In the ever-evolving world of AI, there is a fascinating realm of AI companions. It is reshaping our daily lives through virtual assistants, chatbots, and more. But what impact these AI companions have on our daily interactions and emotional well-being is something really interesting to think about.

In this week's talk, Amit and Rinat talk about AI Companion, what they are, their diverse roles, ethical implications and a lot more!!!

Are you single? or maybe double? Whatever it is. Today's topic, we're going to talk about something that might be relevant to you. We're going to talk about AI companion today. And it's really interesting topic and AI companion can mean many things, whether you're single or not. That's, you know, maybe less of an issue, or less of a concern when we talk about AI companion because it's so much more than just the relationship stuff. There is a lot more to it and how the whole thing works, and how it's sort of impacting our society and the way we behave and make decisions are quite important to know about because right now people are doing it and you might know someone, or you might be someone who is interested in knowing more or experiencing this new advance on technology. So thanks Amit for coming up with this topic. AI companion. Very intriguing. I would very much like to talk to you. And I hope our audience will enjoy our conversation today. So what is AI companion?

So thanks again Rinat for that great introduction. I couldn't have done it myself that well. So, I mean, when you talk about a companion, the companion is normally a partner, you have close relationship, but a companion could be many other things that the companion could be a friend. A friend, it's a companion. A companion could be work colleagues, a companion could be someone you share your deep dark secrets with. Could be a therapist could be your psychiatrist. Could be could be many, many things could be your school teacher. Not in the sense that okay, he's a companion, he or she is a companion but in the sense that they are very close to you about a particular subject, or they teach you about that subject. So a companion can have multiple meanings. And with the advent of ChatGPT and these large language models, what's happening is that you can train these tools to behave in a certain way, like in certain personalities. And that's how AI companion is now becoming even more popular. So the reason I wanted to talk about this topic is it's already very popular. There are a lot of apps about I mean, having AI friends, dating apps where you can actually date AI instead of real person and many other things. Talk to a mental health expert, but that expert is an AI, talk to a private tutor that you there is an AI and there are many other applications and all these things. Tell that okay. This is something to stay because if a technology like this is being used in this way and people are actually using it, it means it has its own advantages, and it's popular, so I thought okay, let's talk about it. Let's discuss what are the pros and cons etc.

Yes, absolutely. It is. It is. I mean, the more I think about it, the more I sort of realize how, you know, how big it can extend the use and also the impact in our society, etc, etc. And as you correctly mentioned, it's not just the romantic relationships that you have, but also it can expand to any kind of relationship. Controversial relation, sorry, conversational relationship with any entity. And now, the more I think about it, I feel like I want to go back to the beginning. And if, you know, right now, it's becoming more and more sort of worth it, if you'd like with the conversational AI models that we have recently, like ChatGPT and the rest, but technically, we've been having AI companions for just a little bit more earlier than that. For example, Siri or Alexa or these kind of conversational sort of technologies has been there. But the main problem that these faced is that it couldn't hold up a conversation more than one response.

So you ask a question, it gives you an answer, and then it forgets the history of what just happened. And then you have to ask a new question and then so on and so on. Now that posed a obvious problem with the realistic mess of having a proper conversation like you would do with a human. Now. Nowadays, I've noticed Alexa has also started to become more kind of if we want to call it realistic, because that may not be the right thing to say. But yeah, so it is becoming that but again, it's still you know, none of those models were nowhere, nowhere near even in those cases, we have heard news and various anecdotes in social media that people were kind of, you know, sort of escaping loneliness by talking to Siri and the rest. So even then, you know, when people are lonely or going through mental health challenges, it is very difficult and they need anything that that would sort of even temporarily occupy them out of loneliness. And this is very difficult, and, you know, any help is something. And because of that, I think, at that time, even though that wasn't really sort of having conversation, it's just asking question, I'm getting an answer, yet still. It was something but now I am kind of keen to see this space and how it expands to, you know, everyday life. As it as now we have the technology to have continuous conversation with the memory of that conversation from the starting point. And that's a big deal.

It is a big deal. And it's interesting, you'd spoke about Alexa, Siri, Google Assistant. Basically, initially, they were used as Voice Search voice commands. So instead of doing a Google search through typing something on your computer, you do a voice search. So you trigger the agent through a voice command. And then you ask a question, and the question comes, they give you a response. So now they're called actually now they're called Voice assistants because they can now assist you in doing certain things like switching off the light, switching on your TV, shutting your door switching on the camera, etc, etc. So those are become voice assistants and they are also like an AI companion because as you mentioned, people find it comforting to talk to these assistants. There's actually a very old film that's called her a “HER”, and that was about a character that speaks to an AI.

I have Yeah, it is very, it was very interesting in the way how the whole thing turned out as the computer used. It's like the, you know, the power of, you know, sort of talking to or sort of generating multiple responses like 1000 times or 10s of 1000 times more, and it would get bored with just one person. Right? That was that was very interesting, and it's a good movie to watch if you are thinking about but again, it is fictional and we're you know, our current technologies nowhere near to that fictional scenario in the movie.

Well, so what happens in real life so, say I'm talking to you, okay, we are talking over zoom, but when we meet, We talk in person, so you have a face, you have a voice. So even if I close my eyes, and I listen to this podcast later, I can recognise that that's Rinat. So I can associate that voice with that person. And then you have a particular style of writing text. So you use certain words in your sentences. So that indicates that okay, this is your style of writing. So that is again an identity and AI companion can take up that personality. And these personalities can be real person. So I recently read again, in one of the articles that I was reading that TikTok influencer, what she did is she was getting a lot of requests from people to, like, start talking to her. So she thought, Okay, let's do an experiment. Let's let me train an AI model with my voice on 1000s of questions. So 1000s of questions were asked to her and she responded using her voice. Okay? Now imagine, you create a bot or chat application, okay? And you ask people to register to this chat application and you get a private session with this TikTok influencer and you talk to that person in their voice. So you ask a question, and you get a response back in a WhatsApp voice message, not a chat, but a WhatsApp voice message. And then it feels even more personal. So now it's not about text. It's like in the movie Her that now it's more personal. So now, you connect more because now you have the voice. So you can keep listening to that voice. And that's her. That's the next level. And

you know what? Yeah, absolutely. It's happening right now. And yeah, how much it will expand that. That's for us to see. Now, what you were what you were just saying right now maybe think of this thing like how many senses do we have five senses, and whenever we experience something, it is ultimately these five senses that we're using. So now when we see a person in person, there are you know we can I mean obviously would be kind of weird, but we can smell them or we can touch them. These are the two things that happens differently as opposed to meeting them. For example, in a Zoom meeting, like we are right now. I mean, apart from these two senses, the other three senses are fully, fully okay and with better quality, etc, etc. The experiences as if we were I mean, they were nothing sort of restricting us to communicate in different ways. I mean, you know, they're even the nonverbal communication for example, hand gestures or you know, with really good connection and really good sort of setup you won't even know whether if I'm nervous and if I'm breathing heavily or you know, my, even the subtle changes in body language or expressions etc. So apart from smell, which is not really part of regular conversation he'll do with everyone and touch which is also not really a conversation point. It is to some degree to really close companions, but not always. The rest of all of the experiences can actually now be I mean, you mentioned audio, right? audio and text. These two are already very much but video is also underway. I mean, right now you can create videos. Right now you can create sort of create a professional image of yourself by just uploading a regular image of yourself. It could also create a 3d model of your head by uploading your few images from different angles. So the video part and right now on kind of real time video AI update is not there yet, as far as I'm aware, but manipulating videos with AI is also available in potentially with GPT-4 I think you can kind of Yes, yeah. So

So there are there are other tools like RunwayML and they have Gen2 model and that can create a video from text prompt or takes an input video and then using your text prompt, it edits that video in that style. So suppose you want a style of a video so you upload a style and then it will create the video from your text prompt in that style. And what you mentioned about the people now having a what we call it a digital avatar, right? So that digital avatar can be created down because you can scan yourself so you can scan yourself using a tool. The whole face from all 360 degree and even your body if you want and you can then upload it on the internet and you can then give your voice to it. And now if someone wants to talk to you, they can talk to your digital avatar we discussed in the last podcast about digital immortality, like how a digital version is created of a person who's going to pass away so that their relatives who are grieving can still be in touch with these people and imagine now this can be replicated. So currently what's happening is instead of a real avatar, like my holographic image or my complete self, it's a digital cartoonish kind of Avatar, so like you have on the metaverse in the meta universe, some horizon worlds. So in that you have a digital avatar of Mark Zuckerberg, but that's not like the real person. It's like a cartoon character. So those cartoon characters are currently there on dating apps and you can actually date these avatars.

Yeah, yeah, exactly. Yeah, I mean, AI generated video content is already there. The only thing that isn't there, although I don't know for sure it might just be there from yesterday or tomorrow. The only thing that isn't there is real time AI generated video content. So right now I can't sort of just type and my AI avatar would be with you in a zoom call. I don't know. If that's possible just yet. But once that is possible, then that means digital communication with a sort of AI being is very much possible to a degree that you are happy to do. With a human like you know, we are talking right now at work. Most of us work from home nowadays. And you know, many other things that you know, we are doing our from as long as if there is video there it just complete the full circle, but even without the video, text and audio is already there, and a lot of people are sort of using it to have an AI companion.

Absolutely. And I think I mean, with the AI companion we have to see that. People are getting comfortable. As you mentioned people are getting comfortable talking to this AI companion, more so than a human being. So a lot of there are a lot of introverts and especially people now who are so used to technology. They are more comfortable talking or sitting in front of a computer than actually sitting in front of a person. And that creates a whole different social dynamics. And these people find more comforting and there are robots are actually so there are like robot dolls that you can get so that I mean it might be socially awkward to take those robots out. But at your home, you can actually have a robot with you who can maybe do the household duties, who can be your companion who can listen to you who can respond to you empathetically, etc. And these AI companions that we are talking about I mean we are focusing on dating. But let's say example you have a mental health issue. And you don't want to talk about it with real person because you think that they will judge you. So you go to an app and then you start talking about your issues and App then responds in a very empathetic way. And then you've you find that comforting because it's very different from say a human being giving a response because humans get tired humans are erotic. They're not all the time rational. They also have emotions, but with the AI you can control all of these things and they can give you a response that more suitable for your maybe personality and your condition. And that helps a lot of people and they feel more connected to that person. I think that's the same thing happened with the film HER. So in the film HER, The same thing happened the person felt more close to the AI than actual person and they started missing the AI

Yes, yes, absolutely. This is again, I mean, if you guys are following along our podcasts in previous episodes, I have always been sort of clarifying it over and over again, is that chat GPT and all that other sort of conversational AI is and not they don't have general intelligence, so they don't actually understand you and all of those things and I kind of thought was trying to make a point that look, these are not real. These are sort of just mimicking human conversation, but yet, even knowing that this is the case, I think I would still be sort of inclined to learn, say, for example, a new skill. I don't know, programming languages very well. And I could ask ChatGPT In fact, I was sort of getting help during a few projects that I was doing. And, you know, it was more kind of natural to me to talk to it as a conversation. And I think it also responds better because that's what it's trained on as well. But I can totally see that it doesn't even have to be real or that part that we think that this is a fundamental requirement of having a meaningful conversation That the other person has to understand my feeling. That's not always the case.

This is this is where we might go a little bit off topic, but let's just think about when you're having a conversation with a human, you're thinking you're conveying a particular set of emotion and an idea, but are you sure that they're receiving it that way? I mean, it's a very, very simple, simple sort of test that you can do. I mean, you know, write down like three words like sometimes, often, seldom and there could be 10 words, similar words, and tell them to write what percentage of things occurring is sometimes, you know, you might say it's 50%. You might say seldom is 10%, and you might say often is 90%. Or 80%. But and tell them to write down the same thing and you will see the difference and you will really make you think that anything that I've added in you know, they are being received the way the same emotions I'm telling the person the human person, and then when they're saying they're also thinking about how they felt but without you experiencing their sort of what happened during their day. There is no real way for you to fully understand the way they want you to understand. So at the end of the day, when you are talking you're really just talking to yourself. And whether the other person is an AI or a human, as long as they're responding in a way that guides you towards your goals, then it's actually less important not saying it's not important at all, it is important it is important for us to know that the other person can sort of understand my feelings why I've done this thing as opposed to the other thing. And they can relate to a degree it's not completely going nowhere it is being received and the general idea is being understood. So it is important, but it's not a fundamental requirement as much as we think it is. And that's what I want you guys to sort of ponder about. After you listen to us today because it is an interesting way of looking at our traction.

That's a profound way even I was taken aback with that thought like you're talking to yourself. I think when we pray, and I'm not a believer in God. So when you pray, you actually close your eyes and you talk in your subconscious mind and you pray for something. So normally, if I look at it in a way that okay, if there is no God, who am I praying? Who am I talking to? And I came to the conclusion and by the way, disclaimer, this is purely my opinion. I'm not forcing my opinions or beliefs on anyone. You're free to pray to any GOD. It's just my belief and my opinions. So please don't take it any other way. So, when I pray and I talk to when I used to pray, and I used to pray to a God, I realised that actually I am just talking in my head, and only I can listen to it. So basically, I'm talking to myself, as you rightly mentioned, that I'm just talking to myself and I'm calming myself saying that things will be fine. Things will be okay. You will get this, you will get that. you're just convincing your mind. It's like positive thoughts, reinforcement, which now makes a lot a lot of mental health experts. They tell that okay, don't worry, be calm. Don't be anxious. So it's a way of calming and that's why meditation is picked up because meditation is also in your head just you talking to yourself or just focusing on one single thought it could be breathing, it could be a game, it could be just one single thought you just focus on it you calm your mind. You relax yourself, and then you just focus on it.

So I think AI companion. Yes, you're right. It doesn't have to understand what you're actually saying. It just has to give a good enough response for you to think that it understands you. And that's enough and that's why I think ChatGPT has become so powerful because it is good enough for most of the people to interact with thinking that okay, it is there to help. But it's not it's just a tool, and it's just randomly blabbering words if you look at them, courting them, and then it's not Algorithm, the training model and add it and you think that it's intelligent when it's not. So you always have to do fact checking, but it's good enough to help you understand difficult concepts like a virtual tutor. So instead of now paying someone a lot of money to get coaching in a programming language or in mathematics or physics or chemistry or biology, history, geography, you can now have a virtual tutor. That tutor is a companion now they are a learning companion and you are learning with that person. Instead of watching a video on YouTube on which you can't ask questions. Now you can say that okay, let's watch this video together and I have these questions. Can you answer that for me?

So that's an AI companion as a virtual tutor. Same thing when a mental health expert, I feel lonely, I feel depressed. What do I do? How do I get back? How do I live my life? What is what's the purpose of my life, etc? That person, that companion, that expert will say, Okay, do this, do this. And then same with the dating app, same with many other application virtual assistants. So I want you to understand that okay, this is where I'm coming from. I want you to set up a meeting do this, do that. So the virtual assistant understand, and these are all very, very powerful applications. And, of course, one thing that we haven't spoken yet is the ethical aspects to it. So now you are actually talking over the two model that is there on a server, and whatever you're seeing is getting recorded. So your deep dark secrets, your private emotions, your thoughts, they're all getting recorded on a server, even if they say that it's not getting recorded, it is getting recorded, for sure. Because no one can say that it's not getting recorded.

I mean they have to, I mean for the for the system, they have to record the right guidance. they might have chatted anonymously, but you know, Anonymous is as good as however careful.

Yeah, exactly. So even ChatGPT records, everything that you say to it. So that's why it is very important that you anonymize the data that you are putting it so you don't put personally identifiable information. Maybe we'll talk we'll do a talk on this as well bi. I think that's a really interesting topic. But you should make sure that you're not talking about a lot of personal things that can identify you as a person like Amit Sarkar is talking about religion and non believer, and I can identify that person by what they're saying, or to ChatGPT or by asking what questions.

But here's the thing. I mean, if someone is sort of considering an AI companion, and they might I mean, as we mentioned, at the beginning of the target companion could be anyone you know, may be a virtual assistant or a teacher, but it could be also very close personal, personal relationships like a therapist or a girlfriend or Boyfriend. A romantic companion. And when you're kind of having these types of relationship like a therapist or a partner, romantic partner, then you don't really want to I mean that the whole point is with these people, you don't have to think about repercussions you can just open yourself fully up. Now judgments, Yeah, no judgments and then the most safest place you know your partner and your therapist. These are kind of you know, thought to be and then if you now have to sort of be cautious about what I can or can't say, then that does take away from the experience though, doesn't it? Even the however realistic their responses are it does, it does kind of

Yes. I think you are right. If you can't ask like a real person, like I'm talking to you, and if I can't confide in you, or the AI companion in a way that more real, and every time I have to think what I'm seeing then yeah, it creates it defeats the purpose of the companion. Then said you're just treating it as a tool.

Exactly. So now this is actually quite important and relevant now. And I think we should do another episode on this as well, which is how to create your own ChatGPT. So this is something I think we should we should talk about in a later episode, because nowadays there are free tools available in GitHub and I think I suppose I could be wrong, but Facebook has also released a sort of AI model that you can basically install it in your computer and then sort of train it with various agent. Now the first question anyone who understands a little bit about AI, they would say that oh my god, my personal computer doesn't have enough spec to run an AI model. You know, you would need a graphics card this much and it wouldn't be possible in person, but that submit I'm here to tell him in this episode, but we'll go into it in more detail in later episodes. When you know, I've kind of done it myself. I have that in my task list to do. But that's a missed you can come to like as close to only 1% gap, you could come to 99% Close to ChatGPT level responses with using these free AI models that you can use it in your computer in your personal computer or laptop.

And the level of closeness it will read is you know the difference becomes negligible even if it's not one person even if it's 10% difference in you have 90% capability of an AI model. And then you can teach and you can you can sort of train it with the data that you would like to be trained on so you trained it generic everything first. And then even give it specific data that you've collected about yourself, maybe your chat history, maybe your email history or everything about you. So then it knows you a lot more so it could also mimic a different persona that you know we discussed earlier today. So it is very much powerful and I think that might safeguard you to keep in your data within your personal computing space. And that's why I thought it's important that we bring it up now. So you guys know that there is a way and you know, it is a little bit tacky. It's not impossible. There are YouTube videos with step by step guide. But yeah, something that is possible and available right now. But we might want to do another episode on this on its own so

So the only thing or only drawback which I see with that is that you can't take it's not portable so you always have to have your laptop with you. But most of them they carry a smartphone.

That is true. I don't think it's smartphones are powerful enough to do that.

But smartphones are not powerful enough. But smartphones are there so that you can get an app which has these AI companions and can use those apps login whenever you want. Say you are in a meeting say you are in your community. Say you are in a cinema hall say you are with a friend. So you're at a cafe watching or hiking. So you have the power wherever you have internet and whenever you have your smartphone with you. You can talk to that app instantaneously without like worrying about whether I have my laptop, and I think that's the power right? I mean, we can do so many things on our computer, the laptop, everything but the smartphone is the single most device that the person interacts with the most in a day. I mean even if you have computers you interact more with a smartphone than anything because it's portable. I take my smartphone everywhere. I mean, I don't carry my wallet anymore. I just use my Smartphone. So it's that powerful. And it's like a house key now so I just carry it everywhere. I can forget my keys but I don't forget my smartphone.

And Nowadays with Smart Lock,

Yeah, smart lock. I don't trust that technology yet. But, yeah, that is true. So I think that's the important bit that I think differentiates between a desktop application or a laptop application to smartphone app. And that's where the I think that's why a lot of companies they push for an app because it's just easy and people can interact with even if they're in the bed. I mean I don't want to take my heavy laptop and on my bed and then type on it sitting I can just lie down on my bed and then using my fingers. I can just quickly type something in the app and get a response. So I think that's a very it's a very important differentiator. And that's why these things are getting popular now. Because these models, they're all Hosted on Cloud servers. And you can interact them using an app. And it's so easily available and you can modify them, as we discussed in the beginning that you can modify them different personas. I mean, right now in ChatGPT, if you want to do a proper prompt, you tell them that okay, you are an AI expert, or you are a computer expert, or you are say you're applying for a job. So you say that you're a technique. You're a technical recruiter, and you and ask me questions for a particular subject. So you give them the subject and they'll ask you questions and then you respond to them. So they are a technical recruiter. Then you say you are a Java experts, you ask them a Java specific programming problem and they can ask reply in that particular with that particular knowledge. If you just ask if you just give them a simple question and a problem, they will respond you but not as that persona. So that persona is very important to define. If you want a specific response or very good response from ChatGPT.

Yes, absolutely. And you know, then we delve into more into how to promote it better in a way that you get the right kind of and we were you know, I would tell you guys to sort of go check out our prompt engineering episode because we've covered a lot more and also some tips on how to prompt better to not just chat GPT about any of these conversational AI models, which I might think that which I think that it might be really helpful in nowadays, even if you're, you know, using it for job or just for fun. That episode we cover a lot in terms of how to prompt better.

Yeah, and I think I mean if we come back to AI companion, it's not about prompting now. So it's not like you have a specific persona in mind. You are just talking to AI. Normally, you are not putting any context you're just saying okay, I just want to talk about this today. I want some help in that aspect, like we are talking on this podcast. So something similar. So you don't define any personas. You don't define anything. It just, it's actually pre built. The persona is pre built in the tool that you're using.

It actually reminds me something, sorry to interrupt you Amit. This actually reminds me of something a very like tiny thing that I've experienced. So a few days ago I was kind of I had a sort of a coding problem and I was sort of taking the help of ChatGPT asking it. I want to do this how can I do this please help me with the code. And then when it wasn't working, so it was actually quite late and I've been doing that the whole day and it was like three o'clock in the morning and I was you know already kind of you know, quite impatient, etc. And then after one of the questions it kind of responded with you have to, you know, gave me a list of things that I have to do in order to sort of get to the next stage of my project. And then instinctively I had this response saying that I'm tired, right? Like I didn't feel like doing all of those things and I couldn't even feel like telling it that or break it down in smaller chunks or whatever I just wrote. I'm tired. And this is the thing this is the best part of this kind of language model which I have not a specific, you know, specifically skilled So straightaway it responded with all the things that I can do to make me feel better. I could go for a walk I could have a drink of water, and not just that way else I could go to help who I can call and all of these things which are actually quite helpful and someone what I can imagine someone with mental health issues and loneliness and other kinds of sort of need to converse. They would be so massively helped by such a tool that it is amazing and to be able to sort of tailor it to the persona that you would like to converse with. That is a huge ability and I am quite positive about how this technology will expand.

Very interesting because my wife did a similar thing. She asked a question that typically responded and then she said thank you. You thank someone right in the in the real world. So she responded Thank you instinctively and ChatGPT then responded Oh’ it's a pleasure….This/that….. So it was it was really nice like okay, ChatGPT understand what you're saying. I mean, of course it says it taking thank you as important trying to see what are the best output based on the training data.

It's not that relevant whether it understands the exactly good enough. Yeah, yeah. It's yeah, I would say it's even more than good enough because for you as a as the human part of the conversation. You understood, and you got the response. And it is up to you and this this reminds me of a really sort of powerful quote that I really liked very much is that you can't do anything that happens to you. You cannot sorry. You can't control what happens to you but you can control how you respond to it. So I mean, you it is in your head, you can decide whether this response is meaningful or not. And it doesn't have to be like thinking feeling, understanding human on the other side of the conversation for it to be meaningful. I mean, a lot of these responses that it generates for work for whatever, they are quite meaningful responses. And yeah, I mean, I would say that I very much sort of for the progression of this and looking forward to what it brings.

Yes, and I think I'm also looking forward to what the technology unfolds into because I mean, there are I think many other applications that we haven't even thought about, like, an expert. I mean, like a debugging experts say you're an engineer in outer space and you have to fix something. You don't have a human being you have an AI companion who can guide you on how to fix things in outer space. Maybe deep submarine. I don't know. I mean, you're on a hike and you need some navigation help, you don't know how to read a map. So

these are some of the other use cases that I actually even didn't even think about lonely.

I am just taking it out loud.

As when you're setting sail off to the seas and when you are in space. That is a big issue people goes in can go insane to just be by themselves and I when I heard the first time they international spaces, space station is one person every six months and that is it. They just have to be by themselves for six months, which is quite crazy to think about. And

yeah, they do a lot of experiments with they keep themselves mentally busy, but you're right. I mean, not talking to someone and just being by yourself for six months. Well you were I mean we all were right in a similar situation during COVID.

That is a good point. Yeah, but

It resulted in a lot of mental health issues for a lot of people because they were lonely. They were depressed. They were not able to talk to anyone. Because it was not great for a lot of people. So yeah, I understand where you're coming from, but yes again. So these AI companions could be very useful for these long journeys, long trips. And not just like, just companion to like for comfort purposes but also like an expert like as I mentioned navigation, solving issues, mechanical problems. So instead of now having called support team for any issues, you have a companion app, and that companion app can solve all the problems for you. So yeah,

yeah, it is fascinating and intriguing. Yeah, but, but fun. It was quite fun to talk to you about this and yeah, hopefully our audience enjoyed it as well. And we look forward to getting your guys's feedback and get in touch and let us know what else you'd like us to cover. But yeah, this was this was a good conversation. Amit, I very much enjoyed it.

Thank you so much Rinat. It was really good. And see you guys next week.

Yeah, see you guys. Bye

bye.