PyTorch Community Voices: Interview with Refik Anadol Studio (Full Video)
Hi, my name is Suraj and I’m a developer advocate at PyTorch. My name is Justin and I’m also a developer advocate at PyTorch. And today we are excited to be speaking with Rafik and Christian from the Rafik and Adoles studio. Welcome guys, so good to have you here. Hello, take very much. How are you? Thanks very much for having us. Yeah, thanks for being here. Rafik, you and your team are award-winning computational artists and what I find so fascinating about your work is that you use data, AI and digital technologies to create something that is very uniquely human. Tell us a little more about your artistic background and maybe things that have influenced you to do what you do. Yes, happy the first of all, so great to be here with Christian together with you because I think we as a studio on the shoulders of many wonderful things like your incredible libraries, approach, and research altogether. So we are grateful first of all for the community. And so I’m a media artist at my journey almost 14 years ago. As an undergrad, I guess studying photography, videography, web design, 3D graphics, computer graphics, eventually visual programming language. I think my four years at my undergrad in Istanbul, Turkey where I born was a pretty much open-minded process of like how to use creativity in any context. 2008, and a wonderful professor from auto university, came to our school to teach us a program called Pure Data, which was like a visual programming language for mostly musicians or people who are practicing with sensors as like an instrument. It was very first time I saw how actually a sensor that is completely invisible can transform the world around us into a visible patterns. And I think that’s where in that mini class, I coined the term most likely data painting. The idea was transformed that sensor data into a visual language of super minimal like points and lines. And I think that obsession became and grow a studio practice, then came to Los Angeles for working with incredible pioneers in the media art scene, specifically KZRYAS, who invented processing with Ben Fry. And I think started the open source programming language that allowed any creators who wants to work with code to imagine. So I was very fortunate 10 years ago, start working with him, learn how to the program from a Java site. But before that, I am using a software called the VVVV, which is another visual programming language that allows you to make many amazing things. I think my mind is more inspired by the visual programming language. And then I found my limitations quickly, and I was always inspired by the idea of make art for anyone in the world. So never inspired by solo, egocentric research of an art, but more like how we can collectively imagine beyond who we are and how we can make art for anyone in the world. Truly, truly anyone in the world. That is a huge responsibility and a research. So decided to became a studio practice eight years ago, right after the MFA graduation from UCLA design media arts and start the journey. Now we are 15 people can speak 15 language and represent 10 countries. And very fortunate, one more checkpoints we start our work as a public art. So not for galleries, not for museum, not for private, but for people that anyone in the world can explore it. Now I’m happy to say that we have 16 permanent pieces around the world. And we traveled more than 64 steves with our works, free and open. And then we practice later in the museum and the private world. So we always first public first studio. I’ve seen you mention before that sci-fi has had an influence on your art. Yes, may I share my screen maybe an image that really resonates a lot? Yes, it is. I was eight years old super fortunate moment happened. And back in time, as you may remember, VHS cassettes and the movie the ballet runner. And I was eight years old one day randomly, completely randomly. My mother is always like brought home this, you know, movies and sci-fi things. And that was the weekend I remember super inspired by this movie. And I as a child, I don’t believe you see anything this opian, but it opian. I think that that that was the beauty of the future completely embedded in the mind in a way that like how the future of Los Angeles or real. I don’t know like imagination of the cities in the future. I think became an evidence in my mind, but I think very much. It’s time has a huge influence like it’s a beautiful city where the left and right, you know, west and east past and future and where the truly the two continents connect and honestly. And that idea and the future and learning like you know programming language very early, I believe that journey like all connected very early ages. I have a question that’s kind of like a twister, but if you could go back in time, like maybe 10,000 years ago and make a cave painting to confuse everyone in the future. What would you draw? Oh, it’s a very high-talent question, but I may I show when I talking or current work, maybe as a studio so that people have a some visual correlation. I would like to share our mini show real if you can see it while I’m talking because I believe this body of work you are seeing here is our cave paintings, I guess, for the future. So I mean, as I mentioned to me, what was the three inspiration was the idea of data is not numbers is not just, you know, mathematical representation of things, but also a form of a memory. And I think if you think about a memory, which is a very fundamental, I guess, our organic, the basic model of our memories and the consciousness, I found that actually we can completely see the data as a form of memory and that memory can take any shape and form. And I don’t believe that our work is not necessarily relying on a physical reality, but also it can use light as a material. And I do find that in our research that I mean light is one of the most divine material or physically scientifically, like, you know, we can talk about light in many different contexts, but I found that actually light makes invisible visible. And the algorithms that are allowing us to like imagine to see beyond what we can imagine became our I think I guess cave paintings. 2016, I would like to say that an important start point for this imagination is our residency at the Google AMI department again, thanks to blaze, Agura Arcas and Mike Tyka and Karen McDowell. I think six, seven years is like a six to seven years in AI research as we all know, I think that was a very incredible moment because what was amazing in these workshops in this, you know, residency, I was like imagining maybe a shaman and an artist and an AI engineer together imagining things that are beyond just, you know, scientific and mathematical reality, but also as a spiritual context in life. So I found that our work that is emotionally creating this resonation is probably coming from that that belief of Jack supposing many disciplines together. Yeah, it’s very interesting you say that because many cultures do worship light for the same reasons that it illuminates what is not and what is not visible or what cannot be directly experienced as is. And I also read somewhere how you mentioned your art is about revealing that what is unseen yes and you have a series called data pigments which I thought is very interesting because you look at data as memories you look at them as pigments but what they really hold are nuggets of information. Yes, and you also said things like buildings can dream and colors may be heard and some can be seen. So I’m interested to learn how do how do you use AI over here. Yes, so I mean, Chris also with us here, I’m sure he will share amazing also his personal journey in the studio, but I will say so first of all the early days of data pigmentation, which I’m calling it was very much obviously focusing on environmental data, which again, great for case areas and incredible mentorship 2012 I start plotting very early win data through this open source sensors. And then there was an amazing learning about like how to think about information comes from a sensor about sensing the nature right the weather data humidity temperature, win gust and direction. So that was a very early studies and then later on we with with the AI it completely change the imagination and I want to go to 2016 because that project to be honest in my mind and in our research is one of the most important one maybe we can quickly go that one. First of all, the project is heavily inspired by this incredible book by Argentinian writer, Borgas and I was completely fascinated when I read this story way before what AI means, but the idea of like one day, hopefully creating this library of everything. And it’s a dream that one day every single data in the world is accessible to anyone in the world and so this was a story that I think inspired so 2016 we start researching about the archives and thanks to was of curtain who is a well known curator who challenge artists like you know see beyond what we can. And that’s the idea of like you know turning the information into a knowledge in front of a library in a live library where human watches AI like reconstructing the sorting the images and you know giving a meaning to this 1.7 million documents. So the obsession with I think archives started with this project and I do believe that this has a very significant importance in the AI arts as far as I know 2016 this is what an only first piece publicly available that allows you to like you know interact with this 1.7 million documents at yes any data universe in real time or also in in the VR as well so not necessarily has to be a physical existential thinking. But I will say this project triggered another idea called machine hallucinations which was the idea of if a machine can learn can a dream. I mean any gamers I’m pretty confident like talking about or anyone loving science fiction like political I really give some narratives. But to me what was amazing on the left side we are seeing a DC again. And result again 2016 is a very early days for you know also in good fellows incredible you know unveiling the idea to me this was the moment that the another like research started like how can we take this what AI can learn from in this case 1.7 million documents but how can we turn it into like a data pigmentation. So we start like using fluid dynamics specifically on top of this machine decisions or the latent walks and last I will say 6 years we train and again Christian can explain more but we have more than 3 billion images and and train more than 100 AI models and again thanks and we’re the friends that I think also join the support process. We were able to like work with PG and style again to a D.A. tree and of course recent studies with Dali to an open and a wonderful research but what I want to say is in this research we learned that from the culture nature space urban and we look at. Idea and the archive of the humanity is collective memories like as you may see it’s all about like seeing this new world and have to say that this research unfold into a whole new body of work and last 6 years as you may see here we explore all these outcomes in a very different and try to be. Corrector is unique to look at this patterns of how machine you know can represent the dreams. When you when you use data as your raw material I think it really helps to use AI models which can. I actually want to talk to you about the machine hallucinations piece as well because it almost seems very cynetic and it’s nature like you’re taking your taking information and one domain and you’re using models to represent them for humans to experience them in a completely different way. What emotions or what responses are you trying to invoke with such arts or such pieces of art. I think the the main narrative I’m seeing extreme things so first of all I think like money right if you look at it’s a story of art it’s always like happen right in many different you know centuries and decades but if you look at money as a like an example like how he inspired from the atmosphere right and how he was depicting like the you know the unseen atmospheric world or landscapes. I don’t believe it’s so different to think that in this case like we have a thinking brush and I think this brush truly we can dip into the mind of a machine this art feel mind of a machine and really paint with that you know thinking brush. And as you can see here even though they look maybe similar algorithms as you may feel that the outcome is extremely different and I think like this question of how you know machines could simulate unconscious and you know subconscious events like such as dreaming like. Remembering and hallucinating that became like a fundamental feeling and I believe that these pieces are kind of creating that in different context. And of course let’s also remember that we are just watching a two dimensional representation what gets much inspiring to me is when we take this research and bring into architectural domain and that’s I think where we start to see what happens if architecture goes beyond concrete still all glass. What happens if you like the balls which has barely you know in a way to function than just you know new Tony like or like some you know fundamental like physical reality of a building how can we go beyond that what we see is I think when we connect the art architecture and neuroscience art AI and technology like when they all combine I feel like that’s where the and so and deputy happens. It all no sounds like your perhaps using here to and throw more fies what are non-living objects really is a perfect moment because I feel like I love this so much what you mentioned because I think what we do is finding human in non human. Is that how your workers interpreted generally or has any have that I mean your art is so it strikes so many different synapses that I wouldn’t be surprised to learn of any interpretations of this that were completely unexpected for you. I mean then I have a very great example so first of all maybe also Christian can also explain our process little bit but what is really inspiring to me is once we like diving to this massive archives ideally we always learn the latent space right like at the end is all amazing archive of images when they are clean and process which Christian love to share I’m sure the process we always land in this latent space right we always find our self it in this you know and I’m not. You know and I mentioned the environments where automatically represented but for the public you know for anyone on the like you know public space they have pretty much like a no context right unless we have an experience or a question so that’s I think where the public are triggers. I’m pretty sure you want to like talk about like our process because it’s really also inspiring to me like how we start and how we like you know go through amazing algorithms if you want Christian. Yeah definitely well hi by the way my name is Christian I’m a lead data scientist that we’re thinking of all studio I’ve been working with repeat now I think since 2017 or 2018 kind of right when all of the machine hallucinations research kicked off I actually. I actually began as a data excavator you know I can say very very proudly that we put together some great pipelines and since 2017 of downloaded nearly 3 billion images. You know it’s an amazing process to work with all of these images right because we go when we think about doing a project like I’m pretty sure New York City was the first project I worked on and when we go and we and we collect these images from public resources you know the intention isn’t to do anything. Crazy right but it’s rather collect all of the collective memories from all of the people who have gone and experienced New York I think it’s kind of why it feels so relatable to people even when you know they’re all computer generated which is you know what we’re really seeing or. The visions of the collective memories of everyone who’s ever been in New York and taking a picture. I think it’s very very powerful so Christian is an amazing example because for us this article installation in New York I think. In 2018 as a research and 19 as an output I felt like it was a very rewarding experiment because I’ve been calling it experiments by the way I don’t think these pieces are like finished or say like that’s it because you know the data is alive and see the live and. There is no way to like say this frozen reality of a modular represents you know anything but just an experiment so what was it really inspiring to us is like as you remember like this last you know 76 years of like. Data pigmentation or fluid dynamic research on the neural networks and became here physically visible like what we felt in this exhibition by the way people staying for five hours. Made it 18 like representing some beautiful responses all like recently last year we were able to take another building in Istanbul 120 years old building but instead of using archaeology like a sort of like a. Restoring the building we thought with you know damaging the history we thought that just adding you know bunch of projectors we can project the memories of the building which is an old theater and we even like you know do this six dimensional you may study is you know from the sound to image and text is really our input for our I. Research and we always find this incredible you know joy in how we can represent this you know six seven dimensional plottings in different forms and I think like even like here maybe Christian you can also talk about like we export every single movie. We are seeing like we look at this data in a way that you know every single phase in 150 movies ever played in this historic building like somehow I feel like this you know feeling the past living now but what we feature may happen right that kind of a three three three experiments in a given time happens in this context. Absolutely I mean what’s been really exciting for me personally to work on with all of these projects is like how do you take some complex abstract. You know mathematical machine learning models and apply them to things that are realistic and that people you know feel and understand and then show the results of these amazing amazing technological advance we have to people in like a. Communicable and understandable way it’s really really interesting and by the way we are also when there’s a lot I think is what I would like to say is presenting the technology in a way that’s not only just you know it’s on context but for example this piece I believe is a emotionally important right we have around 75 million flower photos and from the cement zone archives of 16000 speeches ever so far apparently. So we’re going to find in this model like yes we had a beautiful you know a visual representation of an AI can you know could we construct any flowers but also we thought that what will happen with the sound of the forest right there’s a lot of incredible open source. And what will happen if we can also the sense of the scent of this AI and for example for this one we even created this special scent in gin in collaboration with firm in it and we were able to let their AI called Charlie trade on half million cent molecules and dream a scent called rain forest and even project on this room. Oh. That’s a multi century experience. Exactly and I do believe that so so so this specific research like we start in the studio outreach and amazing scientific research but at the end land in a free public art where any generation and culture background can open this door and feel this. You know unique unique experience where AI construct the sound and the image and the scent as well. So to me this is an exciting journey how we can push the boundaries. Yeah so one of the things that I’ve just some of the thoughts that come up in my head and you sort of lightly touched Christian on the process and collecting the data and visualizing these abstract concepts. I’m really curious to know like what is your creative process like what went into like you know deciding to build like this this experiment on like what like how is that process look like what do you what do you do are you is there an idea that you’re trying to do and then you like see what data you want to collect and then you like build these models is there and like you have any expectations because some of these these things are really intricate and really interesting and it’s like how did you. Like how did that come to that like I’m really curious to know about this process great question I think we have many amazing collaborators across institutions nonprofits you know amazing you know engineers and researchers and like you know we are really on by us about collaborating you know and then and that’s I think the trigger point like opening that first dialogue and very open on us like what is important for them for us like what is the common ground what is not seen what it says be. It’s been done before you know what is the amount of push and pull you know that very basics but then I think we always lend in this challenge situation and we always find out like okay like now we are downloading I don’t know 300 million nature for us can we get all the thing I like types in the world or you know or then or Stockholm as the collaboration with Stockholm still for our sale the mayor of sales says can you please use the sale as an input and then or like we always get this challenge is like seeing how to represent the city of sale in you know 2020’s and of course yeah Christian and then you know and we dive into this the past memories of us that we are living behind us and then you know clean them as good as possible but very important thing is we spend so much time in curating data so I think there’s a very fundamental particularly we are approaching everything but first of all we never show what is real that’s like interesting context but you know based on this image archives and then we also like have amazing collaborations with institutions such as like a mama or you know hospitals, SpaceX like we donate very significant amount of support to our partners so recently unicef like we don’t just you know create art for us or private but also like you know give back to communities so our creative model is really challenging for us to produce hopefully easy to understand and an approach and educate and inspire and hopefully give somehow meaningful back to the community if we can so as part of your your artistic process do you let the art to sort of unfold like based upon just the way how the models and data is trained or like is there some sort of like oh I’d like it to take shape in this type of way is there some way to nudge it and influence it in that manner so Christian I think it’s a great time to maybe explore like for example so this data universes we are showing a lot and very very enjoying Mozart, Rumi, Zahadid like a mama archive like it’s really you know any image sound and text archive Renaissance right you know the older writings and images and sorry paintings and sculptures so we always like start with this context but then find these new forms of meaning like there is no single way for any human I guess like to remember like millions of images but you know they just somehow represented in our mind right so is you map or lower dimension reduction pipelines to me is one of the most amazing sculpting process maybe Christian you can mention the tools we are using please yeah definitely I mean so from like a from a technical side of things you know one thing that I think that we’ve found a lot of success in doing is actually you know mostly unsupervised learning yes it’s it’s a lot of I feel like any time we take on a project and any time we start researching a new subject what it really is to me more than anything else is data exploration right you know of course we collect the images but it’s way more about what we can learn from the data rather than we can extract from the data you know of course there’s like curating here and there you know if we have way too many pictures of one subject and it leads to an over fit model you know we need to slow down the certain numbers to get good models of course things like that right but definitely you know when we start talking about collecting you know the collective memories of the location or a species of flower or anything like that again you know it really is much more about what we can learn from the data rather than we tell the data and that’s that’s a really inspiring thing so much more exploration than actually actually it’s a great world because unsupervised is our next name of the exhibition at MoMA specifically we learned that this word has a much positive resonance in in understanding the machine process and and I think it’s really interesting to think that even though like there are so many cutting edge algorithms maybe can do incredible like image recognition and and so on but still giving a chance of a machine to co create it is so fun and try to find the absolute truth in any like you know archive and I think that that letting letting go creativity and lowering the barrier of like egocentric thinking versus like how machine can become an extension of mind is more inspiring than you know stuck in just you know certain truth of an algorithm versus like you know even for again trainings right like we are intentionally sometimes you know extreme experimenting with this learning rates and try to be more abstract like not necessarily try to mimic the reality but again it all starts with the data as Christian mentioned at the end the ground truth which is the memory is like you know the inputs it’s actually it’s actually really funny because for the unsupervised project specifically we had a great set of classifiers for this MoMA data and we spent quite a bit of time actually training supervised models and supervised dementia reduction and then through all of it away it’s interesting earlier you mentioned that you know you refer to these as experiments and that they’re constantly evolving and yeah if you’re doing unsupervised learning and you’re sort of letting that art you know take shape and form based upon that data it’s kind of interesting to think that like there might be like unique like oh moments of time where the art is is different so it’s like you’re there you’re looking at it you’re viewing it you’re watching in front of you and then a half an hour later somebody comes and looks at it and it’s a different path because the data obviously yeah and by the way one thing that really you know our you know dialogues with the curators and you know our arts arts into the us we always do something very important I think is is is fundamental for us is we always unveil the name of the algorithm where data comes from researchers behind the algorithm we have this super clear sharp approach in any exhibition we have always a monitor dedicated space where we demystify an unveil the behind the scenes while we got critics oh now I learned everything about this and but it’s not about like hiding the process at all it’s about actually because I think the AI specifically in this age in this specific mission intelligence age is incredibly important like unfold as much as while there’s a significant you know fears and you know potential ramifications predictions but really want to like take this idea of like what else can we do with that and how can we explain this in a way without looking from top but you know just as a simple and honest and open as possible and I think this is a this is one of the most exciting way of sharing yeah since since so much of your work is very multi model in nature you mentioned image or text this must be a really exciting time right yes models coming out the really good stuff coming out but there is also a criticism that is being leveled against these models you know that they are basically harming traditional artists or eating into that space so you have used make a scene which is a model that you you give it a text prompt and you give it a very rough sketch and it sort of interpolates that into a very nice illustration as an artist what is your opinion about these technologies first of all I’m extremely positive at first of all I can’t draw like I can’t even like know how to draw so I’m nearly like this by a side of like if I can’t draw like can a machine help to draw better and like imagine this world so I feel like this creation is extremely exciting I don’t see that problem that some friends are seeing but I also understand them very well because sometimes people from not from the field who have been like you know physically producing and and little generating a lot of you know work and when someone you know suddenly sees spending couple of seconds and producing I can understand that feeling definitely but also I feel like it’s an extension of mind it’s a completely a new ways of imagining so I’m pretty sure that should be a like a balance between you know how to ethically source and generate these models and how to really make them available for public and I mean we saw many great examples in the discord channel to like you know GitHub’s and so on so we love it I mean with as I had architects for example with Patrick Schumacher we explore many multi models together and 3D model outcomes but I saw the joy in the face of an architect I mean they felt like wow this is like producing and promising and offering us a world that is not so hard to imagine so so this moment or like the you know a roomy right Persian poet that there’s no way to we can imagine like things about all the mods like we tried so many different ways of imagining these things that we can not otherwise so to me it’s a serendipity it’s it’s almost like you gain an additional skill where you’re taking the latent representation exactly you know yours and you’re able to bring that out in a very tangible way that other people can also communicate with them so that’s absolutely yeah so one of the things I was curious about is do you think that like AI generated art can become like too perfect like do you think that there’s like an AI equivalent to like you know the big plus happy accidents and are there digital representations of imperfections like for example like when you’re painting having brush hairs get stuck in the paint and then transferring into the onto the canvas I think I will definitely say that this for me is not lying in exactly representing reality I found the more joy and inspiration and imagination more in the serendipity of like things doesn’t go well as expected and lucky accidents or unexpected like findings in the as per wise universe as Christian mentioning to me is much inspiring than try to find you a reality that exactly mimics it I don’t believe mirroring reality is not all the way fun to personally or inspiring because there’s pretty much way long journey to like mimic our written quality and you know the sense senses that are at the moment in the most utmost level but I found the quality of again as I said is this dreaming hallucinating like our conscious behaviors apply to this techniques to me is more inspiring than mimicking the reality immediately cool since art is very open to interpretation and it’s quite subjective by definition do you think when you when you start using AI and other digital technologies to create art do you think there’s a time where it’ll start to become more individualized like for example let’s say I just wear an EEG headset and based on what I’m thinking or what I’m feeling the the art and solution changes and it becomes just bespoke for me so this is an amazing I think situation I love this idea so much because first of all unfortunately our journey with brain started 2017 with an unfortunate event that I learned my uncle was unfortunate diagnosed by Alzheimer’s by the way I’ll be going next week to meet with him after five years his diagnosed and I’m you know researching with him 101 and his brain signals that’s what I started project called mounting memories and what was really incredible is thanks to by the way professor Adam Giesel from UCS he trained me an entire our team like how to use this cap and how to use 800 plus EG research data and we use EG learn algorithm back in time to like classify the moment of remembering so to me it was like a one of the first experiment at least in our research we found that it’s still AI can be this intermediatory like you know process where our most important data I do believe memories can still be you know sub you know encrypted into this art context and and these was a very exciting exhibition that we found that people completely enjoy this idea of like transcoding the data from the you know surface level which is sometimes very noisy data still turn into like a noise algorithm and reconstruct it so this was the first time we saw a beautiful reaction from the audience around the world but now it’s very predictable that multi model process of course understanding space with like brain signals is an absolute joy and speculation still yeah I think this is this is interesting because you know part of you know the subjectivity of art is also to like share your experience with others and to to share that so like you multiple people are wearing those headsets like and and seeing what they’re how that is reacting like you know being able to share and be like this is what I see versus like what do you see and kind of seeing like what that is based upon that moment in time and the interaction that you’re having with that art I think is pretty interesting I’m curious to know you know like how and why did you choose to move all your work to to PyTorch ah-ha Christian please by the way we have a couple of more friends who are not here in the studio but I’m Christian will represent this answer very well absolutely well one of the biggest improvements for us I think I think we’re feet touched on it a little bit when he said accessibility you know really the biggest thing for us is accessibility and these abuse you know PyTorch has really allowed us to speed up our development for a number of reasons I guess the first the first one that comes to my mind would be a torch hub torch hub is a great place for all of us to be on the same page with the latest models you know it’s really easy to just swap out and we have one piece of code where you just pass in a command line flag for like image embedding for example and you know it will just change the it will change the model we use we have access to so many models you don’t have to go online download the weights compile any tools anything like that so torch torches really uh-perch torches really streamlined a lot of our processes and made everybody like make all the models super accessible to people it’s really really just a huge improvement one tool that we worked on recently so I had been using uh-yo-lovy for as like a compiled C tool for a very very long time on like a couple machines and um in the last six months or so we actually converted to your low v5 on uh-py torch and you know now everyone on our team can use the tools you can clone the Git repo you can download it you can run it on your machine you know what the environment is going to look like and it’s absolutely fantastic because now we can take this you know previously compiled tool that was compiled specifically for machines we can go clone the repo put on any machine and run it anywhere we want and it works exactly as expected so accessibility is the number one thing that comes to my mind and then you know also just overall developer experience as well you know py torch has been fantastic you know I you know I remember when we first started doing this how much time and effort you know very honestly we spent trouble shooting a lot of tensor flow stuff a lot of a lot of time there you know py torch really focuses on that developer experience and makes the ease of use really accessible to everyone it’s always very obvious well documented so it’s just fantastic and one more also amazing thing as Christian mentioned our entire again pipelines later space browser and we completely like ported back like the name and I think the speed the efficiency as Christian like compiling between machines which we always exhibit around the world and and it’s just everything became so smooth than before yeah for scalability and accessibility I mean to really improve our quality of life at the studio so so much especially as we grew and you know data science team got bigger made a huge difference yeah and that’s why one more time very much deep thankful and grateful for the everyone who is supporting this incredible journey with orbit out knowing so we are very grateful for everyone at py torch yes all the all the thousands and thousands of contributors to the py torch repo thank you guys yeah promise to make great art for humanity promise yeah no it’s it’s fascinating to hear just how much data you’re you’re taking in and you know not just like just images it’s all different types and creating that machine to to take in the sounds you know one thing I had a friend who always would say like oh I’d like to know what things smell like like looking at a picture and knowing like what you know what is this nature scene smell like and so it’s like have you were done any experimentation with with that then like yeah what are what are your plans for yes I think in the future so the entries are upcoming fun what a challenging project called data land it’s our take on metaphors where we think that exactly this type of ideas where the physical and virtual connects with a and data for any age and background open and culturally you know borderless so we are hoping to make these experiments for public in our space opening in Los Angeles hopefully next year and then inviting you know researchers all around the world to find that type of unique experiences with AI and data cool yes Rasha do you have any other questions yeah I heard something about something happening at mom yeah so very happy to say that our show is opening 18th of November at mom and this is a very important for I hope the entire AI community because it’s the very first time we will be witnessing a machine learning algorithm in real time which I believe in painting will be getting a real time data from the microphone which not recording a camera not recording and a climate data of the museum will be infusing in real time get algorithm train on the entire mom archive and this archive is so unique because I think this archive has pretty much every single time in the arts for you know trying to find a thousand options incredible like heroes of humanity’s arts so it’s a kind of a homage to like you know I think a checkpoint in the art history to remind us them and our future so it will be open and public and free and hope to be there giving workshops together sharing behind the scenes educational context and it’s free and open to everyone it’s a public art at mom Agnes and thank you Michelle and Paula to the indicators to let us generate this amazing work of art cool when you say it’s a workshop like you know do people have like hands on experience with learning how to create art with AI or is it like interacting and yeah yes hope to hope to create as much as educational context we can in this four or five months of open exhibition time yeah that’s that’s really cool it’s a really interesting and great opportunity like if you’re interested in in AI art and learn one of learn how to get started I think it’s a cool opportunity one of the things that just an observation that I think is really fascinating about this is when I think back it’s like you know and we talked about like you know can art become too perfect and you know I think that like AI art is another like medium of art right instead of using you know a palette and a paintbrush and you know representing shapes and images or abstract ideas on a physical canvas you’re you’re taking a bunch of different things you’re taking you’re not just limited by paint and pigments like you’re you’re able to create your paints your pigments through whatever you can in that digital realm you know smell data sound data climate data image assets you know you name it you can if you take it you can put it together and train it and yeah you can do lots of different things and so you know I think from just our conversation and it’s really interesting to see like this whole new art form sort of take shape thank you very much very grateful to be in the journey all together with the team and everyone thank you thank you Rafiq and Christian for joining us today that was such a fascinating conversation I cannot wait to go and participate at the moment exhibit if I may say so I think that’s a very pioneering homage to the pioneers and I’d love to experience that in person and who was watching this you should definitely make it that sounds pretty incredible yes we must not again when it’s the event 18th of November opening public all right 18th November free and open table and you said it’s it starts so how long will the event last at least four to four months at the moment yeah great thanks so much thank you guys all right thank you all right cool so thank you very much yeah I am before you leave I sent you an invite to pack church conference I know we talked about this very briefly in the very beginning do you want to attend and would you like to do a talk absolutely we will we will love to do that and that will be amazing by the way I may be in the other zoom call very happy to be both me and Christian I mean we’ll be honored to be there will be in question and we’ll work at the details and perfect okay thank you again amazing this was fantastic great also thank you so much so much for the great questions and everything yeah thanks so much guys yeah appreciate it great day thanks.