Media Week NYC theCUBE + NYSE Wired | Cole Swain, Tomorrow IO

Clips
More from Media Week NYC theCUBE + NYSE Wired

Cole Swain

VP, Strategy & Growth

Tomorrow.io

play_circle_outline Tomorrow.io: Enhancing Global Weather Predictions Through AI and Earth Observation Data Analysis

play_circle_outline Advancements in technology: Cloud compute, IoT sensors, remote sensing, AI, new space frontier.

play_circle_outline Digital twins: Simulating data to analyze and predict weather patterns more effectively.

play_circle_outline Building Climate Resilience Through Sustainable Tech Design: Tomorrow.io's Hiring Approach and Consultative Services

Info
Transcript

Cole Swain, Tomorrow IO

Cole Swain

VP, Strategy & Growth Tomorrow.io

Cole Swain, VP of strategy and growth, discusses Tomorrow.io, a weather prediction company that uses AI and satellite technology to provide global weather observations and predictive analytics for enterprises and governments. They aim to improve resiliency and help navigate weather impacts. Technological advancements, such as cloud computing, AI, and CubeSats, have enabled more accurate and real-time weather predictions. The ability to create digital twins and analyze data has revolutionized the industry by allowing for more informed decision-making and risk ... Read more

explore Keep Exploring

What does Tomorrow's weather prediction company do? add

What are some of the challenges in data and computing that have changed in the past decade, and how does current technology compare to a few years ago? add

What is the significance of digital twins in understanding and predicting the behavior of hurricanes around the world? add

What are some of the key areas of focus for Tomorrow.io and how do they plan to continue growing their business? add

bolt Powered by CUBE AI

Cole Swain, Tomorrow IO

search

>> Hello and welcome back to theCUBE's coverage here of day two, all day, bringing all the action here on the floor of the New York Stock Exchange. I'm John Furrier, host of theCUBE. We've been breaking down all the action for the UN General Assembly week. It's been climate week. Of course, it's media week with the NYSE Wired community. Big event going on, a tech talk at lunch tomorrow with Brian Baumann here at the NYSE. Check it out. Cole Swain, VP of strategy and growth, is here with Tomorrow.io. They're doing some amazing work and glad to have you on. Hey, thanks for coming in, Cole. Appreciate it.

Cole Swain

>> Yeah. Thanks, John. Thanks for having me.

>> Climate week's going on and I'm sure you get this a lot with the word tomorrow in your name because when I say, "Hey, are you going to come on theCUBE tomorrow?" Like, "Oh yeah, is Tomorrow coming on or is it tomorrow coming on theCUBE?" Love the name. Let's set the table. What's the context? What is Tomorrow.io doing and what's the keys to what you guys are working on?

Cole Swain

>> Yeah. Tomorrow's a weather prediction company. We use AI machine learning techniques to be able to predict what's going to happen next as it relates to weather prediction and we deploy our own constellation of satellites to create earth observation of space-based sounder and space-based radar technology, providing the world with global coverage to create observations for inference and training to improve weather prediction globally. Then we package all this up into a predictive analytics platform for hundreds of enterprises and governments worldwide, empowering them to deliver insights that are actionable relative to the context of their operations, helping them build resiliency programs to navigate the impacts that weather's creating for them every day.

>> Yeah, I want thank Stephen Orban over at Google Cloud and the team there. I talked with him last Friday. He runs some marketplace. I said, "Hey, I looking for some great, good guests to talk about AI." Apparently, I wasn't on top of the climate week as much as they were and they recommended you come in. I'm so glad because one of the big themes in this shift we're on right now is that all these hard problems that were not solvable, even a decade ago, are now solvable with large scale computing in the cloud. Basically supercompute being democratized right now, as with all the advances we're seeing in silicon and hardware and then the data problems are now getting solved. Again, so you're starting to see these hard problems. You're in that category where now you can do things differently with tech. You're solving hard problems. Can you scope the magnitude of, to get a sense of the telemetry of the earth and the world we live in and atmosphere? I mean, weather is a dynamic data set and I can imagine it's always changing. Scope the data challenge and the compute challenge that you have to look at and compare to how you would have to do it if it was like a few years ago.

Cole Swain

>> Thanks to the Google Cloud Ready sustainability folks, Denise Pearl, Jill Higgins, for bringing us on board. Appreciate that. Shout out. It is a big problem and you're spot on. It was very different about a decade ago than it is right now. In part, a few categories. You mentioned a few: cloud compute, IoT sensor and device and remote sensing capabilities, AI, which is a big change. The whole industry right now is going through a lot of change, as well as the new space frontier, effectively, which allows us to be able to provide our CubeSats in lower earth orbit that if you were to go back a decade ago, it's just not a possibility. When you think about your point about weather, it is a big problem. I mean, every corner point around the world, about 100 to 200 different parameters that you're modeling out. These models are refreshing, in some cases every three to five minutes, in some cases, every few hours and at some point some of these parameters go up to 50,000 feet at altitude. Some go into the depths of soil depth moisture, for instance. The data is massive and it wouldn't be able to be done without cloud compute, acceleration and we're using a tremendous amount of different types of technologies like GPUs to be able to run this multiple times to create what's called ensemble forecasting, to create risk management solutions for our customers to understand the probability of these predictions actually coming out to be true. It is a huge data set-

>> The value you can get from just disruption to life, society, business and then optimizing supply chains, to forecasting all kinds of predictive analytics models that may have been based on bad data. I mean, I can imagine huge possibilities, but I want to get the space angle because I think what's interesting is that you mentioned theCUBESats. It's been cheaper to launch than never before to get satellites up in space. There's congestion and contention for the space. Is that a problem that you see or is that just the way the world is? Does that contribute to more data or more noise? How would you react to that because we've been having some conversations with Cube. We don't talk a lot about satellites, but it is an IOT device. It's beyond the edge. It's in space, so technically it's in another distributed computing environment if you want to geek out on it, but it's providing great data. You've got visual data, but also communication data.

Cole Swain

>> The sensors that we're putting up there, they're on the radar side. They're the first of their kind in the lower earth orbit and it's providing a full atmospheric profiling, full column of the atmosphere in a way that's really never been provided before. It's taking effectively what you would consider a photograph or an X-ray into an MRI. The best analogy I can give as to the complexity of that problem, to your question earlier, it's like trying to count the pebbles on a racetrack while driving 200 miles per hour in a race car. You have to get all these individual scans and condense them all into and downlink them to your right downlink stations to then post-process this information into your prediction models, which that in its own right is a complexity. That is just one part of the whole space program from launch to manufacturing and putting these actually into orbit. It's definitely-

>> It's definitely technically fun.

Cole Swain

>> Technically fun.

>> Challenging because challenge is fun for people who like to engineer, but you're getting data and you're actually looking at problem sets that have been hard. Take us through what it's like. Take us through an example of some of the analysis, what the workflow looks like because this is mind blowing because now you're seeing things before the rest of the picture. You're going in and getting deeper layered data that you can compute in real time and/or just do processing on, nevermind getting the AI to come back in to do some assistance as well. Take us through some of the fun examples.

Cole Swain

>> The most exciting points were when our Pathfinders were up there and the company's about around 200 employees at this point, so we had just launched. We're all waiting for the downlink. The satellites are setting themselves up slowly. A few days are passing as we're unfolding them and getting them into position and all of a sudden we have an overpass over a specific type of storm that we can validate with current radar instruments that are deployed in the US, for instance. Had a great parity to make sure that what we're observing is actually correct and our team is bringing it back in and posting it on the Slack channel and we had our first observations, which that in its own right... Now, you have a different type of data set that models aren't used to ingesting and so now you have to figure out how do you actually bring that into a model to create influence on the prediction. It really was just the start of the problem, but we had set up all of our engineering and infrastructure in anticipation of that to make sure that as every additional bird goes up, we're able to improve the predictions over time.

>> I was just talking when Joseph Nelson, he's the co-founder of RoboFlow. They provide development with computer vision stuff, tooling and we're talking about the visual value of the data. If you think about the visual observation space, that's a term that we don't use very often, but if you think about from a data perspective, the observation space, it's everywhere. It's so much data, one, but then the ability to narrow it down at any given time, to find that pebble while you're going 200, you got to do a little Matrix move where you slow everything down. You can do that though with simulation, so you got digital twin capability. Take us through how you guys think about that because we just wrote an article in the Economist, our Cube research team, around digital twins. It's not just manufacturing. You can create digital twins and use them to augment analysis. It's very much Matrix like.

Cole Swain

>> Yeah, it is.

>> It's like slow it down.

Cole Swain

>> It is. It's a great point.

>> I mean, take us though-

Cole Swain

>> Digital twins, it's a key topic because let's date ourselves back, had to be about a year and a half ago at this point, hurricane notice that hit Acapulco Mexico. It was the one that rapidly accelerated over a 12 hour period, from about 70 miles an hour to about 170 miles per hour. Now, there was a hurricane hunter mission that flew into that. The same type of sensors that are on those hurricane hunters are the same type of sensor that we're putting into low earth orbit and what they were able to do is scan the hurricane and effectively start to understand its current behavior, but what that doesn't give you is what is the expected next behavior. The United States is the only country that has a hurricane hunter program flying off the coast of the Atlantic and off the coast of the Pacific, yet these types of storms are happening every day around the world. This one got a tremendous amount of attention because it hit a major city and the National Hurricane Center, for instance and many of the other global organizations are tracking these storms around the world, but they don't have the level of fidelity. They don't have the ability to create a full-profiled, three-dimensional digital twin and a 400 kilometer swath overpass to recreate that in real time, to then understand what's causing these changes to happen and that's what's really exciting about what we're doing is because when we're refreshing every hour for every corner point around the world, we're collecting a data set that is going to be instrumental to training these AI weather models and for inference on these weather models. That is just fundamentally changing the industry as a whole.

>> The reinforced learning will come from additional data you might be getting off your analysis. This is kind of where the tech goes. Again, back to what my original thesis from when Dave and I riffing on our CUBE pod, this is categorically new applications, that's first generation because there are now problems to solve. We're kind of living in the world where if you're an entrepreneur or you're an engineer, you're in a field day here.

Cole Swain

>> In the age of this generative AI, it's changing the industry over the last two years in many ways, which used to be very formulaic in a math-based equation to predict what the weather's going to be. Now it's a neural network style and it's the convergence of observations and AI together at the same time that's accelerating a lot of this.

>> Well, climate week is interesting here because I've always been saying, I've been very vocal on theCUBE over the past 15 years. Dave, he always rolls his eye because I kind go on a rant, but sustainability has been kind of a bolt on, kind of lip service. That's my words, but it's gotten better over the years, but it's been like a white paper, but I think companies like AWS, like Google and others are seriously building in sustainability into the design process. Now for the first time, we got the capability to make sustainability stuff. Whether it's if you're affecting the environment, whether you're building a data center or working on these big projects. To build in IP with large scale supercomputing capability in the cloud, to do sustainability. There's no excuse not to. It's like security. Why wouldn't you not build it from day one? What's your answer to that? If you can give navigation guidance to folks about how to make that work better, how do you build some of this kind of big tech? You guys are doing large scale big data tech. You're doing large scale, big data cloud, doing large scale space. I mean, you guys are doing a lot of things.

Cole Swain

>> There's a lot of pieces.

>> There's a lot of hard parts there. So how do you advise a simple organization to build in something that's going to be helpful for the environment?

Cole Swain

>> The sustainability category was one we were always put into and it made a lot of sense because we're doing earth observation, we're helping many organizations around the world with our observations to improve their sustainability practices, but for us, it was always a category that made it a lot of sense, but it didn't make sense in a way because we're not driving critical emissions improvements directly through our product.

>> You're indirect impact. You're enabling more impact from others.

Cole Swain

>> Where it started to connect more recently is this topic of resiliency that's becoming-

>> Explain what you mean by that.

Cole Swain

>> Helping organizations be prepared to mitigate disruptions that they would otherwise get slowed down or cause tremendous amount of challenges to their operation, by being proactive and being able to get around them with predictions that they're prepared for. That category of resiliency, it seems to be a hot topic where-

>> What resilience would it be? Would it be physical resilience? Workflow? Operational resilience? All of the above?

Cole Swain

>> If you were to think about how do you quantify what climate risk is for a nation or for an enterprise.

>> It's situational.

Cole Swain

>> It is.

>> It depends on who the customer is. If you're in shipping and you're in supply chain logistics, you will be impacted and disrupted by weather.

Cole Swain

>> You will be.

>> You would let that same modeling around, hey, let's take the slow route rather than going over the Sierra Mountains.

Cole Swain

>> Do you have an action plan? Do you have your schedules watching it? Do you have all your communication programs in place and do you have the best in class predictions?

>> They don't call it Donner Pass for nothing over there in Silicon Valley. If they had weather predictions, they probably would be alive. That was an example, but this is disruption, so weather is huge. Economics. Food. We've covered this a lot with IBM and the weather when they had the weather company, but you're getting more technical with it now. You're getting enabled. What are some of tech that you guys have seen that's helped you? Can you talk about some of the advances in cloud and compute that you guys take advantage of? What are some of the key hardware fail opportunities?

Cole Swain

>> Without it, we wouldn't be able to modularize and deploy models globally and scale them up and experiment and sandbox what the best techniques and tactics are, globally, as well as, again, I keep saying it, but this convergence of AI is very helpful because it's key is it's all about how much you're observing the earth to train the models and that's what's changing how we think about it is, the earth observation we create allows us to run training in high levels of compute and inference to improve predictions for anyone who's building a weather model.

>> Cole, I'm not going to lie, I really like what you guys are doing. It's really kind of exciting. If I was in college right now, I'd be like, I want to work there. This is an area that's going to have a lot of headroom that you're working on because it's got a lot of exciting elements to it. You're solving new problems, you're pioneering new territory. You have great impact, mission driven, checks all the boxes. For the folks watching, what are you guys looking for? What's it like to work there? What kind of people do you hire? If someone wants to get involved in Tomorrow.io, what do they do? Talk to the folks out there. Give the plug.

Cole Swain

>> Yeah, you got to be ready to immerse yourself in a lot of scientific aspects that you might not already be prepared for. You got to be resilient. You got to be persistent. You got to be tenacious. Leave the ego at the door. Be curious and really just come in ready to work because that's what we're doing because we're doing a lot. We're in space. We're running our own weather prediction models. We're delivering insights and analytics. We're doing all three. We fully vertically integrated solution, driving weather and climate infrastructure and intelligence for the world. We are the acting national weather service for countries around the world. That's just a small dollop of the impact that we're able to make.

>> You're at 100, 200 people, you said?

Cole Swain

>> Just shy of 200.

>> You're the hiring?

Cole Swain

>> Of course.

>> Any areas in particular you want to shout out to? Engineering?

Cole Swain

>> Anything go to market. Anything space and anything engineering. We're always looking for top talent.

>> All right, great. What's next for you guys? What's on the business plan?

Cole Swain

>> More launches. More customers. What's really interesting for us is a consultative service of how we work with our clients and helping them understand the cost of their operations when they're not paying attention to weather and how we actually put a number behind this because not everybody has a budget for weather intelligence solutions today, but they do have budgets for KPIs and weather has a role in improving those KPIs.

>> AI can give them. You can get them a dashboard to understand what's going on. We'll have to follow up and have another conversation on cyber security in space. One of my other favorite topics. You can't do break fix in space either.

Cole Swain

>> You can't do break fix in space.

>> And how do you manage any kind of malware? I mean, it's a whole other level.

Cole Swain

>> That's it, man.

>> Cole, thank you so much.

Cole Swain

>> Yeah, thank you John.

>> Okay. Cole's here inside theCUBE. We're breaking. Sometimes I feel like I'm in space with all this action and all this intoxicating tech conversation here on theCUBE at the New York Stock Exchange show floor. I'm John Furrier. You're watching theCUBE. We'll be back tomorrow for more media week with the NYSE wire community and theCUBE wire community coming together. Silicon Valley, New York, bringing all the action to you. Thanks for watching.