How Deep Learning Is Helping Blind People with Karthik Kannan, Founder of Envision

(Intro Narrator at 00:00:03) Hello, my friends. Today, Joel is talking to Karthik, founder of Envision, and they discuss how Envision is combining deep learning software with Google Glass to help blind people see, their strategy for getting this tech into the hands of as many people as possible, and how Karthik is motivated to build technology by the impact he can have on people's lives. All of this right here, right now, on the Modern CTO podcast.

(Karthik at 00:00:32) Here we go.

(Joel Beasley at 00:00:33) This is the Modern CTO podcast. I'm excited to talk about the technology that you're building. What exactly have you built?

(Karthik at 00:00:49) So what we've built with Envision is a smart glass solution that helps people with a visual impairment to live more independently using computer vision. So the Envision Glasses is the main product, and it's a tool that helps a visually impaired person to read text from any printed surface, be it a document, be it a computer screen, be it a Coca-Cola can, for example, in over 60 different languages from pretty much any font that you can think of. It can help them recognize faces of their friends and family, so you can teach it to recognize faces. You can also teach it objects around you, so it can recognize objects like your coffee mug and things like that.

(Karthik at 00:01:29) And apart from all the AI stuff, the glasses also allow you to make a video call directly from the glasses to a friend or a family member. So when you're wearing the glasses, it's got a camera in there, so the person on the other end of the video call sees your first-person perspective. They get your audio, they get your video, and then the visually impaired glasses user gets the audio.

(Karthik at 00:01:52) So they can make video calls directly from the glasses as well. And the main idea is that you don't really have to hold a phone in your hand, because when you have a phone, you also have a cane or a guide dog on the other hand, and your hands aren't free anymore. You're just pointing stuff around, and it's an incredibly frustrating, cumbersome experience. And smart glasses really solve that problem of having your hands free and being able to do things more easily. That's what we do.

(Joel Beasley at 00:02:19) That's amazing. And why this? There's so many problems out in the world. Why solve this one?

(Karthik at 00:02:24) So it all started about six years ago, before deep learning and artificial intelligence was really the hot topic of the day. Before all of that, at the very, very beginning, I was really into machine learning at a very early stage, simply because I had a lot of friends who worked at NVIDIA at that time who were really into deep learning, and I was also doing some part-time work for them and consulting and things like that. So it really got to a point where I was hearing about this from everyone around me, and I started to slowly get into it myself. And it was at that moment, around that time, that I had a chance to go to a blind school in my hometown in India to talk to high school kids who were just graduating. They're visually impaired.

(Karthik at 00:03:13) They'd been in the school all their life, and they're just going out into the real world. And so one of my friends invited me there, and she said, you know, just have a chat with them about what you do as a developer, as a researcher, as a developer, what do you actually do, right? And it was supposed to be a 30-minute conversation.

(Karthik at 00:03:29) I remember this so clearly. I remember going there, talking to these kids about being a developer or a designer. It's just about solving problems. I wake up every day in the morning, I solve problems for a living, right? And I just volleyed a question back to the kids there in the audience about what kind of problems would they like to solve when they step out of the classroom and go out into the real world, right? And a lot of them started talking about being able to do things more independently. A lot of them said they want to read a book independently.

(Karthik at 00:04:05) They wanted to be able to go out independently, make impromptu plans, right? Like, just if you see a nice sunny day, they want to be able to step out, go wherever they want to go entirely by themselves. That whole 30-minute conversation snowballed into a two-hour, sort of life-changing experience for me, talking to these kids and understanding how independence was something that sighted people took for granted so much, and how for a lot of the kids in that audience, independence was such a prized thing. And especially in a country like India where there is not a lot of awareness around accessibility, right?

(Karthik at 00:04:40) It was a huge, insurmountable obstacle for them. And it was around that time that I had a chance—I do have my cofounder who's also called Karthik, and he's a designer, an industrial designer. And right after that conversation, the two of us met, and I was just talking to him about how there are many aspects of deep learning that were starting to become as good as humans or even better than humans in some instances. For example, OCR, optical character recognition, or being able to recognize text.

(Karthik at 00:05:10) Computers were really outperforming humans in some areas there. And I was talking about how we could actually take these advances, combine that with really good design, and see if we can build something that could help people with a visual impairment, right? And in the very beginning, the two of us started working on this, not with the aim of becoming a company. That was never actually the aim.

(Karthik at 00:05:33) I was in between jobs at that time. I was going to get married, so I just took a six-month sabbatical. And I thought this was a really interesting problem to work on. And the idea was just make it into a project, put it out into the world, and see how there is—and just raise some awareness around this, try to showcase AI as a thing that can help people with disabilities or AI and accessibility could work together. And so that's how the Envision project started, and I started working on it.

(Karthik at 00:06:01) I very clearly remember, in February 2017. And the initial idea was that my cofounder, he was studying industrial design in the Netherlands. So he was here in the Netherlands already, and he would do all the design and the user research part, and I would do all the prototyping and actual building of it. And so our aim was just create a simple prototype app, throw in image captioning, throw in OCR and a few other things, and then just showcase how AI can work really well for accessibility. And so we started working on it.

(Karthik at 00:06:34) And over a period of time, over like six months, we had around 400, 500 people on our TestFlight beta who were just using the app, using this early clunky prototype app that had like a 40% crash rate. And there was no framework or no software at that time, like how there is today, to deploy machine learning models on device. So we had to build a lot of it by hand. And I remember just spending a lot of time on it and working on it, and people absolutely loved it, right? It organically grew to 500 odd people.

(Karthik at 00:07:05) And at the end of six months, I still remember we ran out of funding from the university. I was a bit overdue on my credit card bill, so I was like, okay, I'm going to wrap this thing down. And I still remember, on a Friday evening, I sat down and wrote out an email to all these 500 people, thanking them for being a part of this project and saying, hey, we're going to shut this thing down. It's been amazing. You guys have shown that AI and accessibility could be a great fit, and we're going to take all of this stuff, we're going to try to put this out there and try to showcase to the world that this is something really cool.

(Karthik at 00:07:42) And I remember waking up the next morning. It was a weekend. Almost all of the 500 people got back to me. Like, every single one of those people actually wrote to me and said, why are you guys shutting this down? I've been using this to do my schoolwork.

(Karthik at 00:07:55) I've been using this to read stuff. I've been using this to operate my washing machine independently. And I've been able to do so many things around my life independently. It's just stupid that you guys are shutting it down. If you're serious about this stuff, you should actually build this as a company, right? And it was at that point that it actually struck us that there are people out there who really are benefiting from this kind of stuff. And if we're serious about this—and it's a huge passion for me to be able to build technology that can have impact on people.

(Karthik at 00:08:28) It's been this itch that I've had ever since I first encountered programming as a kid. And I finally have this opportunity to scratch that itch, and I just thought, yeah, screw it, let's do it. And I moved from India all the way to the Netherlands to start the company, and we've been doing it for almost five years now ever since.

(Joel Beasley at 00:08:47) That's unbelievable. So how did you end up getting money for it to get started?

(Karthik at 00:08:53) So in the beginning, I had a bit of savings from my job back in India, so I took all that. I just got married back then. I was barely married for a month, so huge props to my wife for letting me actually do this, because it's insane that someone would actually go on a whim from India to the Netherlands to start a company building—accessibility, AI was all very new, but I wanted to just take a stab at it, right? So I moved to the Netherlands, and initially we got some additional funding from the university that kept us going.

(Karthik at 00:09:24) And I still remember the first two months that I was here in the Netherlands, we worked really hard. I think seven days a week, almost 12, 14 hours a day, trying to complete the app. And our aim was to try and use the revenues from the app. So we made the app a subscription model, and we started to charge users €5 per month for using the app. So that initial revenue that we were able to bootstrap ourselves, that kept us going.

(Karthik at 00:09:51) And of course, we also took a personal loan. We were part of an incubator in the Netherlands. And through the incubator, we were able to take a loan from a bank, and that kept us afloat. It kept us going personally. We didn't draw salary for the first year, year and a half.

(Karthik at 00:10:05) I just took the personal loan and then used that for my day-to-day expenses here in the Netherlands and bootstrapped our way in the initial couple of years. And then we raised a small seed round of funding and so on. But it was initially totally bootstrapped by people in the community, people who just came to us and willingly gave us money for the app that we were building, and that's how we kept us afloat.

(Joel Beasley at 00:10:30) And where is the company at today?

(Karthik at 00:10:32) Well, the company is great. We have tens of thousands of people using the app. In fact, we just made the app free. It used to be a subscription-based app, but we've been able to really bring the technology to a point within the app where we can offer it for free.

(Karthik at 00:10:48) We now have the Envision Glasses, and we're about 30 people right now, 30 people from over 25 different nationalities. Yeah, it's been an amazing journey so far.

(Joel Beasley at 00:10:57) I think you use the Google Glass hardware, right?

(Karthik at 00:11:02) Yeah.

(Joel Beasley at 00:11:02) At what point did you use that? Was it from day one, or did you use something else at first?

(Karthik at 00:11:07) Yeah, actually, we wanted to make Envision as a wearable from day one, because when we were doing our user research, there is actual scientific research to show that visually impaired people have better head control than they have hand control. And when we started talking about this idea, even as just a concept, every single person was like, this has to be a wearable, right?

(Karthik at 00:11:29) And in fact, the earliest design of the Envision Glasses was this AirPods-like design that my cofounder came up with. So completely unrealistic, but we just imagined how it would be a wearable. So the idea was always to make it a wearable. But in 2016, 2017, wearables were like this really alien concept. It's still a very alien concept.

(Karthik at 00:11:50) I mean, a lot of people—wearables aren't like something that we use day to day. Smart glasses are still kind of science fiction-y stuff. But back then, when we found smart glasses, there were two problems. One, they were too stigmatizing.

(Karthik at 00:12:05) A lot of visually impaired people felt in the early days that why would I want to wear like a 500-gram headset on my face and make myself look like a Robocop, right? That's the kind of smart glasses over there. Or if the smart glasses looked good, they weren't really powerful enough to run our software. So we had to try and find like a middle ground, right? Something that wasn't stigmatizing but still was powerful enough.

(Karthik at 00:12:26) And so we just bided our time and tried to keep ourselves afloat with the app and with the funding and all those things until we could try to get to a point where smart glasses were really good enough to put our software on. And that's when the Google Play Award happened. So we won the Google Play Award in 2019 for the best accessibility experience, and that's when I had a chance to go to Mountain View headquarters at Google.

(Karthik at 00:12:54) And I remember taking my award to Google I/O and then just really exploiting the hell out of it, trying to get myself in front of every single Google AR, VR person with the same pitch. Like, hey, we've got this tech. It's won this award. People really want to put this on smart glasses.

(Karthik at 00:13:12) If there's something that you guys have in the works, please tell me about it. We'd love to do something with it, right? And I remember—I was there for a week, and I remember having like 20 meetings or something, just packing my day, trying to meet whoever would listen to me. And at some point, I spoke to someone higher up in the AR/VR food chain at Google, and they said, okay.

(Karthik at 00:13:33) We're actually working on a new version of the Google Glass, the device that everybody loves to hate. But we're not going to make it for consumers. We're going to make it for enterprises. So we're going to make it such that it would be helpful for someone like a DHL that has a huge warehouse, and people could wear the stuff, scan barcodes. But since you guys are super passionate about wanting to build this for visually impaired people, we think it's a cool idea.

(Karthik at 00:13:55) And they made me sign a bunch of NDAs right there. And then they said, okay, take these two Google Glasses with you back to the Netherlands. Don't pay us for it. Just take it.

(Karthik at 00:14:05) Do whatever you want to do with it. And if it's good enough, let me know. And I remember on that flight just coming back, I was reading all the developer documentation on the flight itself, and I came back to the Netherlands. The moment I landed, I just took the glasses out, started putting the Envision Android app on the Google Glass and making some tweaks to the underlying software itself, because the Google Glass doesn't come with accessibility software. It's just stock Android.

(Karthik at 00:14:35) So we had to build the accessibility layer and all those things. So we started working on it, and I remember somewhere around December 2019 that I just remember giving a pair of Envision Glasses to a visually impaired person who had come to the office and didn't want to give it back to me. So that's when I knew we had something, something solid, right? And that's how the Envision Glasses started, basically.

(Joel Beasley at 00:15:01) That's amazing. That's so cool. Tell me, I want to walk through some of the features. What do they

(Karthik at 00:15:07) So the thing that it's most used for is to read text. That's one of the most revelatory things for me as a sighted person, because when you think about it, every single thing around you is text. The UI on your phone screen, on your computer screen, the stuff that you see around you, the products—every single piece of information out there in the visual world is text.

(Karthik at 00:15:32) And I realized how important it is for people to be able to read text from any printed surface possible. It could be a curved surface like a Coke can or a water bottle, right? So we started to focus on building computer vision tech that could really read pretty much any printed surface or any kind of text possible, regardless of the font, regardless of it. So that's what the Envision Glasses does really, really well.

(Karthik at 00:15:58) So if you are someone who is trying to read a handwritten document, if you have cursive handwriting, we put a lot of effort into making the AI really good at reading cursive handwriting. Or if you're reading stuff from a curved surface, or if you're trying to read a multi-language document. So the glasses are really good at reading text, and that's what people mostly use it for. Almost 80% of our users or 80% of the glasses usage is focused on reading text. And with reading text, there are two parts. One is being able to scan a piece of text correctly, and the second thing is understanding the layout of the text itself.

(Karthik at 00:16:34) Because if you're a sighted person and you're wearing glasses that have cameras on them, right, you can know more or less where to hold the document and take a picture. But if you are someone who's visually impaired and you have no concept of sight ever since you were born, right, how do you really guide them to take the best possible picture with the glasses? So we focused a lot on building AI that guides a user on how to take a picture of a document and automatically captures it for them when the document or the book is fully in frame. The second thing is documents or any piece of text has a lot to do with the layout of the information.

(Karthik at 00:17:11) For example, if you're reading a magazine article, right, a magazine article or newspaper article has columns in them. Now if you just pass that particular image to a regular OCR engine, it just reads everything from left to right. It has no concept of what is a column.

(Karthik at 00:17:26) Right? What's a magazine article? What's a newspaper article? What's a table? So we had to build a lot of intelligence into recognizing the layout of a document and reading it in the exact manner as intended.

(Karthik at 00:17:38) Right? So if you're reading a magazine article, the OCR engine should understand that it contains columns, so it needs to read the columns correctly instead of reading it just left to right. If you're reading a letter which doesn't contain columns and it's just regular text, it needs to read it out in the correct fashion. So reading text is one big aspect of it. The second thing is recognizing faces.

(Karthik at 00:17:58) So we had to figure out a way to allow users to teach faces of their friends or family members with the glasses. So once you teach the glasses that, okay, this is Joel, anytime you're in the frame, it can speak your name. We're also working on a feature where it can tell you how far you are and so on, right?

(Karthik at 00:18:16) So recognizing faces is one thing. Recognizing objects is also another aspect of it, because there are so many objects in your everyday life that visually impaired people use. So we worked a lot on improving the object recognition capability of it, and so that is another aspect. And the last part is the video calling feature. And that was actually the most surprising thing.

(Karthik at 00:18:37) Because when we built the video calling feature, we thought, okay, people already use FaceTime, they use WhatsApp. We thought people would like using this, but so far, ever since October 2020 when we shipped the first pair of glasses to this date, June 2022, there have been more than 70,000 minutes of video calling that happened through the glasses.

(Karthik at 00:18:58) So people absolutely love this feature. They make video calls every single day from the glasses. That's when we realized how crucial this whole hands-free thing was, because people use it to, for example, help them fill out their tax forms. They call a friend or a family member and help fill out forms online. Or if they're trying to go ahead and cook something, right?

(Karthik at 00:19:19) They call a friend or a family member and just ask them questions on, oh, am I adding the right amount of salt, or can you read out this thing to me? And wherever the AI cannot really help or wherever the AI fails, humans can step in and fill in the gap. And so that's what it is. And of course, there are other features as well. You can recognize currency with the glasses.

(Karthik at 00:19:38) You can detect the amount of light in the room. You can go ahead and detect colors. So there are all these other things, but primarily focused on reading, recognizing faces, objects, and the video calling.

(Joel Beasley at 00:19:50) That is so amazing, man. This is exciting. I know I'm saying a lot of words that are positive, but it just takes some time for it all to sink in, as you're talking about the amount of light in a room. Right? Yeah.

(Joel Beasley at 00:20:01) We don't think about being able to detect that on a daily basis to live our lives.

(Karthik at 00:20:07) Yeah, it's so true because, you know, you walk into a room and then you know, okay, I turned on the light in this room or I've turned off the light in this room. You can sense that so easily. But for someone who is, again, and especially for a lot of young visually impaired people, you know, they want to get out of their homes. They want to live independently. And all of these really simple things are a huge factor in making them feel more independent. If you can tell for yourself or know for yourself if you've turned off the light in a room or turned it on, or if you can read the mail that you're getting from the municipality, the government, or if you could go ahead and know if a particular object is there in front of you—if you have left your white cane on the table or if it's on the floor, or if you're looking at your particular yellow coffee mug or whatever it is, right?

(Karthik at 00:21:08) So all these small things are the ones that seem very easy for someone with sight, but then for a person with a visual impairment, it's more like, wow, okay, I've never been able to do these things, but now I can do them with artificial intelligence.

(Joel Beasley at 00:21:13) Yeah, I was watching one of the YouTube videos that a reviewer was reviewing the product, and they pressed a button to find their laptop and they scanned around the room, and it started beeping at them when they were looking at their laptop. And I thought, wow, that is so cool.

(Karthik at 00:21:30) Yeah, yeah.

(Joel Beasley at 00:21:31) How many blind people are there in the world? Do you know?

(Karthik at 00:21:34) There are about 380 million visually impaired people in the world. So they're people with some form of non-correctable visual impairment. But if you look at people who have the entire spectrum of people who are visually impaired or people with some kind of visual disability, that's around 700 million people, potential people. So that also includes, for example, people with dyslexia, aphasia, and all these other kinds of visual disabilities that are not exactly related to visual impairment, but then our technology or artificial intelligence in general or computer vision can truly help those people as well.

(Joel Beasley at 00:22:13) Do you have any ability in a similar neighborhood as seeing light in the room? Do you have any ability to read the emotions on other people's faces? I've seen some of the AI models doing that.

(Karthik at 00:22:26) Yeah, there are a lot of things that we know we can do, but we just don't want to do it because it's a very tricky place to be in. A lot of people come to us and tell us, oh, I would love to be able to tell the emotion on someone's face. And we experimented a lot with those kinds of things in the early days. And my opinion of those kinds of things is, you know, it's not there yet. It's really not there yet. And we try to focus on the problems that we can solve really, really well and that are really important for people. And then I know some of the things, for example, there are some AI models out there that claim to tell the age of a person. That's the absolute worst kind of a use case for computer vision because it's impossible.

(Karthik at 00:23:10) Today I might look like I'm 35, and tomorrow, if I lose a ton of weight and then shave my beard, might look like 25. And it doesn't really help a visually impaired person at all. So we try to steer away from models or from tech that's not there yet. And so we just say, okay, no, we don't want to do this. But yes, I'm totally aware of emotion recognition and all those things.

(Joel Beasley at 00:23:36) What's the layout detection feature?

(Karthik at 00:23:38) So the layout detection feature is basically the feature that, given a document, tells you if a document contains headings, if it contains columns. And we're also going to be extending this to work with tables, you know, because again, for someone to be able to read their own account statement without somebody else stepping in to help them, even if it's a friend or a family member—it's your private life at the end of the day, right?

(Karthik at 00:24:04) And if you're trying to read a blood report or something like that, you should be able to do that independently. And being able to recognize tables, regardless of whether they have borders or borderless—it could be any kind of table. That's something that we are working on. And my hope is that in the near future, we would be able to provide alt text to images that are there in printed documents as well.

(Karthik at 00:24:25) Because, for example, you're trying to read a magazine article and it has a picture of a person wearing glasses. Now that information is something that you can't add an alt text to. You can write a caption there, but how many people actually write a caption for an image in a printed document? You can do that for an image on Twitter or Facebook. You can provide alt text, but for printed documents, you can't, right? Similarly, we're working on interpreting mathematical equations and graphs. It's something again that, you know, a lot of kids who are visually impaired in school, especially with the whole COVID situation, they started to get these PDFs. But PDFs are some of the least accessible documents that are out there.

(Karthik at 00:25:09) Most people, when they make a PDF of a book, they make something called an image PDF. And an image PDF is just nothing but a PDF document that has an image in them, and it doesn't have any kind of alt text. It has no semantic structure, right? So that makes it really hard for kids, especially who are wanting to interpret a mathematical graph or an equation. They can't do that. So we're working on tech that can also interpret math equations and mathematical graphs and at least give them some basic context that, okay, this is a math equation and this is what the equation says, or this is a graph and this is sloping upwards, and trying to give them some context to what they see. That's what layout detection is, you know, just trying to add more intelligence to documents and books in general.

(Joel Beasley at 00:25:55) Where do you see these glasses in ten years?

(Karthik at 00:25:59) I just strongly believe that smart glasses are going to be the next wave of computing. It's something that I hope to see come to fruition. And at that point, Envision Glasses is going to be hopefully the most popular solution on smart glasses for people with any kind of visual disability, right? We're today starting off with people who are having visual impairment, non-correctable visual impairment that can't be corrected.

(Karthik at 00:26:29) But my hope is that we can take this tech and also help people with dyslexia, with aphasia. We're already starting early research in that area right now. We have some testing going on to expand our tech to work with people with other visual disabilities. And I hope that we're able to get the glasses, the tech, into the hands of almost every single visually impaired person, regardless of their background, their country, their economic situation. Just trying to get this tech out into the hands of as many people as possible.

(Joel Beasley at 00:27:01) Do you ever think that it will connect at all with the brain computer interfaces?

(Karthik at 00:27:07) I hope so. I hope.

(Joel Beasley at 00:27:09) Or is that too far off?

(Karthik at 00:27:10) Maybe the next ten years. It's hard to say, and I think there's so many cool things going on in this space. But I hope that we can reduce the level of latency that's currently there right now with wearing your glasses and looking at the camera and pointing the camera in the right direction, getting the output in audio format. And if we could just reduce that latency with some kind of a brain computer interface, that would be amazing. But I think that's probably sci-fi, probably in the future. But I hope that smart glasses, at least in the next five years, become better, more powerful, and we can try to take this tech and put that on as many types of glasses as possible.

(Joel Beasley at 00:27:53) Yes, absolutely. Do you have competitors? Are there other people doing this?

(Karthik at 00:27:58) Yeah, we do have one competitor. It's an Israeli company called OrCam. The primary way we differ from OrCam is that OrCam makes hardware and software. They basically make a mini computer that can stick to regular reading glasses, and that does some of the things that we do.

(Karthik at 00:28:17) For example, it primarily focuses also on reading and recognizing faces. So these are the two things that it can do, and the software runs completely offline. The approach that we took is very different in the sense we wanted from the beginning to only be a software company. We didn't want to go into the hardware space at all. Again, driven primarily by our belief that smart glasses will be more relevant in the future.

(Karthik at 00:28:43) So, you know, we would rather be building the software for it rather than working on the hardware and seeing ourselves becoming obsolete in five years' time. So we took a software-first approach, and our glasses software is also platform agnostic. It's running on Android, but every single piece of code that we have written is completely platform agnostic. In a sense, we primarily use C++ for all the computer vision stuff that we do. So we can always put all of that stuff onto a new platform whenever that becomes possible.

(Karthik at 00:29:11) And we also look at the Envision Glasses as a super app. We are in talks with other accessibility companies, and we are providing them access to the Envision Glasses so that they can put their applications that are really popular in the accessibility and the visually impaired community to also come on the Envision Glasses platform. So when someone buys a pair of Envision Glasses, they don't just get Envision software, they also get some of their other favorite apps also on the Envision Glasses as well.

(Karthik at 00:29:43) So it becomes something like a super app, right? Something like Line or all those other types of apps where you have multiple apps that you can pick and choose from and making it more like an App Store or Play Store kind of model. So that's the approach that we've taken versus trying to build our own hardware. It has its trade-offs, it has its pluses, but it has its minuses as well.

(Joel Beasley at 00:30:06) No, it's smart, though, and I call it the infrastructure play where, you know how Stripe becomes part of someone's business. And so if you get these other technologies to start building apps on top of yours, that just makes yours more resilient. Hey, how do you make money if you don't do subscriptions anymore?

(Karthik at 00:30:24) So primarily, one, by keeping the operational cost of the app really low and relying on the glasses as the main source of revenue. So right now, we sell the glasses. People go on the website and they can buy our glasses for $2,500. And we also are covered by insurance in a lot of states in the US. There's possibility to apply for funding and subsidies in Europe and UK and all these different places. So glasses is the primary sort of revenue generation model for the company right now, and the app is more of a way for people to sort of get accustomed to the Envision ecosystem as a whole.

(Karthik at 00:31:05) Right? It's more like a lead generation tool for us at this point. And also, it's sort of part of my core belief that, yes, we are a company. We have to go out there and make revenue to be able to build more of the things that we want to build. But whenever there's an opportunity to go ahead and offer this tech to as many people as possible, because I know that something like $2,500 might be really expensive for an average Indian, for example, or any of those kids who are studying, who are part of that high school that I went to and initially spoke about.

(Karthik at 00:31:37) Right? Those guys can't really afford the $2,500 price tag of the glasses right now. And for those people, the app is completely free, so they can just use it even on their $50 Android smartphone. Right? And we put a lot of effort into making sure that the app is available on as many devices as possible.

(Karthik at 00:31:57) In fact, anyone having an Android device that runs Android 6 and above, which I believe came out seven, eight years ago, but even those people can use the Envision app. So it's about just making the tech more accessible.

(Joel Beasley at 00:32:09) Yeah. Can Apple users use it?

(Karthik at 00:32:12) Yeah. It's available on iOS and Android, and we're also working on a desktop version as well.

(Joel Beasley at 00:32:19) So I can download this for free right now?

(Karthik at 00:32:21) Yeah. You can download it for free right now. You can just go to the Play Store or the App Store, install the app on your phone, and then just start using it.

(Joel Beasley at 00:32:29) What do you type in for the name of the app when you're searching?

(Karthik at 00:32:32) You just search for Envision AI, then it lands you directly on the app itself. You can search for it on the Play Store or the App Store, install it, and then you can start using it.

(Joel Beasley at 00:32:42) That's amazing. Alright. Well, I'm gonna install it and play with it this afternoon. We're leaving in about an hour to travel down to Florida. Gonna go back and visit some family, and I think it'll be fun to play with this and show the kids.

(Joel Beasley at 00:32:56) And yeah, so thank you for giving me an activity this afternoon.

(Karthik at 00:33:01) Yeah. Yeah. We've just wanted to ensure that the tech is there in as many hands as possible at the end of the day, and I think the app is a great platform to reach a lot of people right now. And hopefully, you know, with the glasses taking off, we'll be able to continue the development of the app for many, many more years to come, hopefully.

(Joel Beasley at 00:33:22) Thank you so much for listening. And if you found this episode useful, please share it with a friend or colleague who you think would get value from it. And if you have topics that you would like to hear discussed on the podcast, either add me on LinkedIn or send me an email, [email protected]. Every time I get an email or LinkedIn message, it absolutely makes my day and inspires me to keep going.