This Artificial Intelligence is Disrupting Cannabis Plant Care

Max Unfried Deepgreen

What if cannabis cultivators could use artificial intelligence (AI) to help them automatically identify a problem with their plants before the problem is even visible? Our guest today is Max Unfried chief AI officer for Deepgreen.ai. Max walks us through how AI is being leveraged in the cannabis space for the maximum benefit of growers and business owners.

Learn more:
https://deepgreen.ai

Important:
Get your free cheat sheet “The Five Trends Disrupting The Cannabis Industry” here https://www.cannainsider.com/trends

Read Full Transcript

Matthew: What if cannabis cultivators could use artificial intelligence to help them automatically identify a problem with their plants before the problem even starts? Max Unfried, Chief AI Officer at Deep Green is here to tell us how sophisticated cultivators can leverage AI to help their grow. Max, welcome to CannaInsider.

Max: Thanks for having me, Matt.

Matthew: Give us a sense of geography. Where in the world are you today?

Max: Well, at the moment we're, like, based out of beautiful Boulder in Colorado. So, next to the Rocky Mountains.

Matthew: Okay. And what is Deep Green at a high level?

Max: At a high level, Deep Green is an AI platform that uses computer vision to give plant analytics. And, specifically, we do diagnostics of diseases and pests and detect those on crops and we also do yield estimation.

Matthew: Okay. Can you share a little bit about your background and journey and what were you doing before Deep Green and where you're from and how you got to this point?

Max: So, by trade, I'm a scientist. I studied physics and branched into artificial intelligence at university. After that I did, like, some guest research in Taiwan about, like, detecting facial expressions in humans with the help of computer vision. And then, after that I landed my first job out of university in Vietnam where I worked for a Swiss FinTech start up. And I was building machine learning systems and algorithms to analyze tweets and news, and try to extract insights from them about what the crowd is thinking about certain stocks and help investors in a way make, like, a choice based on those tweets. And that's actually where I met my co-founder and our CTO, Maxine.

Matthew: Okay. Okay. And where are you from originally, Max?

Max: I grew up in a small, small town in the German countryside close to Stuttgart.

Matthew: Okay. That's nice. Cool. And let's dig in here. Can you tell us how most cannabis cultivators are looking for problems with their plants now, and what you're doing differently with Deep Green to help them get a sense of what the problem is and then what's possible with Deep Green?

Max: Of course, I'd love to. So, I mean, everybody who has been, like, in cultivation so far, they see all those plants. Like, thousands of them, right? And, currently, like, cannabis cultivators walk through the rows and inspect each of those plants manually, like two or three times a week, and check if there's, like, any kind of disease or pest or, like, nutrient deficiency. And just imagine how much work it is, like, just checking every plant and every leaf.

Especially, now there is, like, double stacks become more popular that you actually have, like, a stack of cannabis plants above each other. So, you have to move your ladder all the time if you wanna check it properly. And, obviously, cannabis cultivators, they have their hands full, they have, like, other things to do. They cannot spend all the time working through those rows. And that's where Deep Green is coming in.

The problem is with the diseases in plants, they usually start very small and, like, are often hard to capture for the human eye. But they can grow rapidly in a grow room, right? And it can be devastating. And what we do at Deep Green, we help cultivators to monitor their plants more efficiently. What we, in a way, provide is imagine that you have a master grower with super-human vision that has an eye on your plants 24/7. And the way we do that is we put, like, cameras on the ceiling of a grow or we attach them to LED lights and those cameras, simple [inaudible 00:04:15] cameras taking a picture an hour.

This picture is then sent to the cloud and it's analyzed by our artificial intelligence. And if we detect any issue we will send a notification to the grower with the exact information about the room number, the plant, and the issue we detected. So, instead of walking the entire room, the grower can go to this location and check if everything is okay or what is wrong. And if anything is wrong he can just do the next steps to prevent damage.

Matthew: Okay. So, is the camera in the ceiling or on the light, does it, like, auto-focus to what it needs to see and then it just snaps a picture? Is that what happens?

Max: Exactly, it just snaps a picture.

Matthew: Okay. It sends it to the cloud and that's when your algorithm kicks in and starts looking at the pixels and going deep into the pixels, and saying, "Okay. Is there something that matches a pattern I have in my database?" Is that pretty much what's happening?

Max: Exactly.

Matthew: Okay. Okay. That makes sense. What's writing an algorithm like? I mean, this is just a mathematical...I mean, do you get to...is there, like, an open source library of algorithms that gives you a starting point where you can kind of take ingredients and make your own algorithm?

Max: So, I mean, the technology we use is, like, called Deep Learning. And it's, in a way, a technology that brought out the most breakthroughs in AI in the recent years. And, obviously, we're like any other research company, right? I mean, Google, Facebook, they have, like, hundreds of AI scientists that already do a lot of research and provide some of those...the idea of those algorithms. By the end of the day, we code them ourselves. We, obviously, use inspiration of what comes out of the research. But at the end of the day, it's like we're, like, sitting in front of a computer and, like, putting the mathematical equations into code that the computer can understand.

Matthew: Did you see that news headline where Amazon's technology, in kind of a publicity stunt, confused people in Congress for criminals? They wanted to show how that could be misused.

Max: I actually did not see that, but it doesn't seem too far away that people in Congress could be criminals, right?

Matthew: Oh, gosh. You got to host a show, Max. That's great. Well, I'll include that in the show notes so people can see that it was kind of a publicity stunt though, because the...it was kind of big in the news for a couple days. But then Amazon said, "Hey, they didn't have the reliability set to 99% and they didn't do all these things right when they used our facial recognition and so forth." But I could see where there's a potential for abuse at least on the human side in terms of, you know, using it to control people. But, in this case, it seems like it's really super useful. Can you tell us what some of the most common problems in the grow room you find with plants?

Max: So, the largest issues that we find when we speak to cultivators is, like, they have usually powdery mildew, which is like a fungal disease. And it attracts spider mites, which are like some very small insects, and then obviously nutrient deficiencies, because I mean, it's just super hard to get the perfect nutrient combination for every single plant or strain. But the biggest pain point is really powdery mildew, because it can spread rapidly through the air and it's a fungus and fungus like a humid environment. So, they're having probably a really good time in those cannabis cultivations.

Matthew: Yeah. Okay. Right. Right. Right. So, how long does it take after the picture is taken? You take a picture every hour, it goes into the cloud, and how long does it take before you get an e-mail or a text message or whatever the grower that you've detected an issue?

Max: Usually, like, our algorithms they can, like...if we take an image, like, it's processed within four seconds. And then we can directly send it back.

Matthew: Okay. So, it's happening almost instantaneously it sounds like.

Max: Yes.

Matthew: Okay. And we talked a little bit about how, you know, the human eye is not necessarily trained to see this until it's a problem that's, you know, manifest large enough for us to say, "Holy crap, there's powdery mildew or spider mites." But how much earlier can a camera recognize this with the algorithm that you have?

Max: So, at the moment we're that far that we can, like, recognize the earliest signs that are visible on a leaf. So, still before a big outbreak is on the way and you can still control the damage and prevent it. The good thing, obviously, about computers is if they look at an image they are good at seeing things human eyes cannot see on photos. And we're currently working really on detecting what is invisible to the human eye.

Matthew: Wow. There are so many applications here. It's unbelievable when I start to open my mind to this. And how about coffee, have you ever tried this with coffee?

Max: Well, it's actually quite funny. I mean, when I spent, like, two years in Vietnam, and Vietnam is like the second largest producer of coffee in the world behind Brazil, so I like running and I've been sometimes running, like, through coffee plantations and I was obviously looking at those plants, like, just out of curiosity for agriculture, because in Europe, usually, we don't have coffee plantations. And sometimes I saw something odd, they're like weird yellow spots and, like, some things that could be powdery mildew outside on there.

But I have no idea how big those coffee plantations are affected by that. I mean, it's like an outdoor crop, right? So, they obviously have, like, a different immune system than, like, the plants indoor. But as long as there's a medium to capture images outside, we can use them and analyze them. And honestly, I expect that coffee farmers of the future will use drones to fly over their plantations. Because, I mean, they're massive, right? I mean, you just cannot walk them. You can run several marathons through them and probably still not be done.

Matthew: What type of plants do you think lend themselves to working with Deep Green the most? Do you think cannabis, any others, hemp? I mean, hemp's outside usually, so...I mean, have you thought about hooking this up to a drone or is that not what you're working on right now?

Max: We have thought about it, and it's on our roadmap, but it's probably, like, two years away. Just because, I mean, farmers not adopting drones that quickly yet. But the way we think about it is like that cannabis is, in a way, the black swan of agriculture, right? You have, like, the young people, you have young farmers compared to general ag and they have a quite large affinity to technology. And you also have, like, a mode from big agriculture, right?

Like, John Deer is not gonna move any time soon into cannabis given its legal status. And that actually allows for great innovation that is driven by, like, smaller startups, and use innovation to transfer to general agriculture. So, we had in the mind from the beginning that we develop this technology in the cannabis industry, given that, like, it's a high-value crop and [inaudible 00:11:53], but then move to general agriculture.

And, actually, we moved quicker to general ag than we would have expected. At the moment, we still do mainly cannabis, but we started out our trials with center pivot irrigation systems that are like those large circles that run over cornfields or alfalfa fields. And we mounted cameras on those, and currently we're detecting on corn and alfalfa fields, we detect weeds. And you brought up hemp and the company we're actually working with, with the center pivot irrigation systems, they just sold their first pivots to hemp farmers. So, that is coming. So, we're looking into putting cameras on pivots that run over hemp fields to help analyze what's going on there.

Matthew: So, you say you also recognize weeds for the general ag farmers?

Max: Yes.

Matthew: And then, once they see that there's a weed there, is there an automated way to remove the weed or are they just like, "Oh, I know there's a weed and I have to go remove it by hand," or what happens?

Max: Obviously, like, removing of it by hand in the alfalfa or cornfield is not scalable, right?

Matthew: Right. That's why it's like. what do I do with that information? I've got a weed, but as long as there's not too many weeds I guess they just look at the picture.

Max: Exactly. That's fine. Obviously, if we have, like, large patches of weed then they probably consider spraying. And that's still good, because they can just spray an area where they know the weeds are instead of, like, laying a carpet of pesticides over the entire field. In a way that saves them money, but it also helps the environment to do, like, in a way, precision agriculture and, like, precision spraying instead of going brute force and, like, just putting chemicals everywhere.

Matthew: Gosh. I could see a future very soon where, you know, there's little automated robots just going through and picking and doing things like this, just general maintenance. That doesn't seem crazy to me at all.

Max: No. I mean, it's getting there. I mean, there are already robots out there that can pick some fruits like strawberries or like apples or just go through and look at the plants and see what's going on, on the fields. I mean, you have, like, robots in cannabis that try to help with the trimming. There's, like, a lot of cool technology coming up in the next few years.

Matthew: Yeah. So, if there's a grower or a business owner that's listening right now, and they're like, "How do I get started with this? I wanna integrate something like this and get notified, you know, before a disease really manifests or a pest causes a problem?" But they're like, "Is this cutting edge, is this bleeding edge? Like, how much learning curve is there?" Can you tell us what would it be like for a business owner or a grower trying to onboard this technology?

Max: So, the first thing would probably be to reach out to us. But, in general, it's like very, very few work for the business owner. Usually, we just come, we put in those cameras, we install them, we need, like, a few sockets where we can plug it in, and then we start automatically taking those images. And the business owner, or like the grower, will be notified of anything is wrong. So, it's, like, not a big barrier to entry. Everybody has a ceiling or, like, usually lights where we can monitor. And yeah, then they're good to go to receive the notifications. And the only thing they have to do is to open the e-mails or the text message to see if anything is wrong.

Matthew: Okay. It'd be really cool to have a way to have this where a security camera tells you if there's an unidentified person in the grow. That would be kind of interesting too, but I guess most grows have so much security now just to get in that I don't know if that's necessary. But it would be nice to know. Do you think that would be pretty easy to make or does that sound difficult?

Max: Compared to what we are trying to do, detecting a person in the room is, like, very easy.

Matthew: Okay. Okay. What do you think? Can you tell us something about someone's mental state, you think, with a Deep Learning algorithm where you could say, "This one matches a profile of being high and wanting to eat Ben & Jerry's ice cream?"

Max: I mean, can you as a human?

Matthew: Not as a human, but, like, is it...I think as a human, sometimes I could, but I mean, is that something you think you could program into an algorithm or does it sound too crazy?

Max: I mean, as a human, you can certainly say if somebody is high, right? If they wanna eat Ben & Jerry's ice cream, that's probably, like, a harder part. But, I mean, given on the facial features you can certainly detect, like, emotions, what they're feeling, what kind of state they're in. And then there's, like, cool stuff coming up with, like, EGs. Like, sensors that measure your brain activity or, like, your brain waves. And, obviously, if you hook together imaging plus brain wave reading you might actually be able to figure out that he really wants to have Ben & Jerry's, because every time he wants Ben & Jerry's, or she, there's some certain part of the brain that's gonna be activated. So, let's see what the future holds.

Matthew: Yeah. Speaking of the future, I mean, I don't know, it seems like in the last five or seven years, technology has just gone on a tear where there's, like, an accelerating acceleration. Like, it used to be where you could kind of see, "Oh, you know, here is another development, but it makes sense that it was a while since the last one." Now, it just seems like it's just absolutely, you know, stage-five "Star Trek" and there's just craziness happening every day. What's your kind of plan to keep iterating this so three to five years from now it'll still be relevant, and still be, you know, top-of-mind for a grower?

Max: So, the way we see this, like adopting our technology, is not a question of if, it's a question of when. AI is disrupting every major industry. Like, it already started. In the next few years, it will only, like, become more and more...it allows for more automation and it found its way into agriculture. And like you mentioned, I mean, these last few years this changed rapidly, right, it's people starting like about self-driving cars and automated weapons and all those things.

Which some of the technology is good, some of that technology is not that good. But the tech we're currently using, some of it didn't exist three, four years ago. Some of it didn't exist, like, even in January this year. So, it's, like, really state of the art and cutting edge. And every few months, like, significant improvements are made on algorithms research and AI research. And that's gonna accelerate more and more, because, like, more and more money is flowing into the AI world, right?

I mean, like, corporations invest heavily in it. All big governments of the world create, like, the AI strategy plans. It's becoming a race and this technology will only become better. And I think it's not gonna become better linearly, but exponentially. Like, so, the algorithms we use today, in 3 years they will not be 3 times better, they will be 30 times better. And, obviously, we can reap the benefits of that and it's gonna be exciting.

Matthew: Oh, yeah. For sure. And what does your family back in Germany think about you being in the cannabis space?

Max: Well, my dad was an entrepreneur. He's, like, in his 60s now. He had his fair amount of experience with cannabis as well. Quite funny, actually. I am a very bad roller. So, when I was actually living in Germany sometimes I actually went down to my dad and asked him if he could roll me a joint. So, he's okay with that. And my little sister as well.

Matthew: Okay. And so, you got to Boulder by way of the Canopy Boulder Accelerator Program. Can you tell us about your experience with that?

Max: Of course. I mean, I'm a scientist by trade, right, so I have barely any business experience. And in that way, like, Canopy is, like, one of those life-changing events if you want. We went there, like, from Vietnam on a very short notice, like, within two weeks we quit our job in Vietnam and moved to Boulder and attended Canopy. And they teach you so much about how to build a business and give you, like, the insights and the knowledge about American cannabis industry, which I didn't know much about.

They bring in all those people from the U.S. and abroad that are, like, in the cannabis industry. Like, business guys, CEOs, investors that you can chat with and you can learn from and pick their brain, which is obviously, like, super helpful when you're, like, just at the beginning of a journey. And they really help you to prepare to pitch investors to...you learn speaking the investor's language, you learn what they wanna see, and how the world of venture capital is actually working.

So, it was, like, super helpful. It's, like, the last year or, like, even through that I just learned so much from different things of business and life that I didn't know even that they existed. And, I mean, Mike and Patrick, they are, like, obviously, through Canopy, our investors, but they also became like mentors and friends during the process. And the great thing is we can always count on them if we need some help or need introduction they are directly there to help us. If we need some mentoring they are there to chat with us. Currently, they actually still let us work a little bit out of their office, which we really appreciate. So, it's, like, a really, really great program to get started in the cannabis industry. And I would definitely do it again.

Matthew: Great. And where are you in the capital raising process?

Max: So, we raised two rounds of capital so far from Canopy and a couple of angel investors. And, you know, as a startup you are, like, always in the fundraising process.

Matthew: Yeah. If there's investors, accredited investors listening and that would like to reach out to you if they're interested, what's the best way to contact you about that?

Max: The best way is probably, like, per e-mail. It's max.unfried@deepgreen.ai. And, yeah...

Matthew: Yeah. And for people that are...people are wondering how to pronounce Max's name, we were talking about this before I hit the record button. And it's, like, unfried chicken. That's how I would describe it. U-N-F-R-I-E-D. So, Max, I want to roll into some personal development questions here to help listeners get a better sense of who you are personally. Is there a book that has had a big impact on your life or way of thinking that you'd like to share with listeners?

Max: Sure. I'm actually, like, an avid reader. And, actually, one book or one group of books I read a few years ago and currently are re-reading is the "Incerto" by Nassim Taleb, which consists of "Fooled by Randomness," "Black Swan," and "Antifragile" and "The Bed of Procrustes." And what really meant for me or helped me, in a way helps you to make smart decisions in a random world full of non-linearities and asymmetries so that you think about randomness and risks.

And, in a way, what I take from it that you wanna make decisions that if things change rapidly and things come that you cannot control, like uncertainty, like probability, that you're not getting crushed by those events. And, certainly, as like an entrepreneur, you have to have, like, an affinity towards risk. But you obviously don't wanna be, like, the sucker of events that can endanger your company or, like, yourself. So, I really appreciate that book in the way of thinking about risk and how you can, in a way, hedge against it.

Matthew: Is there a tool besides Deep Green that you consider vital to your productivity?

Max: For a tech guy, I'm probably not much of a tool guy. I'm actually more, like, the guy who takes away tools to not getting distracted.

Matthew: Ah, focus.

Max: But the focus, actually, like...focus, exactly. And two things come to my mind. One is actually, like, I changed my IDE, my integrated developer environment, to Atom, like by the recommendation of our CTO. And it actually really helps me, it's somehow better or it feels better than the other ones I used before. And it helps just to clean up data quicker and work better. And another thing which I usually do is, like, skipping breakfast. The thing is, you obviously have more time because you don't have, like, to prepare your scrambled eggs and your bacon in the morning. But also it helps me to be more focused, because you just don't have, like, those mood swings based off food until afternoon, which is very nice.

Matthew: Okay. Now, let's just rewind for a second there. What was the first thing you said, because I'm not really familiar with Atom or what you were talking about there? Can you give us a little context of what that means?

Max: Sure. Pretty much, in a way, it's a program on your computer that helps you to write code.

Matthew: Okay. Got it. Got it. It's like a code compiler or something like that?

Max: A compiler is integrated, but it's, in a way...just imagine it a nerdy version of Microsoft Word where you can write code and execute code.

Matthew: Okay. Got it. That makes sense. And as long as, you know, we...I think you might be the...well, maybe the first or second German on the show. I just wanna see if you can translate a couple of words for me.

Max: Of course.

Matthew: What does rumspringa mean?

Max: Rumspringa is just this Amish thing.

Matthew: I wanted to see if you...but what does that literally translate to in German?

Max: Like, a person who jumps around.

Matthew: Who chums around?

Max: Yeah.

Matthew: Okay. So, for people that don't know is that the Amish, before they decide if they wanna continue as an adult in the Amish community, get a period of time where they can go out and drink and smoke and cavort, do what they want. And then they can decide if they wanna live in the world of sin or go back to the Amish. And that period's called rumspringa. And so, I wanted to know what that exactly translates to because I haven't been able to find that at all. And then, one more. What does schadenfreude literally translate into, because I know what it means generally, but what does it literally translate to?

Max: My German is a little bit rusty. How will I translate that? Literally, the translation is like damage happiness, probably.

Matthew: Okay.

Max: So, it translates although the meaning is that you're happy about something that happened to someone, but it's a bad thing that happened to them.

Matthew: Yeah. Yeah. So, schadenfreude is kind of like taking pleasure in the misfortune of others and I'm glad to get the literal translation there. So, thanks for that, Max, for that random trivia.

Max: You're welcome. That was a funny random thing. Yeah. I enjoyed it.

Matthew: Well, as we close, tell listeners how they can learn more about Deep Green and connect with you and find out more.

Max: Sure. So, we have a web page, which is www.deepgreen.ai, where you can get more information. Then you can contact me by e-mail at max.unfried@deepgreen.ai. You can find me on LinkedIn at Maximilian Unfried or if you are gonna be for the MJ Biz Con in Toronto in mid-August, we're gonna be around there as well.

Matthew: Okay. Max, thanks so much for coming on the show. Good luck with everything, and keep us updated.

Max: Thank you, Matt. I really appreciate it. That was, like, a really joyful talk.

Matthew: If you enjoyed the show today, please consider leaving us a review on iTunes, Stitcher, or whatever app you might be using to listen to the show. Every five-star review helps us to bring the best guests to you. Learn more at cannainsider.com/itunes. What are the five disruptive trends that will impact the cannabis industry in the next five years? Find out with your free report at cannainsider.com/trends. Have a suggestion for an awesome guest on CannaInsider? Simply send us an e-mail at feedback@cannainsider.com. We'd love to hear from you. Please do not take any information from CannaInsider or its guests as medical advice. Contact your licensed physician before taking cannabis or using it for medical treatments.

Male: Promotional consideration may be provided by select guests, advertisers, or companies featured on CannaInsider.

Matthew: Lastly, the host or guests on CannaInsider may or may not invest in the companies or entrepreneurs profiled on the show. Please consult your licensed financial advisor before making any investment decisions. Final disclosure to see if you're still paying attention, this little whistle jingle you're listening to will get stuck in your head for the rest of the day. Thanks for listening and look for another CannaInsider episode soon. Take care. Bye-bye.