They discuss how Rainstick mimics natural thunderstorms to create electric fields and frequencies that promote plant growth, challenges and breakthroughs in their research, and their participation in the AWS Compute for Climate Fellowship.
Episode notes:
Rainstick uses electricity to mimic the natural effects of lightning to grow crops bigger, faster, and more sustainably.
Want to learn more about the Compute for Climate program? Check our podcast with Lisbeth Kaufman, Head of Climate Tech at AWS.
Ryan wrote about how software is being applied to agriculture a few years ago.
Connect with Darryl on LinkedIn.
Congrats to Lifeboat badge winner WestCoastProjects for their answer to Test accuracy is greater than train accuracy what to do?.
TRANSCRIPT
[Intro Music]
Ryan Donovan: Hello everyone, and welcome to the Stack Overflow podcast, a place to talk all things software and technology. I'm your host, Ryan Donovan, and today we're gonna be talking about making it rain, actual rain, the technology of it, and using lightning with my guest today, the Chief Rainmaker at Rainstick, calling all the way from Australia. So, thank you for joining me today, Darryl.
Darryl Lyons: Hey, Ryan. Excited to be here.
Ryan Donovan: So, Darryl, tell us a little bit about how you got into software and technology.
Darryl Lyons: My background is probably doing lots of free labor on my parents' farm. And so, by default, kind of went, 'hey, how do I shortcut this labor and bring some check onto the farm?' So, that was probably about 10 years ago, and now I'm into my third agritech company. So, that was basically to get out of doing a hard task, and how I can implement technology into limiting that.
Ryan Donovan: Yeah. I think you'd be surprised how much technology is actually short-cutting work.
Darryl Lyons: Yeah.
Ryan Donovan: So, you founded Rainstick Technologies. Tell us a little bit about what that is and what's interesting about the indigenous technology that you all are using.
Darryl Lyons: So, Rainstick mimics the natural effect that you have in thunderstorms, which creates electric fields and frequencies in the air. So, we've set out and built tech that delivers those frequencies through electric fields into biology, starting with doing that into seeds. And where that links to my heritage – I'm a proud Maiawali man, and my tribe are rainmakers, and they notice that effect and influence thunderstorms, and that grew our grain for tens of thousands of years. I've taken a role as Chief Rainmaker. I've worn around that background with a lot of purpose because I totally believe First Nations knowledge can help us create new technology that can make a massive difference on the planet.
Ryan Donovan: That's interesting. And how do these frequencies actually affect the seeds, I assume, or it's other biology?
Darryl Lyons: Yeah, so, we map what we call 'recipe,' and what different frequencies to get a very targeted biological response. So, we have different frequencies, or combinations in a recipe that can grow more roots or grow more shoots. So, we're aiming to map that to agriculture, and also doing work in nature restoration, and how we can accelerate the growth to improve ROI and profitability for farmers, and then improve establishment in some of those restoration activities.
Ryan Donovan: That's interesting. So, right now, do you have a technology that is, you know, zapping seeds to make 'em grow? Or is it in the prove-out stage?
Darryl Lyons: We're still in prove-out. Officially, we're only two years old; unofficially, just about three years old. So, me and my Co-Founder, Mic Black, who's the tech Co-Founder, we started out just using our own cash to work on this in our garages to see what's happening there, and then found our 'aha moment' around how different frequencies have different effects on different biology, and went, 'wow, there's something here. Let's dig in and form a company.' So, that was just over two years ago.
Ryan Donovan: Mm-hmm.
Darryl Lyons: Now working in seeds, we've done all that in our lab. So, you know, if I pitched to you to go, 'hey, I'm gonna mimic what happens in nature and thunderstorms, and it comes from my ancestors and a tribe. I'm going to use this rain stick, and we're gonna change the world,' a lot of the research community would be like, 'you guys are a little bit out there. It's a bit too crazy. Just go away.'
Ryan Donovan: Yeah.
Darryl Lyons: So, we kind of really had to lean into that and went, 'okay, how do we embrace that skepticism?'
Ryan Donovan: Mm-hmm.
Darryl Lyons: And like, 'why wouldn't it work? Or what would you need to see?' So, we hired a couple of researchers. So, over the last 18 months, we treated a lot of seeds. We've measured phenotypical traits and outcomes of over a hundred thousand seedlings using a bit of machine learning to quicken that process, so we're not stuck there with a ruler and measuring it all. And then now, we're presenting that back to the industry and the research community, and they're like, 'okay, yeah, you guys didn't need a straitjacket back then. There's definitely something.' So, we've got a lot more interest in what we're doing. So, now we're proving that out with larger commercial corporates around the world and a lot of the research community to validate what's going on. So, we're doing, you know, more validation trials over the next 12 months, not trying to rush it to get commercial too early.
Ryan Donovan: Yeah. So, how much of this sort of frequency testing is, you know, to put it jokingly, tasing seeds in a lab, and how much of it is running simulations in physics spaces, but without actual physical seeds?
Darryl Lyons: Little bit of both, I guess. This is a new emerging field of science called bioelectricity, and this is kicking off probably in the States and globally to start, and I think it's gonna have a profound effect over the next couple of decades. We believe we've probably got the biggest database of the effect of bioelectricity on plants. We've probably got about 5-6 million data points on that at the moment, hence why we're in this 'Compute for Climate' fellowship and program to really put that into models to understand those effects. So, we're just starting with mapping that now, and mapping our range of frequencies. We have 50,000 frequencies we can choose from.
Ryan Donovan: Oh, wow.
Darryl Lyons: Nine different waveforms we can deliver those in.
Ryan Donovan: Mm-hmm.
Darryl Lyons: And then any time duration, from a nanosecond to a full day. And so, that's why we're needing a lot of Compute to kind of work out, 'what is that?' And then, you've got all of this environmental data and how the plant grows, weather data, all of the genome data, and epigenetic data. How do we start bringing that in to then understand which frequencies are getting an output, and a cascade of effects from the genetic output, all the way through to an output they want in improving yield?
Ryan Donovan: Right. Yeah. It sounds like it's a 'cog in the toric' nightmare with– what is it, 50,000 frequencies, nine, and then all the physical variables there?
Darryl Lyons: Yeah.
Ryan Donovan: You came to us through the AWS 'Compute for Climate' fellowship. Thanks to them to sending you over. What kind of Compute are you using? I assume you're not just, you know, on some MacBook somewhere running tests. You need a little bit beefier systems for that, right?
Darryl Lyons: Yeah, we're excited to get in the fellowship and excited that AWS has actually recognized that indigenous knowledge has a potential, and I guess this is the reason I've really leaned into it. On one side of my family, I've seen my grandfather lose his farm to drought, [which] affected him a lot, and my dad finally got on the farm and then, you know, a decade or really bad years. So, with climate variation coming, we've got this huge problem with producers around the world. So, really keen to see how we can make a difference. And then on my other side of the family, from my mom's side, I just really observed that there's hardly any traditional knowledge that's really come into farming, and practices globally. So, how can we link those two together? So, yeah, we've got a lot of work to do improving how we do it. So, I guess we've picked tomatoes as our key species to work with the Compute and use the data we've got. So, we've got a whole lot of Compute power in this fellowship to work on. So, tomatoes have a lot of public data sets. We are mapping genetically what our up and down regulation of genes are from each frequency suite we put in; then we map that to where we're pushing. We have early data that, when we have this more vigorous seedling, it's actually taking up more nitrogen, and nutrients, and then converting that synthetic nitrogen into the leaves, and then photosynthesizing better. So, there's a lot of public data sets that show this cascade effect to get to this output what all producers are wanting – this nitrogen usage deficiency. So, that's what we're mapping now on that gene regulation to pull a model to go in this grow environment. We want to pull the trigger and upregulate these genes to get this cascade effect. We need a lot of Compute for that.
Ryan Donovan: Okay, and you said you're using machine learning. How difficult has it been to get accurate results out of the machine learning data? Because, like you said, there's a lot of physical variability. How much is there a feedback loop in learning on how the model should work based on physical results?
Darryl Lyons: Just at the start. So, we're just kicking it off with this program. It's always been our goal. We're lucky enough to be selected. I guess we're probably one of the earliest-stage companies in the Compute for Climate cohort. So yeah, we're early stages. I guess we've got a physicist on our team who does a lot of the theory around what's happening when we're putting those frequencies in, and what potentially the cascade effect in the seat. So, now we're gonna combine all of that knowledge, and then put it into this model, and then look at how we get phenotypical data, the epigenetic data, and the genome data and transcriptomics to understand how we can get our first pass of what this is when, you know, we've got three months in this fellowship to do this, but we know this is a five year journey for us. And then, agriculture's really starting to lean in to really get all these massive data sets and put them into LLMs. So, we think it's the right timing. It'll be a huge opportunity to link all of those to then get to the output we want.
Ryan Donovan: Mm-hmm. Yeah, using the indigenous technologies is an interesting one. I think there is a bias towards certain folks, at least in the US, to be all-in on anything that is old knowledge or Eastern knowledge, sort of an anti-corporate way of thinking. But there's a lot of those old knowledges that have been picked up and processed by science, like the bark of the willow tree is Aspirin, right? How did you all decide to pick this technology?
Darryl Lyons: I guess this is probably something a little bit different 'cause there's thousands of examples of that and trying to extract something out of plants. So, this is more around their observation. I believe my– well, this wasn't handed down to me because in Australia we had a pretty rough history–
Ryan Donovan: Sure.
Darryl Lyons: And a lot of that knowledge has been lost. So, you know, I believe mine were really in what's called water cycling. So, how they could actually link into the biodiversity, and get that rain, and get that water cycle to bring life. So, you know, what we're triggering is one part of that, and what's the reaction at that cell level? 'Cause we believe biology is electric, every cell is electric, and we're demonstrating it can be influenced to change its reaction through frequencies. So, it's a unique example. I haven't seen a lot of– how do you come up with this base knowledge, or overview, or concept, and then take that in, and create some new tech around that, and then do that? We've got a special class to share that goes to the Maiawali Foundation that can never get diluted, and we're deep tech company, so we have to take in VC money to build this technology that's in recognition of that traditional knowledge. It's in our shareholder agreements, all in our customers, everyone has to agree, and acknowledge it, and abide by it. So, yeah, I hope– I've seen a lot of different traditional knowledge in Australia, and that's a huge opportunity to expand that rather than, I guess, in Western society, where we're boxed in to go, 'cool, we're just gonna take aspirin from the bark, or this molecule from the plant.'
Ryan Donovan: Right. Are there any sort of risks to humans or other biological life forms? Like, is there, you know, a possibility of using this for pest control or something?
Darryl Lyons: Yeah, so, there's a bit of a history in the early 19 hundreds, where people were using frequencies to heal certain health ailments in humans, but it kind of got ridiculed in the 1930s, and got put away. But now it's coming out in this brand new field—we're calling it brand new—field of bioelectricity. So, they think it's gonna have a profound effect [on] it. And again, my belief is it's been known for a long time from my ancestors with the use of the rain stick, and understanding that effect. So yeah, we tested here, so you can go a bit far – certain recipes we have actually send plants backwards. Plus, we've also done some work in mycelium, and microbes, and some mold where we can inhibit mold growths. So, I believe it could be a form of pest control, as well.
Ryan Donovan: Yeah. I'm sure you know when you started approaching people, anything that's like using electricity to grow crops, or to heal, has a little whiff of woo to it. What did you have to do to convince people that this was a real technology?
Darryl Lyons: On top of that, we add in the frequencies, and I guess through the 70s, you got the hippie vibe of the frequencies, as well. So, it's like totally double words. So, we really needed to embrace that cynicism, or more skepticism, and go and really question, what would you need to do? Or, how would you need– whatever. So, that put us on a trajectory where we really measured the phenotypical outcome of the traits of the plants growing bigger and faster, and then presented that back. So, we've done lots of experiments. We've measured over a hundred thousand seedlings and presented that data back to the industry and the research community. And people go, 'okay, yeah, there's definitely something going on here. How do we get involved?' And 'let's explore it further.'
Ryan Donovan: And I am sure that that data collection – that's part of what you use the cloud competing for, I'd imagine, yeah?
Darryl Lyons: Totally. So, initially, we first started with hand measuring. We built our own tools and our own bio prospecting lab to automate a lot of that process with machine learning, but there's so many variables in large data sets in agriculture. So, you've got the genome of the plants, there's heaps of varieties of every species, the epigenetics, the grow environments. So, there's truckloads of data sets, hence we're in this Compute for Climate program with AWS to help us bring in those data sets to understand how we can improve our recipes to get a better outputs.
Ryan Donovan: Yeah. Thanks again to Elizabeth Kaufman and her team for connecting us. So, you must have a large range of scientists there as part of your team – physicists, biologists, and do you have software engineers, too?
Darryl Lyons: We do, but it's not very large, Ryan. We've only got a team of five people, so we tick off for each of those categories. My Co-Founder is a very ice-breaking tech genius, so he's basically come up and built the technology with a physicist doing a lot of physics and chemistry. And then, he's also built some LLMs in his prior companies and prior work. So, that's helping us automate and bring in larger data sets. And we're expanding that team as we go on this mission to meet our longer-term goals, which [is] really, basically, creating a recipe engine, and be able to pull in all these data sets to understand what is the right recipe for the right result for that species, and variety, and grow environment.
Ryan Donovan: That's interesting. So, I know, you had a five-year plan and three months with the Compute for Climate fellowship. What do you hope to get at the end of both the three months and then the five years?
Darryl Lyons: We're grateful for being in the program and the access to those credits to be able to start this process. It's always been a north star and a goal to be in three or five years to have this recipe engine. But I guess, with being a small team and only an early-stage deep tech company, to have those resources to be able to start that process. So, at the end of three months, we're actually aiming to create that recipe engine to be able to predict what the right recipe is for, probably, three varieties of tomatoes, to improve that nitrogen uptake efficiency. And then off the back of that, we can continually keep working on that, and we'll be bringing in a team of software devs to work on that very hard, and continually go on that process to map that gene expression in different species to understand how we can do that pathway. 'Cause there's a huge effort globally as we need to transition agriculture more sustainably, putting less nitrogen or having the plants more nitrogen-use-efficient is a huge opportunity, and a huge opportunity to reduce agriculture's impact. So, yeah. That's where we'd like to be in three to five years. Can we have that recipe engine and create that for multiple species across agriculture?
Ryan Donovan: Okay.
Darryl Lyons: I can't mention the name of the company, but we've just signed a contract with a global mining company to do mine restoration. So, it's always been a hunch that this tech will work on native seeds. So, we're really keen to work on nature restoration. So, that's something we're just adding in, and we've– in the Pilbara is where Australia probably exports a lot of iron ore.
Ryan Donovan: Mm-hmm.
Darryl Lyons: And it's a really, really dry environment, and some of the hardest to germinate plants in Australia are in that environment, and some of the hardest in the world, and we've been doing some work in the lab showing that this tech can have a profound effect on that. So, we're about to start field trials and work on that. So, it's another huge, big opportunity that we're very passionate that the technology can make a difference.
Ryan Donovan: And that's a whole new plant to gather data on, isn't it?
Darryl Lyons: Yeah. And again, those plants only have subsets of the map gene, so most of those species aren't really mapped like in agriculture, where they're trying to create new varieties. Like in, you know, every year with large companies putting lots of money in. A lot of native species don't have any large data sets, so I guess that goes into your earlier point and how we can potentially create and use Compute to create models, and put stuff together to get improved output, and establishment, and more vigorous plants growing.
Ryan Donovan: Hmm. Well, that's exciting. As we say in the software industry, that's a greenfield project.
Darryl Lyons: Yeah.
Ryan Donovan: All right, ladies and gentlemen, it's that time of the show again where we shout out somebody who came onto Stack Overflow, dropped a little knowledge, shared some curiosity, and earned themselves a badge. Today, we're shouting out the winner of a lifeboat badge – somebody who found a question that was drowning with a score of negative three or less, and they answered it so well that they got themselves 20 points and brought the question up to three. So, today, congrats to WestCoastProjects for answering 'Test accuracy is greater than train accuracy what to do?' If you're curious about that, we'll have the answer in the show notes. I'm Ryan Donvan. I edit the blog, host the podcast here at Stack Overflow. If you have questions, concerns, comments, topics to cover, et cetera, please email me at podcast@stackoverflow.com, and if you wanna reach out to me directly, you can find me on LinkedIn.
Darryl Lyons: Hi, my name's Darryl Lyons. I'm the Chief Rainmaker at Rainstick. You can find myself on LinkedIn or on rainstick.com.au, an Australian-based company.
Ryan Donovan: All right. Thank you very much for listening, everyone, and we'll talk to you next time.