ChatGPT in AgTech: The dumbest smart person I know
Feb 28th - My 29th Edition to Easy Observations in AgTech
Visit my website www.easyobservations.com for more info about me or follow me on LinkedIn and Twitter
Opening Thoughts:
Okay, Okay…I’m late at this weekly thing again. In my defense, between a dad and daughter(s) vacation, blizzards, and sickness it basically took two weeks from me. Anyway, here I am again.
With that said, I figured I would follow up on a Linkedln post I put down well two weeks ago when I thought I had more time to write. Here it is for anyone who didn’t see it or is crazy enough not to follow me yet.
Basically I asked ChatGPT a simple question, “Why does Agtech have a tough time showing value?”
While I wasn’t all that surprised in the results as I’ve seen and heard all these points before many times, I did like the speed and simplicity. In some realm maybe its answer might have come from some snippet I’ve even written about in the past. That or likely some other colleague of mine. We’ll never know…
For this weeks edition though I thought I’d break down what it gave me and dive deeper into the nuance of its answers. Also, it’s not to say I think ChatGPT is dumb or other AI/ML processes in all of tech. It’s more or less the smartest way right now to come up with the simplest dumbest answers we already know or know how to create.
Maybe I shouldn’t say dumbest smart person, but fastest dumb person I know. I don’t know and in all reality I wrote the headline to grab some attention so yeah, dumb too.
What did it say???
Here is what it exactly said for reference….
Me: Why does Agtech have a tough time showing value?
It: There are several reasons why agtech may have a tough time showing value:
1: Lack of standardization: The agriculture industry is diverse and fragmented, making it difficult for agtech solutions to be standardized and universally applicable. This makes it challenging to accurately measure the impact and value of these technologies.
2: Complexity of agriculture: Agriculture is a complex and dynamic system that is influenced by many factors, including weather, pests, soil conditions, and more. This makes it difficult to attribute changes in crop yields or other metrics to a specific technology or intervention.
3: Long-term ROI: Many agtech solutions have long-term benefits, such as improved soil health or reduced input costs, which may not be immediately apparent or quantifiable. This makes it challenging to show a return on investment (ROI) in the short term.
4: Data challenges: Agriculture generates large amounts of data, but the quality and availability of this data can be inconsistent and unreliable. This can make it difficult for agtech solutions to demonstrate their value based on data-driven insights and analytics.
5: Limited adoption: Adoption of agtech solutions is often slow, due to factors such as limited access to capital and technology, lack of awareness and education, and resistance to change. This can limit the impact and value that agtech solutions are able to demonstrate.
Despite these challenges, the potential benefits of agtech are significant, including increased efficiency, reduced waste, improved crop yields, and more sustainable and environmentally responsible practices. By addressing these challenges and working to overcome them, the agtech industry can continue to drive innovation and demonstrate the value of its solutions to farmers and the broader agriculture industry.
So just that, not bad, not special, overall okay.
What does it all mean??
Well, let’s break that down some and overall that is what this edition is all about. Something like ChatGPT can yes grab all the information and compile it in a nice little snippet like above. That is awesome and all, but I don’t think it can show the nuance and reasons these things actually are the way they are.
Most people can’t really understand that in any realm and explain it and even if they do writing about it to do just that is very hard and complex to do as well. You need many things to understand that nuance and that comes from time, experience, blood, sweat, tears, and much much more.
Maybe with that all said I can take what it said and simplify why these things are true in a simple way much like ChatGPT just stated. Here it goes…
My Interpretation:
fMost could dive into their own thoughts on these to an unhealthy extent I would say. So, I will try to call out the simplest issue(s) and reason for each of these “AgTech Adoption Issues” our computer overlords are trying to explain to us.
1: Lack of standardization: The agriculture industry is diverse and fragmented, making it difficult for agtech solutions to be standardized and universally applicable. This makes it challenging to accurately measure the impact and value of these technologies.
In my days working with the group AgIntegrated, now Telus Ag, this was the reason the company existed and got acquired. We would help companies translate like 300+ different types of Ag data formats from all sides of the equation. We worked with AgGateway to help standardize workflows and come up with standard practices.
You’d think we’d all be better off by now. We aren’t and won’t be for some time and why is fairly simple. Most of Ag still uses older legacy systems and hardware. The other part is many Ag companies have different ways their equipment, sensors, and beyond create data and they want to keep a unique advantage. This isn’t Apple vs Microsoft type data formats, this is horses vs cars vs planes vs boats vs trains type of thing. So many different things and variants to come together. It’s tough let’s just say, but some like Telus Ag and Leaf are trying to help in the ways they can. It’s getting better-ish?
2: Complexity of agriculture: Agriculture is a complex and dynamic system that is influenced by many factors, including weather, pests, soil conditions, and more. This makes it difficult to attribute changes in crop yields or other metrics to a specific technology or intervention.
Sounds pretty good at first, but where it misses the beat is the literal complexity of culture within the word itself of Agri-culture. I’ve talked about before that Ag is more complex overall than understanding the human body via medicine many times. It’s not a perfect one to one, but Ag is incredibly biologically complex and that goes way beyond the basics of say soil, microbes, crops, pests, and beyond.
How and why people farm or why the ones that help them and how they do it vary greatly even between different towns 10 miles away or neighbor to neighbor. How each one thinks and uses different practices between those factors of weather, pests, soils, and more as it says becomes exponentially more complex. Add in the other top 5s and yikes, yeah complex.
3: Long-term ROI: Many agtech solutions have long-term benefits, such as improved soil health or reduced input costs, which may not be immediately apparent or quantifiable. This makes it challenging to show a return on investment (ROI) in the short term.
Now ROI and AgTech can mean a lot of different things. Probably the biggest factors that influence ROI in general for most farmers are not what most actually think beyond one, which is weather. The other is really around marketing the sale of the crop. Just look at current food and commodity prices and you’ll understand. No equipment, precision agronomy, genetics, biologicals, or other special recommendations can compare to the craziness that market prices can swing your ROI in such a short time. That also is corrupted by weather naturally. In the last two years for instance commodity prices have basically doubled. Name something that has a larger affect in what you make in money in that short of time in Ag beyond weather.
So, while many tackle those other things I mentioned they just can’t deal with the craziness of prices and the weather that can influence it. That effects how everyone, especially farmers think about everything. Yes some are tackling the marketing of crops better, but it’s still the wild west. Basically, how a rando on the Chicago or Kansas board of trade feels one day about a hunch will do more of a shift in ROI than any AgTech add-on in a year. It’s really a crap shoot. Now, hedging your risk is what those other things give you in helping you control better the things you can while giving you a better chance at success. Regardless, hard to interpret and get growers and others to think through at times.
4: Data challenges: Agriculture generates large amounts of data, but the quality and availability of this data can be inconsistent and unreliable. This can make it difficult for agtech solutions to demonstrate their value based on data-driven insights and analytics.
So this one I think also is combined with number 1 as well as you can’t put together all this data if you can’t have an easy way to put it all together and that takes translation and standardized practices to do just that. With that said, data challenges is very broad and while getting it from someone or having good data is a huge problem I think there is more to it.
Data challenges in Ag and AgTech in general are a problem because most don’t care about data after the fact. Most farmers for instance love to look at data, but mainly in real time and they have short attention spans around it. They want to instantly see the yield, moisture, seed count, down pressure, fuel level, tire slippage, seed/chem/fertilizer rate, irrigation running or not, temp and wind speed now, and on and on. They want it now, they don’t want to go over the old numbers for the most part afterwards and go through all of that. It’s not that they don’t think of it or work with it, but most in Ag are used to the result just being what it is. They’ll shift things a little next year and keep going.
Farmers as a whole like to see things when they happen as in that moment it makes the most sense to them as they also have a hundred other things to deal with. They will take that moment when they saw that data when it sort of happened and store it for later in their mind. That is sort of the evolution of farmers and how many work, not all but I’d say more than you think or realize. Something to think about.
5: Limited adoption: Adoption of agtech solutions is often slow, due to factors such as limited access to capital and technology, lack of awareness and education, and resistance to change. This can limit the impact and value that agtech solutions are able to demonstrate.
This one is sort of generic so I’ll break down each of the pieces it mentioned sort of. I don’t think it is lack of access to capital or tech, for smaller growers and certain niches yes that can be true but I don’t think as a whole for the majority of production. I think most are aware, though they probably are more ignorant of the tech or just don’t care to deal with it and learn. That of course pushes into the awareness, education, and resistance. Those are very true and for very different reasons. I’ll break that down a little.
Awareness is an issue as so much is changing so fast and people just can’t keep up. That dives into education as once they learn something new in AgTech that is promoted to them or its something they purchased there is already something new and better that is there and being pushed to them again. This leads to resistance to change as once they have something they know and are educated in they become comfortable and don’t want to change what they know. Being a farmer that is uncomfortable in how they farm is probably the biggest thing a farmer tries to avoid. Think of that for a bit and the rest should fall into place.
In Closing:
And that is about it for the most part. I mean, I could do an edition of this newsletter on each of those points easily, but I think I’ve done enough for now. I more wanted to bring something up I don’t think most think of regarding each of those realistic points in why AgTech has value issues. It is so much more than just a paragraph and sentence. While ChatGPT you are a formable thing to help bring up things, you just don’t get it all yet. That is fine by me.
For how “smart” ChatGPT and other AI/ML processes are perceived in this world, they just are scaleable ways that copies the processes of many many things dump people can do more or less now. Think of it this way, if you had a 60k person stadium full of generally “dumb” (ignorant or non-knowledgeable people for you complainers out there) people asked the same question that you had to Google search as a whole and showcase all of the top five answers you’d get what you see above more or less.
That is where I think it is at this point. Of course though, with some nuance…
Thanks for reading Easy Observations! Please Share, Subscribe, and Comment if you would be so kind. I’d also be happy to get together and meet if you’d like to talk more in-depth about AgTech or Earth Observation in Ag. Feel free to contact me at my website. All the best!