Ag Geek Speak
GK Technology Inc Team Members, Jodi Boe and Sarah Lovas talk about precision agriculture, agriculture mapping, agronomy and drainage.
Ag Geek Speak
23. Talking AI with The Man Behind the Curtain: Travis Yeik pt. 2
In this episode, we explore the cutting-edge innovations shaping precision agriculture. We dive deeper into ADMS with software developer Travis Yeik, talk turning digital signals into hydraulic actions, and dive deep into artificial intelligence (AI).
And now it's time for a Geek Speak with GK Technologies, Sarah and Jody friends and I can't wait to get in the fields again.
Speaker 2:No, I can't wait to get in the fields again.
Speaker 3:Welcome back to the second part of our conversation with the man behind the curtain, mr Travis Yike out of Wyoming. We are so happy to have him back here on the podcast to talk about and geek out about so many different things. I'm not even going to list him here because, as you'll see in the episode, we're going to talk about a lot of things, but it is such a pleasure to have Travis here.
Speaker 1:I think the last time where we left off we left off with Travis had just gotten hired by Darren off with Travis, had just gotten hired by Darren, didn't even have to turn in a resume to get hired, which is well shouldn't be surprising if you listen to that whole episode. He's done a lot of stuff with remote sensing, so let's talk about this. What exactly do you do at GK Technology right now?
Speaker 2:That is a good question, and sometimes I ask myself that every day, you and me both.
Speaker 2:And sometimes I ask myself that every day. Yeah, I've been involved in quite a few of the projects that we have at GK, obviously the ADMS software, which is our main desktop software and it does all of our precision ag and GIS coding. I'm also involved in our SD drain and SD ditch and SD tile and you would probably hear on a previous podcast Paul Fuller talked about some of our SD stuff and how we're doing there, and so I went from one end of the spectrum working with valley irrigation in places that didn't have enough water to places that had too much water, and this is a totally foreign concept to me from a guy from Wyoming that you know we were struggling for, you know eight inches of rain a year.
Speaker 3:Let's remember that I had asked the question of why Travis had focused on doing zoning for irrigated fields in Wyoming and his answer was we grow crops that are under irrigation. That's where they're grown is under irrigation.
Speaker 1:So I agree, I come from.
Speaker 3:Western North Dakota and the concept of tile is just like. Why do we want to get rid of the water?
Speaker 1:But there are obvious reasons. I can't imagine a farm that didn't own a scraper.
Speaker 3:But anyway. So again, I mean you've done a ton for SD drain. I mean can you tell us about, like where, when you started on working on SD Drain, where it was, and then what came to, or like where it is now and your involvement with that?
Speaker 2:Yeah, I had to learn everything. We started with Ditch, and so I had to learn everything about why as you said, why we have to get the water off the field.
Speaker 2:Yeah, and which was neat, yeah, which was neat. The big part of that is learning GPS and doing RTK and understanding all the code behind GPS, which is a whole new concept, and to understand and relate that to how we implement that with the actual hardware was a real step forward in how precision agriculture is quite useful in that way is to get it into hands, into the technology that people can use it. That way.
Speaker 3:At the end of the day, once we get maps into that program. That program is then telling either scrapers or tile plows where exactly they need to be in the field to get water to flow. Can you talk more about what you have to know and how you get a computer program to talk to an implement?
Speaker 2:Yeah, maybe I'll go back when I actually first started school. I went into civil engineering and I did that, I think, for a week, and I was like I don't want to sit at a desk and do math all day. That is all I do now. Sorry about that.
Speaker 2:Sorry about that. Yeah, so it's a lot of trigonometry and that's one of the math subjects I guess that I was good at and Kim's back and haunts me to this day but is doing that. And we work with a great company. We work with Rust Sales and guys there. They actually make the that controls the hydraulics behind everything, and I talked to that DAC to tell it how to where to send the implement or send the scraper blades up and down.
Speaker 1:So, travis, what is a DAC? I mean, we talk about this all the time with our customers that use, you know, go to Rust sales and get their SD drain systems. What exactly is that DAC that goes in the DAC, in the cab?
Speaker 2:Yeah, so DAC stands for digital to analog control, and it's what changes from what I send it as code in digital form, from what I send it as code in digital form, and it changes that to the voltages which then converts it into how much the hydraulics need to apply the pressure to make the scraper blade go up and down.
Speaker 1:So essentially, it really is the place where you're sending your computer code and that's how we're ultimately voltage first, but then it's the hydraulics.
Speaker 2:It's the little box that makes the magic work. Yeah, there it is, that's right.
Speaker 1:It's the magic box, the man behind the curtain with the magic box. That's pretty neat. We've had so much success as a company with SD drain and SD tile, and part of it is just because the system is programmed so well and it works so well and the farmers love it. The farmers love getting the maps in the background that work so well with it as well. In addition, we have a great system for making surface drainage maps and a great system for making drain tile maps. Thank you, travis. I just think it's important for everybody to know that it really is quite revolutionary, and Rust Sales does have incredible customer support. If you need help figuring out how those systems work in the tractor cab, they just have incredible customer support. We've been so lucky to partner with them. Now, did you do most of the programming on SD drain and SD tile, travis?
Speaker 2:Yes, yep, so some of the map controls we were able to reuse those from ADMS, but then the rest of it is from scratch that we put together and it's been in. How long have we had SD here? Since 2014, I think.
Speaker 1:It was a long time before I came to work here.
Speaker 2:So it's been out there now for 10 years and I think that's one of the things that I was concerned, I guess, or trying to get across with Precision Ag is that I think there's a few things in Precision Ag that we all struggle with, and that's the cost of it. Sometimes it can be costly to implement and especially if you have like a small acreages or even, you know, just getting the cost down so that everyday users can use it. And then the second major problem is the technology of it. Is that there needs to be, it needs to be simplified so that everyday people can use it in a way that's In ADMS. Don't get me wrong. Adms is a wonderful product and it takes a lot of skill to use. There's not an easy button at all in ADMS, but in SD a lot of farmers they need just a two or three button clicks to get to everything that they need to do and I think that's important with the future of precision ag is to make that available to people.
Speaker 3:I'll just say thank you so much for making it a simple two, three button click and SD drain and SD ditch. Because, absolutely, as a practitioner of precision agriculture, as a farmer, right, there are so many things that you need to know as a farmer that it is very difficult to sit down and say, hey, I'm going to learn how to do GIS and understand you know how all these maps fit together, and then go out and do this. It's just something that when, in the grand scheme of all the decisions that have to be made as a farmer, when it's like, okay, I've got to take a couple weeks off to learn how to this completely new concept, it's that you're right, that cost, it's a big barrier. So the simplicity of it is huge, and that probably took a lot of time on your end and thinking about how do I make this process that's agile enough, that can be done, and two or three button clicks. So thank you for taking the time to think about that while you were developing and putting together those systems.
Speaker 2:And it's probably more of the reason that or just that it didn't come from a drainage background is that, you know, I had to understand it and come back through it, so I wasn't able to look at what was out there already. And it's and I think maybe that's another reason of the success for it in trying to dumb it down is that I was understanding how the drainage process works at the same time I was developing it.
Speaker 3:That's really interesting, right? Because a person would think like, oh, you'd have to really know how these other programs work in order to make our own. But your point is is that because you you didn't have like a quote-unquote bias of how the systems already worked, so you could, you know, make it your own with the mind of simplicity as or like, with the point of simplicity being your end goal?
Speaker 1:yeah, that's huge. So question for you um, in the ADMS software that we have, we have the drainage window and we've got watershed modeling and we've got tile design. Did you program those or did Darren program those?
Speaker 2:Darren, yes, we originally had a drainage and a tile and Darren had done a lot of that in watershed modeling and Darren had done a lot of that in watershed modeling and when we moved over to the SD product I redid a lot of that and we've had two different now drainage models in ADMS. You know, that interesting kind of looking at the research articles and putting research articles in the code was difficult but it's fun at the same time.
Speaker 3:This is kind of a weird question. Do you have a favorite research article that you reference a lot or think is written very simply?
Speaker 2:No, not a favorite. There's so many research articles and they cover so many topics, as does ADMS covers so many topics, and so it's hard to pinpoint one down. I guess, as I go into some of the new stuff that I'm learning with artificial intelligence and whatnot, that a lot of those articles I could not find hardly a single article on using AI in precision agriculture. A lot of it is looking at, maybe like the USDA on saying hey, this is soybeans and this is corn, and doing a broad spectrum of average and using AI for that. But as for more of the precision ag stuff, I had to scour. I guess all the money for AI is going into technology and entertainments and into the medical field.
Speaker 1:Interesting. You're working on a number of different projects right now in the background things that are going to be coming out into the future for GK Technology. Do you want to share with us without sharing too much? What kind of fun new toys are we going to get to play with Travis?
Speaker 2:Well, so, as I was saying, part of Precision Ag and the thing that I think about as a developer in the background is understanding how to get costs out lower, I guess, for people and to make the technology widely available available, and so I so right now we're coming out with, with a phone app like as, where, um you know, farmers can have precision ag gis um in in their palm of their hand with their phones and hopefully use it that way, and in this case it'll it'll be used a lot for for surveying and for record keeping and for sharing data back and forth between our main program, ADMS, and I think that'll be super helpful to our customers.
Speaker 1:And soil sampling.
Speaker 2:Absolutely soil sampling.
Speaker 1:We've got to have it in the soil sample rig. It's going to be a lot of fun, and so if you're looking for that way that you can seamlessly data exchange between this app and your ADMS software, it's definitely something to check out. I think it's going to be a lot of fun. What other fun projects are you working on right now?
Speaker 2:Yeah, sd Drain is always expanding. Right now we are testing a shaping product which will be coming out, hopefully, either end of this year or next year, after we do a bunch of testing on that.
Speaker 1:And that's land shaping right.
Speaker 2:That is land shaping, so taking an entire field and getting rid of your potholes and making it so that water runs downhill Yep.
Speaker 3:One of the questions I get when I go to my fiance's home state is about terracing. Has there been any consideration to land shaping to do like terracing?
Speaker 2:Yeah, absolutely, and I think that's what our shaping program will be used is for some of these things like terracing or creating. You know if you're doing rice paddies or we are just leveling areas off. Yeah, yeah.
Speaker 1:Because really the focus of SD drain and into itself, when we think about drainage, surface drainage is just kind of to get rid of that pothole that's out there, right, and get the ditch clean. But whereas with land shaping you are, you are literally able to create a dam. If you need a dam in a place or get the water to flow, you can literally shape the land.
Speaker 2:Or a golf course.
Speaker 1:Yeah, there we go. I like that. That's pretty fun, so that's going to be something that's going to be new. Coming out is land shaping.
Speaker 2:You've got that app, so you're working on all kinds of fun projects then Anything else use, whether it's copying a politician's voice or creating inappropriate images or whatever it is, and there is some taboo around it, but I think there is a lot of potential in the future that we can use artificial intelligence in precision ag to help us make better decision-making processes.
Speaker 2:Part of the problem with precision ag, I guess, is that, yeah, there are a lot of variables, right. We have soil data and the quality and the types of soils, and we have you know how that relates to fertility of the plants and the soils themselves, whether it's nutrients or water management, and then we got environmental factors such as rain or hail, or pests or weeds even, and so agriculture has a lot of variables and and there's like disciplines, right, that like focus in on each one of these variables and to expect a farmer or a consultant or anybody rather to to know any and all of this stuff is you've got to have four or five different guys in the field to do that, and then to wrap that all up into a precision agriculture software package that people can utilize and understand is difficult to do as well. Intelligence may be a bridge between all of these different variables, to understand and make sense of some of these variables that are hard to model.
Speaker 3:So GK sent you back to school, right?
Speaker 2:Actually, when I was in undergrad there AI is not a new thing they're called Markovian decision models and these were created back in like the 1960s or so, and so AI has been around for a long time. Even when I was in undergrad there, they had computer programs for remote sensing that could model and do classification coverage using AI. It's only within like the last I don't know even five or eight years that it's become popular, and I think some of the, the, the coding Google has some code out there that's open source, that that the everyday public could use and and and to understand and create their own AI models that way, and and then now we have ChatGPT and Gemini and all these other things. That AI is kind of blown out of proportion now, but so it's been out there for a while and we're just now starting to understand and utilize it better in our everyday practices as AI develops more and people get more interested in it.
Speaker 3:So what exactly is artificial intelligence, ai, what exactly this is and what does it mean when we say something is AI?
Speaker 2:Yeah, that's a good question and it's hard to. One of the things I learned in one of the internships I had with the Department of Energy was to create a grandma speech right. This is the speech that you tell your grandma when she asks you, hey, what do you do? And you have to break it down into terms that she can understand it right.
Speaker 3:I love. I'm sorry, I love. It's like you learned about the grandma speech working for the DOE.
Speaker 1:I just like that a lot.
Speaker 3:I get a kick out of that, anyways.
Speaker 2:And so I guess here's the way I would, I would define ai. Let's take a giant plinko game and you have a different maybe you got a little round ball at the top and you drop it down and it goes through all the different pegs, right and uh. And then at the bottom maybe is we have, uh, you, our bins at the bottom, and these are kind of like our decision making process, right. And so as we make a decision, as we sit here as we talk or as we think about a solution or a problem to something, or problem or solution to something I said that backwards that we have a Plinko that goes through and it goes through all of our different nerves in our head and it bounces off of each of the little pegs and finally it comes down to the bins at the bottom where we make a decision.
Speaker 2:And so as we AI is I'm going to make the analogy here that if we change the maybe the diameter of each of those little pegs in that Plinko game and we roll it down, that it will we can make it so that that little ball can roll in the right bin that we want it to.
Speaker 2:And AI is learning that. Hey, if I change this way of this peg differently than this peg, then we can make it roll down into the right bin and we have different inputs. So we may not even have a circle, we may have a square or an octagon or some geoid that we roll down the Plinko and after a while we could have a million different little pegs that it can, which is all the variables. So some of those variables I talked about, such as soil or water, nutrients or weeds, and if each one of those models was talked about, such as soil or water, nutrients or weeds, and if each one of those models was a different little pig that it could roll differently based on how that geoid was shaped, and roll it out into the right bin, does that make sense?
Speaker 3:I think the way I'm understanding this is that you're giving the right amount of weight to different explanatory variables to account for more of the variability that affects an outcome, and you're accounting for more of it. I think AI is.
Speaker 1:I think your description made sense and I think AI is going to have a fit in agriculture and I think AI is going to have a fit in agriculture. I know that there's a lot of conversations around different ideas that we can do with AI and I think some of the decisions that are some of the places where we may need help modeling. You know we think about diseases and weeds and insects and you know I certainly hope into the future that we can find a way to remote sense those problems. So often in today's agriculture, at least around here, we deal with wet field conditions quite frequently and I can remember doing crop scouting and there was no way I was ever going to get that four-wheeler to go across the field because it was just a slop hole, it was so wet. So if you can remote sense those things with a drone, with a plane, with a satellite, with something, so that you don't actually have to physically drive across it first of all, you're going to be saving that agronomist a lot of time because hopefully you get some remote sensing data back that you can interpret and if you can have AI helping you interpret that, that can make those decisions that much faster.
Speaker 1:But I know there's a lot of work that has to be done before we have those reliable models out there, and I do think it is important to remember in agriculture that we have so many variables because it's a natural system. Mother nature always wins, and one of my favorite sayings in agriculture and life sciences and then an enzyme happened. We always have to remember that. You know, we've got these chemical equations that don't balance because there's enzymes. It's a natural system and so hopefully AI will help us shore up some of these things that, quite frankly, algorithms haven't been able to reliably do, and it's going to be fun to see where it goes into the future. I think and, travis, I think you did get some extra training in AI. Just a quick question in agriculture where do you think are some watchouts and things that we should be careful with AI as we're going forward and as we're implementing AI into precision agriculture?
Speaker 2:Yeah, that's a good question and I think, you know, I think AI is going to be lagging behind in agriculture, you know, as compared to some of the other more profitable things such as such as the medical or whatnot. But yeah, and as I, as I kind of mentioned, there are a lot, of, a lot of variables. We're we're modeling an entire ecosystem with farming, and to do that and to make it so that it's it's robust and intelligent and works in every scenario is challenging, I think, and I don't know if it'll ever get there.
Speaker 1:You know, we hope so and yeah, so, as we sit here and we're having this conversation, when you think about your career and the things that you've done and you look forward to the future and thinking about where things could go, what, what are some future things that that you might think would be super fun to work on, or what do you think the future of precision agriculture looks like?
Speaker 2:boys that loaded question, I know, right, there's no answer.
Speaker 3:You can't get it wrong.
Speaker 2:That's right. Yes, I love remote sensing and I love the research behind it and understanding how the processes work, and I think that's you know. As for me as a self, as a developer, that's what I hope. As a developer, that's what I hope. As for precision agriculture in general, I kind of alluded to it before but, yeah, we have these problems that are preventing precision ag from being where it can be in the future, and some of those are the cost and the technology behind it, as well as the support. And that's what I love about the gk technology is that we do have the support, that that goes behind that technology, and that's super important to get people involved and to get people to understand how, how it works and where we're going with that. And so, in the future, for for precision ag, that's what I want to focus on is making kind of those easy buttons, just like an SD drain, so that everyday people can use it and understand the value of precision ag and how we get there agronomy and a background in farming.
Speaker 1:I just have to say that I think it's really exciting to hear a developer, you know, with the ambitions of making sure that we can get technology into the hands of people and that cost shouldn't be a barrier and that it should be simple enough for everybody to use. And hopefully, if Jodi and I do our job right, and everybody else will make sure that we're there to support it, which is pretty easy to do when you've got good products. So anyway, that's very fun. This has been a really fun conversation and, again, I don't think there's a lot of people out there that really know about Travis Yike, because he's kind of a guy that is behind the curtain, but he really is a lot of brain power that goes into a lot of the products that we get the opportunity to work with. So thank you so much for being here with us, travis. It was a very fun conversation. Who knows, maybe we'll have to have you back on again sometime. Thanks for being here, thank you.
Speaker 2:Travisvis, I would love it. Thank you so much again.
Speaker 3:Thank you guys, all so much for listening this week. And remember, with gk technology we have a map and an app for that.