perhaps if they could be blended and used as a base for bread or cakes, but I have no idea if they're safe in large quantities.

[xefer]

I wonder if there will come a time when buying food harvested by stoop labor will be considered immoral. Perhaps human diets should target foods that can be produced through mechanization where possible.

[mikepurvis]

But lettuce and tomatoes are delicious. Hence the interest in figuring out to do it robotically, either in the field or greenhouse.

[njarboe]

Some types of tomatoes can be pick with machines. Here is a video from New Zealand: https://www.youtube.com/watch?v=R3EpFTyN26E

[njarboe]

The Salinas valley is where most of the lettuce comes from. Quite a bit cooler than the central valley. Baby greens harvesting is highly automated. A nice picture here: http://www.hortech.it/en/cp/agricultural-machineries-for-har...

[KGIII]

Blueberry costs are about 50% labor.

[nine_k]

It's a very nice proof of concept. Also I bet these guys collected a lot of info important to develop robots for more finicky crops. This could be the most valuable part of the project.

[stephengillie]

This is essentially the argument against automation - that a human is so versatile, we can be "programmed" (convinced, paid, trained, etc.) to perform many disparate tasks. Our exception handling is so robust (what happens if it floods?) that it's a problem solving engine unto itself. As with a grocery store, the self-check machines can't sweep the floor or stock shelves when business is slow.

[jamesrcole]

> You do need guys, however, for when things do go wrong, and for all the other farm tasks (equipment maintenance, fencing, maintaining irrigation, and so on).

How far away is automation of those kinds of tasks, I wonder? I don't have much of any idea, myself.

[tinco]

Very far away. Computers can't even drive cars autonomously yet, which is a far simpler task. As it stands computers can't deal with any sort of irregularity that humans would deal with through basic creativity. I've also not seen robots plan fine manipulation yet, even though I feel that should be achievable with current technology.

[ge96]

Wasn't there a huge multi-rotor that cost $15,000 haha, it was water proof I think and had radar... crazy. I really like the concept of little robots running around doing their thing. Like those under water drones that are in the oceans.

[oh_sigh]

True, but that is because the farmers have access to huge fields which let them deploy their equipment efficiently. Autonomous farming would let you theoretically use more marginal lands which are not currently suitable for manual farming.

[randomdata]

I expect you are suggesting that small parcels of land that are far away from large fields are not worth the travel expenses? I'm not sure that robots really improve on that, as fuel, wear and tear, etc. going field to field remain a big detractor. There are plenty of farmers who do it for fun. They'd be happy to spend the time in the machine if labour costs were the only reason to avoid those fields.

I already farm a field that is one acre in size and one that is three acres because they are near to a large field that does justify the travel costs. Any that fit that description are likely to be already in production.

[oh_sigh]

Not necessarily small - mostly marginal(I should have mentioned that it wasn't necessarily sheer size of field, but something like size*quality in my original post). I don't agree with your characterization about farmers doing it for fun. I'm sure there are a lot of people, say, within 10 miles of a major city who would farm for fun, but outside of that radius you mostly have people who either already have farms or manually farm on their own house plot.

[randomdata]

> I should have mentioned that it wasn't necessarily sheer size of field, but something like sizequality in my original postl*

Understood. Marginal traditionally refers to the quality, but I wasn't sure how you thought the math would work. And since you mentioned size before, I went in that direction.

I am still unsure of how the math would work out. I have a farm that is already bordering on marginal and even ignoring labour costs it is difficult to turn a profit. Honestly, labour costs are a drop in the bucket. It costs hundreds of dollars per acre for input costs (seed, fertilizer, fuel, etc.), there is a significant capital cost to having the land and the machinery, while the labour if I paid myself minimum wage would be maybe $10 per acre over the course of the year. And I think that's pushing it. That is only 2% of the input costs when my other input costs are $500.

If a field is only 2% shy of not being marginal, I'm certain there is already someone trying to make it work.

> I don't agree with your characterization about farmers doing it for fun.

Well, that's why I do it, so I have some experience there. If it was about the money, I know software development pays far better.

> but outside of that radius you mostly have people who either already have farms

I was assuming that those who farm for fun would already have their own farm. However, now that you mention it, I always see the retired farmers are still keen to get back into the equipment, so it is not strictly limited to those who have their own farms. Living hours away from the nearest city, I don't really see it being an activity of those near the city.

[jsilence]

Also leaving an expensive robot on a small remote plot might not be a good idea because of theft.

[agumonkey]

How much thoughts have been put to improve farming beside fertilizers and pesticides ?

I do wish for better food, and better life for farmers. Also I'm quite annoyed by mainstream tech (find it often uselessly improving the wrong parts of life).

[imtringued]

There is still the time advantage. Having two autonomous tractors is twice as fast as operating one by hand.