Matt and Sean talk about farming techniques that incorporate AI, solar panels, and a whole lot of science. Agrivoltaics is a promising technique … with some nuance.
Watch the Undecided with Matt Ferrell episode, “How Solar Panels Are Changing Agriculture – Agrivoltaics Revisited” https://youtu.be/ww-_U7_oQbY?list=PLnTSM-ORSgi7uzySCXq8VXhodHB5B5OiQ
YouTube version of the podcast: https://www.youtube.com/stilltbdpodcast
Get in touch: https://undecidedmf.com/podcast-feedback
Support the show: https://pod.fan/still-to-be-determined
Follow us on Twitter: @stilltbdfm @byseanferrell @mattferrell or @undecidedmf
Undecided with Matt Ferrell: https://www.youtube.com/undecidedmf
Let me give you the Cliffs Notes version. My beard. It is gone. So we’ve entered an unnerving era for our viewers on YouTube who are currently wondering if they should be wearing 3D glasses because the left frame and the right frame, well, they’re pretty identical at this point. They’re pretty similar.
Pretty similar. Creates an unnerving double image right down to the fact that Matthew and I, independent of one another. Go to Warby Parker and pick out basically the same pair of glasses. I, it doesn’t, I don’t know how to explain this any more than that. Before we get into our discussion, which as always is focused on the videos from Matt’s main channel, which is undecided with Matt Ferrell.
Wanted to share some comments from our previous episode. This was on our discussion from Aquaponics. Or on aquaponics if I knew how to speak properly. This is from our discussion on Aquaponics and there was this interesting comment from Reckless who wrote, I am in the process of building, designing my own system for indoor aquaponics with the intent of going commercial after my garage.
Prototype is fully functional. I’ve been working on the design for over a year. A variety of my design features are not part of any system I have seen so far. So it’s both thrilling to see the topic here on the channel and terrifying to be reminded that the patent talk, the patent clock is ticking. If I make it to the commercial pilot plant phase and don’t flame out like so many others have, I’ll reach out to you guys.
I thought that was rather interesting. Reckless, best of luck and do keep us Yeah. Appraised of what you’re doing. And I would even suggest this, even if this doesn’t work out, I think Matt would probably be interested in getting a, a note from you Yeah. About what you’re doing and how and why would be an interesting conversation to.
Regardless of how this pans out. Not to say that I hope to hear from you if you are in a sad place, but best of luck. Keep us luck. Keep us in the loop as far as how things are going for you. So onto our current discussion about Matt’s most recent video. This is his March 28th, 2023 episode, how solar panels are changing agriculture.
Voltaic revisited this. A part two of a topic that Matt talked about. Was it roughly a year ago that it would’ve been, it was a little over a year ago. Yeah. Yeah. And this is the, this is the science of using solar panels and agriculture together. So you have panels above farmland. And the developments that Matt has been talking about in this most recent episode is to revisit, you know, the concept in its entirety, more recent research, but also how AI.
Is being incorporated into it. So his episode talks a little bit, not just about Agri Voltaic, but just about AI in general. So what did you find in your research about the AI that was the most surprising? Well, it’s
one thing I wanna bring up is, and this is something I’m guilty of, so I’m putting my finger at myself, we use the words.
Interchangeably with machine learning. Yes. And it’s not the same thing. And people have pointed that out in the comments and you’re absolutely right. Yeah. So I just kinda point that out. So when we say AI right now, we may say that from time to time we’re using interchangeably, so just, uh, cut us a little bit of slack.
Yeah. But yeah, the thing that surprised me most was how they’re using this, uh, the, the machine learning algorithms to like figure out how specific crop. Need to grow cuz like whatever solar panels that those solar panels are shielding, whatever crop, it can adjust itself based on how much. And how much rain that specific crop needs to adjust the tilt of the panels and, and it’s not just tracking the sun to get the most production for the solar panel, it’s also taking into account how much of that sun needs to reach the crops to make it go to as big of a yield as it can.
Great Get. So it’s, once again, it’s this dual use. Hey, we can have this symbiotic relationship between everything, where the solar panels are helping reduce the evaporation and the loss of moisture from the plants. And the plants help keep the solar panels cooler, which makes them work better. And the solar panels can shield the, the, the plants so that they don’t like it too much sun, they get just enough.
So it’s, it’s this really cool relationship that you can build. With that machine learning aspect of it and that that was the part that I just thought was so fascinating. It was really, really cool. Yeah, I
was thinking about it as I was watching the video that the terminology around AI really is due for some evolution.
Yeah, in, there’s a big difference between the solar panels in the one part of your video toward the end of the video where it’s the solar panels that are on a mobile rig and they can move around the farm and they learn where gullies and things like that are. There’s a big difference between that. And an interactive AI chat bot.
Yeah. Where you go to it and say, Hey, write me a song about Agri Voltaic, and it writes you a song about a agri voltaic. I wonder if there’s something about like interactive, you know, like, like the interactivity of the ai, the interactivity of the machine learning versus the what seems more like a utility.
And it begins to feel a little bit like we’re talking about the droids from Star Wars. Yeah. Like there’s a big, there’s a big difference between a protocol droid and a GOP droid that, you know, if, for those of our viewers who are familiar with Star Wars, the idea that there’s a droid that walks around and does nothing but offer power.
It’s like a giant mobile battery versus a protocol joint who actually talks, has very sophisticated conversations and is intended to be interactive with humans or other species. Right? But it doesn’t yet feel like we have that kind of terminology or breakdown when we discuss it in these terms where we say, well, this is AI at work and yeah,
it, it’s one of those, you’re creating models like you use this machine learning algorithm.
System to figure out what’s needed to make something work. And then you create that model and then you can put that model on like little computers out in the field, but then just run that model. And so it’s not like it’s sitting there with an actual intelligence figuring things out, right? It’s, it’s just running the model that some supercomputer created.
So it’s like, it’s, that’s kind of what’s going on with this. The solar panel, agri Voltaic system. It’s not like there’s some kind of AI sitting there puzzling out everything every single second. It’s just running a model that they created for that crop and that field and that location, and it’s just running that, that model over and over again.
So it is kinda like that droid that’s going around just trying to power everybody up, right? Not, not have a conversation and, you know, puzzle out things. You know, oh, I’ve got an idea of how this could be even better. Right. It’s, it’s not doing,
And even if it was trying to do that, that’s right now, one of the limitations that the developers of the chat, g p t, the, that whole development, they continue to say it lacks creativity.
It lacks the ability to independently say like, I think I’m gonna write a sonnet today. It’s literally a reflection of our requests. But in your research of this, I asked you already what was the part that you found, you know, sort of most surprising. I’m wondering if there was something in your research that found, you found that was upsetting.
Was there something that you were just like, Ooh, that’s a little leery. Like, did it start singing this on Daisy? Daisy or like
are you talking about did like the machine learning algorithms stuff? The
machine learning aspect of this? Yeah. And then we’ll move on to the actual
farm aspect. Not with this, nothing jumped out for this one?
Okay. Well that’s kind of a relief if we were talking
about open AI chat, G P. Oh, there’s stuff that’s very disturbing there, but that’s a totally different conversation.
So the conversation we’ll focus on in this one will be about the Agri Voltaic as a concept, not the AI I as a concept. There were comments like this from Anile who wrote here in the Netherlands, we have a lot of angry and anxious farmers.
As an aside from. I just wanna say. Mm-hmm. I like that as a statement that the Netherlands is just full of angry and anxious farmers, just like the idea of a lot of very angry and anxious dutchmen running around. So if there is something that can help both farmers and the environment, then I think it’s a good thing.
We also don’t have much nature left and land comes at a premium. So if land can be used for multiple purposes, then it’s a win as. It may also reduce the cost of a solar farm. Since it’s a mutual beneficial cooperation, the people who own the solar farm will benefit from cheap and free land, and the farmers benefit from better crop production.
I’m curious your research as usual, this is a conversation we’re having in the United States. Matt is a citizen of the United States, and when we do mm-hmm. These conversations, and Matt does his research, There is a certain slant toward us information in the US context. I’m curious though, did you see anything that was more global?
And I bring this up because as Anile mentions the, he’s talking about the Netherland context and the Netherland as a country. Mm-hmm. Small, limited farmland. The idea of farmland being at a premium, it’s a little bit different than here in the US where we have swaths of land, where we got lots of nothing.
And yeah, if the United States at, at the federal level decided, Hey, we’re gonna take a hundred square miles and turn it into the an, an enormous solar farm, we would have plenty of places where that could be done without it. Being an issue as far as land use. Right. But other contexts aren’t quite as cut and dried as that.
So I was wondering, did you see anything where there are regions or countries that are specifically looking at things, multi-use goals like this? In this way you see a lot
more of this in, in Europe. In our previous video that we put together on this topic, it was one of my patrons, uh, Robert Heder helped to basically help write that script.
He’s, he’s from the Netherlands, so he was, he brought in a lot of European research and context to it. Mm-hmm. And so for this one I was, we were kind of taking the different tack where we were kind of pulling in a little more US centric view to this, because this is now starting to get kind of tested here where Europe is kind of ahead of us as far of the, as far as this is going.
That’s why this video was a little more slanted. The other direction I was trying to. Balance things out in between the two videos, but you see this far more over in Europe. They’re like testing this everywhere. You know, Germany, Italy. Netherlands. Mm-hmm. You name it, somebody’s testing it over there.
There’s, it’s, it’s incredible because it’s, they have, like you said, they’re a little more landlocked. They’re a little more densely packed, and there’s not a lot of room to do what we can do over here in the United States. So the dual use is incredibly appealing over there. And this ties right back into, like I say this every single time.
I’m a broken record, you know, the right tool for the right job. Yeah. If this is a great solution, it doesn’t mean it’s a great solution every. So it’s like, it’s gonna be very regional and I think this is gonna be something that’s gonna probably take off more in Europe than it will here in the us. And one of the things I thought was really fascinating was that when we pulling the data together of like the news research that’s like, well how much does this cost compared to, you know, just building a solar farm or just, you know, what is the difference in cost?
And we were finding the HA like based on the US research, it was dramatically, actually it was Germany and US, it was dramatically more expensive. Than just doing a straight up solar farm. But in previous research from the last video, we saw almost the opposite. It was only minorly more expensive than. All other alternatives.
Mm-hmm. So it’s like, clearly, I don’t know what it is about how one group built something out to price out the model for what it’s gonna cost versus another one. But it’s like, it, it’s so variable based on where you are in the world, what your needs are, how you’re building it out. It’s one of those, there’s so much more research needed on this because we weren’t finding good.
Universal. Yeah. Consistency. Yeah. It
was, it really based on, it was really based on where you were and it was, it was kind of eye opening, how different the, the results were based on where you were.
I think that there are also aspects like this. Brought up by Tony Vera in his comment around not only the context, but the kind of technology that you might be trying to apply.
Tony writes, vertical Mount solar fencing is the most promising of the Agri Voltaic setups out there because it has the same cost as regular ground mount PV doesn’t use more materials and doesn’t cost much. Doesn’t cast much shade on the crops growing. Moreover, it peaks in power generation in the morning and late afternoon, which is more closely in line with the demand, and it still provides great wind production to crop fields.
This is agri Voltaic that are ready to scale now. So I wonder, did you run across that as a model as well in your research? The, the idea of fencing being the solar.
Oh, yeah, that’s, that’s definitely one that we came across. We talked about that one in the first video. Mm-hmm. We didn’t really talk about much in this one, but that is something we came across a lot.
But again, it depends on who you’re talking to because he’s not wrong in everything. He’s, he’s citing, but we found other research and other papers that we’re saying the complete opposite. So it’s like, It, it, it really depends. But the one thing that was kind of the universal across all of it was if there is an incredible amount of interest, whether it’s vertical, horizontal, these tracking systems, there is so much potential there that people are having success at testing these different systems out and making it, making the math work of like, no, this does make sense.
You’re seeing more and more of that, but. We need more research. We need more places to do this because we need to start to find the patterns that are actually working and which one is the best depending on what region you’re in. So that vertical system might be better in one part of the world versus a horizontal system for a different kind of crop and a different part of the world.
So it’s like this is one of those, you can’t just say blanket, vertical or nothing, right? You can’t say that. So that’s kind of bringing back the right tool, right Implementation for the right crop in the right region. It’s complicated.
To return to the AI aspect of this, there was a comment from Flore Azo that caught my eye.
She wrote, I definitely can see AI being used in the initial setup of agricultural spots that also use renewable energy production, give it weather and geographic data, et cetera, and an AI can place ideal spots that have the best output and other. AI models can absolutely monitor and create a very efficient farm design and nutrient loads.
Considering the solar harvesting tech is always improving, I can easily see not only stationary power food harvest plots, but also mobile farms to provide fresh produce to rural or impoverished communities. I could even imagine mobile systems that use aquaponics with water harvested from the air. AI’s ability to create highly efficient setups is the most exciting aspect.
I think technology that would take years to engineer will be done in mere hours. Fresh produce is so important for health. I can only hope this technology can improve health globally. I wanted to share that just because it’s nice to see such optimism and idealism and a vision in this era of. Doom and gloom.
Flora, I hope that your vision of what could be possible with this can come true. I do like the idea, especially of what if there were the ability to create on a small scale. What if a, a small farm was able to create somewhat mobile farming that would be able to perhaps move from region to region on the landscape where one area of a farm might get more sun than another, and you might end up.
AI that actually moves pallets from place to place in order to spread a little bit of the sunshine across those pallets. The little things like that that I think are fascinating to think about and really scratch the sci-fi itch that I often feel when we have these conversations of when are we gonna get those flying cars?
And Matt, I’m gonna ask you that question right now, when are we gonna get flying cars?
I don’t, I hopefully never. So listeners, before you go on, I just wanna, I wanna just add to that I agree with you completely and we’re heading to that future, I think sooner than people think automation on farming is happening.
We already have drones that can target. Treatments of like, you know, spraying pesticides instead of blanketing your field, it can go and find, oh, the bug is right there, and it’ll like literally spray that one little occasion. It’s like we’re, we’re getting to this place where automation is gonna make farming more and more efficient and be living, basically living with the land instead of fighting with the land.
It’s, it’s, we’re heading towards that future.
I recently was looking at. I mean, this is gonna make me sound very bougie. I was looking at a bottle of wine recently and on the back it was described as being from Italy in the most vertically inclined vineyard in Italy. And the harvesting has to be done via helicopter.
Whoa. And so there are apparently teams of people who are harvesting and then those harvested grapes have to be airlifted by helicopter back to the vineyard. And I couldn’t help but think almost immediately. Is there a video in Matt’s future where this becomes a drone? Production where you have not drone robots, just drone robots bringing bushels of, of grapes back to the vineyard instead of helicopter.
And, you know, the sci-fi side of me suddenly saw all sorts of little climbing robots going down the side of a cliff to reach the, the, the vines. And I couldn’t help but have, Image in my head. So here’s to hoping and like toasting with a little glass of that wine. Here’s to hoping that that future where as Matt put it, working with the land instead of fighting, it can be taking place.
So listeners, what do you think? Our flying cars right around the corner? No, I, what do you think about all these ideas around incorporation of machine learning, ai. Solar panels, multi-use farms. Have you seen any of this in your neighborhood? Or do you think that you live in a place where something like this would make perfect sense?
Let us know in the comments if you’d like to support the show. Don’t forget, you can go leave a review on YouTube, apple, Google, Spotify, wherever it was you found this podcast. Go back there, leave a review. Don’t forget to subscribe and share us with your friends, and if you’d like to more directly support.
You can click the join button on YouTube or you can go to still tbd fm. Click the Become a Supporter Button. It allows you to throw some coins at our heads. The welts Heal and the podcast gets made. Both of those options are great ways to support us. Thank you so much, everybody for listening or watching, and we’ll talk to you next time.