Creativity in the AI Age with Hamza Shaikh

0:00Hamza Shaikh:Students are in the education system for seven years. That's a huge portion of your life. Doing things that aren't necessarily going to actually give you value in the real professional world. Some people say it's a massive waste of time because when you go into profession, you need to be ready for construction drawings and documentation and understanding client's needs. Whereas all you've done at university is do fictional drawings of utopia. Students are not prepared for the professional world. And now we're in a stage where AI has made things even more turbulent. Hey, so before we jump in, I want to say a massive thank you because Space by Stencers has hit the top 25 design shows in the UK.

0:43And that's a really, I mean, obviously, we're so grateful to our guests. They are absolutely amazing. And all of you, the audience tuning in, sharing, following all your engagement on socials. Absolutely. like we're blown away so grateful so we've done something special we have launched show notes and you can go on to stansons.co.uk forward slash space and sign up for show notes where you get a cool little email in your inbox every week just tipping you off on the person that we've spoken to and like key highlights of what they've said some like key headline things just to make quick simple scroll read on your commute on the train so that's stansons.co.uk forward slash space and we want to bring the community on and signing up for our show notes so we'll see you there and now let's get into the conversation.

1:37Hey welcome to another episode of Space by Stansons today I'm chatting to Hamza a digital artist at Gensler and it is the first time I've been speaking to someone about the release of their book, which is very cool. It's around image generation, artificial intelligence, and creativity in design. Hamza has been with Space before on a panel discussion more, so not a podcast episode. And honestly, so easy to talk to. Knowledge just oozes out of him. So this is going to be very, very easy. But I want to just get into the book, why the book, if it's right for you to read, like what you would want to hear from him.

2:17all that type of thing. So let's jump in, shall we? Hello Hamza, welcome to space. Thank you very much. It's great to have you on board.

2:25Hamza Shaikh:Yeah, I'm happy to be here. And have you seen this out and about in the real world a lot? Yeah, I mean, I'll be honest, I've been a bit of a hermit, you know, in the last several months, just because I've had a new baby, like sort of, my baby goes nine months. So, you know, with the first book, I was able to go and do talks and go see people at universities and, and have students come up to me and talk to me about it. But this one, I haven't been out and about as much. So what's the marketing been like? If I'm being honest, I've been pretty bad with the marketing. But I think I'm just reaching a point in my life where I value quality more than quantity.

3:03I've also just become a little bit disillusioned by social media,

3:07Hamza Shaikh:a bit tired of it. Opposite, in my last book, I made a whole social media campaign. so yeah I'm taking my time with this book and you know whatever whatever comes up if it's a quality conversation with you for example I'm I'm gonna be there what is it about social media that I'm hearing this so much yeah what is it is something people are exhausted of it what is it are you feeling that with other people well a with me yeah and b with other people like a lot without just see people like I was chatting something the other day like proper age group but on no social channels no one the frustration with social media has been around for for a long time perhaps our generation is growing up a little bit and realizing that there are more important things um maybe that's one of the factors i think for me specifically i have a social media presence that is specifically focused on creativity and my creative expression and so it can become very draining to continually try and keep that momentum.

4:12Hamza Shaikh:So I think for anyone who's in content creation or even just trying to keep up with social media for their own personal brand, it just gets draining. And, you know, I'm two kids. I have two kids now. Life starts to reprioritize, let's say. I get it. What platform was the best for you when you launched your first book? Well, Instagram's always been my favorite. It's been your best. the favorite thing to use just because I used it in the very early days. And I had started, I would say my journey career-wise, passion-wise actually started on Instagram. Hence my first book was about social media in some way.

4:49Hamza Shaikh:It was called Drawing Attention, Architecture in the Age of Social Media. And so that book was essentially talking about how social media is a very valuable tool, right? Now, I came from the position, I think I still do even in this book, of looking at the zeitgeist of the time or the technologies of the time. I love technology. Surely not. Right. But looking at it with the lens of optimism, in a sense, and thinking, look, the negative things are obvious and I'm not here to negate them with mental health and social media, sustainability issues, with some of the technology emerging today. Those things are obvious to me and they shouldn't be ignored.

5:30Hamza Shaikh:However, I think what's more refreshing is to look at how you can utilize these things in a positive way or what the real utility is of these things. So with me, social media was always like, all right, I'm not going to make it about me and what I ate and how amazing my holidays are. This is going to be about my creative design diary. And so I'd say 10 years or whatever when I started using Instagram, I just started putting my architecture drawings out there. I actually put all the passion and energy that I should have been putting in university at the time into showing the world my ideas rather than my professors and tutors.

6:11Hamza Shaikh:And it actually worked out quite often. It's true, it has got a place, but it just feels like, it's a bit ironic we're talking about a book about AI, but it does feel a bit, with AI the speed of content creation has sped up so much and the volume is just exponential and now I mean it always was about like stopping the scroll and breaking through the noise and all of that like in marketing strategies and social strategies I totally got it but like now that is just like 10 times 100 times more because of the accessibility for more people to create stuff on social media, which, to be honest, can be just trash.

6:52Hamza Shaikh:It's so true. And we have a term for that in the AI world called slop, AI slop. And it was a real crisis and still is in some ways. I think people have become a lot more attuned to seeing the slop. I mean, the average person might look on social media and see something and immediately recognize that it's AI and they don't have to be an AI expert. So I think we're becoming more accustomed to understanding the nuances around AI. However, there's a part in the book there's a chapter in the book that talks about actually it's in the opening chapter that talks about how the the breeding ground for a technology like mid-journey dali and all those early stable diffusion models and something image generation now it's nano banana everyone's crazy over that right but the breeding ground for that was a world that was obsessed with images and content hence the book's called image generation because it's a play on words that yes it's image generation technology that unlocked all of this and we can talk about that a bit more if you want but we're also the generation of people obsessed with images so we are the image generation and so there's a healthy critique in the book as well as well as talking about the optimistic sides of things that's really interesting i was going to ask you where the name came from that's a really type of name i do the play on words uh what i did it in my last book as well drawing attention uh the idea was because it explored how drawings can be your source for good attention if you from your drawing from your drawings right but it's also the idea of like drawing attention to yourself you shouldn't look at that as necessarily a bad thing or a good thing but if you can use it in a way that will progress a you know a positive intent then um then there's value to that amazing and so where did your um inspiration why did you write this book you know i think it just happened i often feel life is a bit of sort of a domino effect now like things are just happening because of 50 things that happened before it yeah maybe that's just everyone and life in in general um but you know the first book i could go back to my childhood but i won't i won't bore you with that but i'd say the first book was essentially trying to capture how a disruptive technology or zeitgeist of the time in that sense architects on social media for some reason getting more followers than you know household brands like gucci like why why is an architectural drawing person on instagram having 600 000 followers to me that was odd right and i weirdly found myself in in a in a cohort let's say of people who were getting a lot of attention for things that, you know, were just drawings.

9:40Hamza Shaikh:And so that started to make me question, like, why is that happening? And is there a value to this? And so for me, identifying these disruptive zeitgeists of the time is something I'm always thinking about. In some ways, I'm a bit of a futurist, or digital transformation really intrigues me. So that first book arguably is a product of exploring digital transformation, right? Because it's looking at the use of social media, the value of it for architects. And so in this same way, when the AI boom happened, I mean, when I first saw Mid Journey, and I was one of the early beta users, like one of like 6 ,000 or whatever.

10:22Hamza Shaikh:And now it's like millions and millions. I remember typing some words in and then seeing this thing go from abstract noise and suddenly making an image of what I wrote. Now today that sounds like, so what? But just remember, three years ago, that was sorcery. That was sorcery because that paradigm was new. You type something into Google, like a pink elephant riding a horse on a cloud falling through a black hole in the jungle. It's such an abstract, weird thing to say. you put that into google google be like like are you okay right you put that into these image generation models you will get it so for dreamers and people who are visually exploring you know a million miles an hour like me i've been the common theme for me is i've been drawing for as long as i can remember it was just mind-blowing and i became obsessed with it i became fearful of it i became panicked because of it but that only pushed me to get more and more deeper deeply involved with understanding even the mechanics of like how is this actually working from a technology standpoint that pushed me to start to look at the open source world and the models that were being released some of the more advanced technology i paid a lot of money to upgrade my computers at home my wife's like what are you doing this is just madness and you know but then i'm showing her like look like you know, we can make images of anything.

12:00Hamza Shaikh:And, you know, she was, she was then scared and excited. And I just went down this route of experimentation and, and I suppose in some ways became, if you could call me an expert, by the way, when I say expert, I'm not an AI expert. Traditionally speaking, haven't studied it, no machine learning. I sort of became a specialist in that specific area because of the early engagement and experimentation. And just so much experimentation. and into it and you do get that experience don't you and yeah that's be if if whenever if I ever do a talk I say to try and make things very simple my passion is experimental drawing methods and that extends back all the way to when I was a kid like I would do a drawing that took three or four days and I would have no issue ripping it up and throwing it in a bin because for me it was never about oh look at this amazing piece to me it was about that cathartic process of doing something where an idea becomes some sort of visual representation to me the process and i think that's probably the reason i've ended up in technology is because creative process now is is in the realm of technology yeah so to answer your question that zeitgeist that blew up of the AI boom and experimentation became then the impetus to say, what does this all mean for architects?

13:26Hamza Shaikh:What does this all mean for designers? And the result of all that experimentation was the only reason, the only way I'm not going to be fearful of this is if I somehow maintain agency and authorship. And that's what that book focuses on, of how to maintain agency and authorship. Yeah, you've spoken about that before. And I really like that approach. You're using it. it's not using you or search if you know what I mean so I think that's and I think that's a really power that's when AI becomes really powerful right it starts to amplify your creativity and amplify your the amount of work you're doing or 100 % and by the way that in of itself is the focus for me right now even in my work at Gensler this is the world's largest architecture firm you know even there the focus is so how do we maintain control how do we maintain agency how are we still allowed to say we are the creative authors behind this work when you type something into an ai platform yeah i had a question i i was thinking i i have thought about that before and i know i've spoken to you about that before but as ai evolves whose image is that um is it yours or is it it's ai's and what trick what something that triggered was um was there a um image generated of donald duck or something at some point through ai and then disney were like hey that's got copyright on it so how can you do that i don't know there was i've read it very loosely and lightly and quickly but yeah it starts to sort of get into copyright as well oh there are a number of cases you know um and uh the american supreme court has as sort of really set the precedent for a lot of these cases.

15:13Hamza Shaikh:And to be honest, it's still in progress. Not resolution. No, the legal precedent is still ongoing. And countries differ in their opinion around it. You know, the UK is a lot more, let's say conservative or strict around copyright, but also training and also protection, you know, safety, online safety, and, you know, the safety of these models. The US is a little bit more open to it, of course. places like japan you know they outright said if you're training ai models on copyrighted materials it's okay which it was you know it's a very bold bold position to take it does feel like that doesn't it like you've got china and then it's just like no limit it's just go for it yeah there's then we're like there's a lot of differences and to be honest it's good uh to somewhat be in the middle ground i think because um you know we need to understand hindsight is going to really show us whether we made the right decisions or not right um and and also these things are going to evolve in a way that we don't know what the right answer is to be to be to be frank um and you know it even gets into the philosophical debate realm the idea that if you made a piece of art is it fair that I've typed something in and it's used 20 % of your art to make that, right?

16:34Hamza Shaikh:You end up going into philosophical discussions about, well, actually forget AI. People take people's ideas anyway, when they draw things, what's the difference that I've used my hand versus a mouse? Like, so there are some really interesting, that's why it's unresolved, right? And as you mentioned, And yeah, there are a number of cases where, you know, Disney and famous actors and actresses are suing or, you know, starting legal cases around people using their likeness and using their characters. Yeah. So that's definitely something that is an issue. But we're in a weird place right now that these models are still being used at an unprecedented rate.

17:20Hamza Shaikh:Right. The amount of people now who know what Nano Banana is, you know, to me is mind-blowing. Because before that, you know, no one knew what stable diffusion was or control was. It was very niche. Yeah, yeah. It's become much more mainstream, especially as people like Apple and Google start integrating it with like phones and HomePods, your laptop in your everyday. And then that's when people are like, it's just exploding, isn't it? It's not like someone like you that's super technical and you've got your amazing computer at home and you've spent all this money and you've dug into the technicalities of how it's working and everything.

17:58It's now just very like kind of just on your watch almost everywhere. Do you know what I mean? And so you're like, that's when things like Nano Banana or even ChatQBT images and that sort of thing are becoming very accessible to the everyday like un-technical user, right? So who's this book for? Is it? Because it's quite deep. I looked at it, I was like, wow, Hamza, this is definitely written by you. It's for clever people.

18:23Hamza Shaikh:Well, I've tried to make it very accessible specifically to students, right? If I'm honest, I even said in the book, you know, in the acknowledgements that, you know, I do a lot, without knowing it sometimes, I do a lot of the projects I do are for students. And I think it's primarily because I sympathize and empathize with students a lot, having been one very recently. And arguably, you know, we're always students in some way. I just feel like in architecture school, the pedagogical journey is so difficult. So difficult. you you know students are in the education system for seven years yes that's a huge portion of your life um doing things that aren't necessarily going to actually give you value in the real professional world and that's me being nice you know it could take a much more extreme take on that some people say it's it's a massive waste of time because when you go into profession You need to be ready for construction drawings and documentation and understanding clients' needs, whereas all you've done at university is do fictional drawings of utopia.

19:36Hamza Shaikh:That's not my take exactly, but that is the correct sentiment, right? Which is students are not prepared for the professional world. And now we're in a stage where AI has made things even more turbulent. Quick breather. Are you enjoying the chit chat? We are really passionate about highlighting the voice of designers and workplace managers. You can help more people and more designers like you by subscribing to the show. So we'll carry on doing our bit and you do yours and we'll make this show seriously, seriously great. And obviously don't forget to subscribe to show notes on stansons.co.uk forward slash space.

20:18Let's keep chatting.

20:19Hamza Shaikh:AI has made massive innovations when it comes to diffusion technology, which is what image generation is, LLMs, which is what GPT is, and Gemini and Lama, all these things. But it hasn't triggered enough innovation, at least for the industry to see and the world to see, in data-rich outcomes, in the idea that, yes, we can generate an image, great, but can you generate an actual built detail? Can you generate an actual built space in Root? No, you can't. And the reason is, is because there's a lot of data engineering, data science that needs to go on in order for you to create those systems. That's one thing.

21:03Hamza Shaikh:The other thing is we still haven't seen enough inference power, which is basically the chips that NVIDIA have made that really have been part of this boom. They actually need to get more stronger, at least in the commercial availability and you know the price of those those chips for everyone to have machines and models that can do very complex 3d data rich generative tasks so we're not there yet and we've seen a boom but only in a very small area of of actually what ai is capable of yeah and it's kind of the start of the process as well it's the start yeah i think the reason you've seen a lot of urgency and i mean i at least i felt a lot of urgency is because i don't if i think if we don't get our foot in now and start to think proactively about even just these technologies that are impacting us then we're not going to be ready for what comes next yeah i was actually listening to you know must this morning at the world economic forum and he was speaking about if you heard him you know i had to see okay he speaks about by the end of this year ai and robotics will be more intelligent than any human yeah and by the end of three years AI and robotic will be more intelligent than the whole of humanity combined, whether you believe him or not.

22:18But I think he's definitely got his head screwed on in that area. And you start to think, wow, that's going to really change how people, whether they like it or not.

22:28Hamza Shaikh:100%. I mean, I'll show you something I'd written in this book, which people can hold me accountable to, was actually a chart and a graph of what I predict to be happening in the next 10 years or so. And we can put this on screen if people want. Yeah, that's right. You know, during the actual video, what you'll see is that from 2012, I sort of give a history of what has happened in terms of the AI impact and looking at this from the lens of creative AI and image generation, but also slightly broader to that. You can obviously see in 2012, there was a deep learning model that really triggered a lot of this during the competition, AlexNet, for example, and that unlocked the world of image generation.

23:11Hamza Shaikh:They made this model. And then if we fast forward, obviously we saw ChatGPT, mid-journey stable diffusion in 2022. And then in 2024, which was at the time I'd published the book, I'd obviously talked about the release of Flux models, which interestingly, they were the people who actually made stable diffusion. They left and started their own thing. And then I knew that was going to be a huge shift in terms of the way things are going to go. Sora came out, VO2s, all these video models started. So then I start to talk about 2026, 27, 28, 29, 30, and then 30 plus. And you can see in 27, I talk about advanced robotics emerge.

23:56Hamza Shaikh:So maybe I'm a year difference to Elon Musk, but the point being there, this is also so relevant to how that's going to impact the design world. because we're talking about the construction industry at large, not just construction, but the industrial industry, industry, manufacturing, everything, the world, is going to completely transform when robotics can take on that level of automation. It's actually on a much different level. I was talking to somebody about robotics and I said every house will have a robot shortly, just like you have Wi-Fi. and they're like don't be stupid i was like of course he will but what would it do so i started really and i thought it would do your ironing go to the shops for you like look after your kids for you look after your pets for you look after the elderly and i'm like can i have it now and i'm oh there you go it didn't take long did it it didn't take long for people to suddenly want one i'll give you another anecdote my wife anti-ai anti-gpt which is funny right because i'm yeah opposites attract yeah but um by the way side note i i love that about her because it gives me balance yeah but now she's she's gpt like crazy how much you are what are they saying that's the thing as soon as like you said with your friend like as soon as you see the value it can actually bring you in your day-to-day life it doesn't make sense to not have it now i will say this we do need to be informed we need to educate ourselves on what it actually what's behind these models right because you have private companies that are essentially taking all of that data and you need to be informed i think data uh proficiency is one of the things i talk about in general in the book and elsewhere that data proficiency is the next big thing for everyone to learn.

25:49Right.

25:50Hamza Shaikh:We need to know what it means to, to give our data, to be in charge of our data. And I think you're going to see a lot more about data ownership emerging this year, what it means to actually be in charge of your own data right now. You don't have access to your data. And when I say your data, you might think like, what would you mean? I'm talking about, you know, the steps that you took to get to the office today, that's your data that has been reported on your apps. It's yours from a moral ethics standpoint, but it's not yours from an ownership standpoint. So I really love to explore the idea of like, what does it mean for us to actually be in charge of all of our data?

26:30Hamza Shaikh:And then we become the vendors to the big companies who actually want our data. To me, that's an interesting concept. It is an interesting concept because historically, in terms of digital Anyway, all of our actions online have been tracked and that sort of thing, which provides those companies data to sell to advertising. But that has been how they've managed from a business model perspective, how they managed. We've had so many free services like Google, which is absolutely brilliant. It's available to everybody because it's free. As soon as we're in control of that data, suddenly would Google be free anymore?

27:11because what's the benefit? Because we've got to sell the data to Google to sell the data. It changes the whole business model of so many apps and services and online everything.

27:23Hamza Shaikh:That's a very fair point. You know, do I expect us to all be, you know, in charge of our data? I mean, not really. Unless there's an incredible startup that emerges and finds a way to do it with some blockchain technologies unable to be, you know, interrupted. maybe but i say that more as a provocation to trying to get a point across about how important it is yeah for us to at least be informed uh that if we do get a robot in the house you know forget the robot even just you using gpt like you the amount of informality that people now have with gpt and talking to them about their life issues and and you know everything and anything you don't know what and how that data is going to be leveraged and monetized and that's worth knowing and also one of the sentiments i really want to get across as well and this is is is also why this there's a reason that this book talks about models and technologies that are now outdated and i knew that would happen and publishers were talking to me about that a long time and i said look, I know these things are going to get outdated, but what I'm trying to teach in this book is not how to use control nets.

28:44Hamza Shaikh:I'm trying to teach you why control nets are important in the first place. And I'll give you another analogy. The people, the graduates who are coming in, who we've identified as talent and working with us, I've seen that the ones who are the most desirable to employ are the ones who have an understanding of legacy ai platforms from the start because they have they have a mature approach to then figuring out what to do with the next technology they have a more well-informed idea about what it means to use ai technology and it might sound like a very very minor nuance but it's huge and the reason i think it's so important is because our roles are now going to be moving towards everyone becomes a director everyone is kind of a director now you've got your creative agency in your pocket you've got your writing employees in your pocket so your job is to direct and you can only direct if you truly understand i think the mechanisms behind who you're directing yeah it goes it goes back to what we're saying about education it's not and often say this and it's not about what you learn it's about learning how to learn and that sounds like a really simple statement but learning how to learn is so key because you could write this book you could write a hundred books and then i'm sure they'd all be fantastic hands but however they will be outdated and and likewise everything in the schooling system the only thing we can pass on to our children is learning how to keep experimenting how to keep learning i'll be gone and i can't so there's nothing point knowing what i know because the world will be different but you need to know how to navigate how to explore how do you know what i mean nail on there and it's about learning how to learn it's such a simple statement but i'm so strong on it in terms of like i've got a three slash four year old boy oh nice yeah and you just think like there's no point me teaching him anything yeah i just need to teach him how to find out love that do you know i mean because it's not i'm old and i'm quickly getting older yeah you know it's interesting it's such a great point it's exactly it's exactly what i think i'm trying to capture in this book and also i have to give credit of course you know we haven't mentioned uh you know i should have mentioned earlier that actually the book is is made up of 10 contributors as well i saw that in the um i saw that in the car so there are 10 other contributors who have asked to write chapters in this book and the way the book is set up is to go through actual practical workflows that capture a moment in time of of a specific approach to using ai not trying to highlight the tool specifically but showing you this tool will use with this tool and that tool and many of these are open source tools.

31:47Hamza Shaikh:It gives students that legacy understanding by the knowledge of students, professionals. There are also interwoven chapters there by contributors who are talking about their own theoretical stance on AI. You know, one of the chapters talks about, you know, the idea of, you know, bias. You know, all of the models have a bias. You know, when you're dealing with a model that is biased, it's actually putting forward ideas that are just not suitable. And this is also a scary thing to think about generally if you think about how there is a greater and greater, let's say, narrowing of where these AI models lie and who's using what models.

32:34Hamza Shaikh:It just means that, for example, if the design industry becomes completely reliant on Nano Banana, then is google actually responsible for the design or are we um i think these kind of questions are important to put forward now right because we're still at the early the building stage of this so so totally totally totally resonate with your point that's interesting and whilst we do talk about learning how to learn and giving them the tools here which is which is great and it is my like underlying most important thing people always want to know what ai techniques or tools you are using so i've got to ask you that because people want to people want to see that so what are you using um or what do you see has the most potential to changing your everyday design process okay so you it's a hard one for me to answer because a few things i wouldn't quite count myself as like a generic user of ai primary reason i would use ai is for my is when i'm trying to experiment with my artwork a lot of that i actually don't even put on instagram because i'm just in my own space and just experimenting another way i would use i use ai is i talk to it a hell of a lot when driving i'll put it on voice mode and i will just reel off things that are stuck in my head i'm I have a lot of things in my head that I need to structure.

34:04Hamza Shaikh:And so I have chats of massive venting sessions that I have with AI. And I'll be honest with you, I'm not trying to advocate or push away the use of any models, but these are my personal opinions. When I used ChatGPT recently, I'm really struggling to have an in-depth conversation. For example, I do think about Claude, Claude from Anthropic. It's a much more favorable LLM model for me, personally speaking, being able to get to that level of depth, especially the Opus model is very advanced. So I love to talk to Claude, organize my thoughts. And then when I'm doing my architectural experiments, I'll use an array of tools, going from old legacy tools like stable diffusion and control nets to nano banana sometimes to put that final finish on things.

35:05Hamza Shaikh:But really, I'd say right now, I'm actually trying to keep AI technology away from my art because I've started to get a bit scared about my reliance on AI for art. you know i've moved it became an experimentation and then i think i started to feel oh my god am i starting to now rely on this to express my art so now i'm moving back to just i'm sketching in a sketching in a notebook you when you do the sketches do you video yourself doing the whole again i'm just going back into artist mode i struggle with putting the turning the camera on it almost it almost takes something away from when i'm drawing and when i'm painting um it feels as if i'm doing it for the wrong reason and i'm contending with this right i i know the value i wrote a book about the value of social media so at the same time i'm i'm trying to figure out how can i do it in a way that is genuine but also you know i can i can express it to other people i'm still trying that's why a lot on my phone is filled up with videos of that they just haven't i haven't done anything with i've literally had this conversation quite quite deep with someone quite recently and it's actually not it's all about what your objective is and what your vision is right because i'm for example i'm doing this podcast it's nothing about me it is truly honestly it's not i couldn't i don't give a flying hoot about me it's just about sharing people's voices helping the design community helping people be more aware of the trends and that sort of thing and so that's what it's all about it's all about everybody else so the whole purpose is it's social and it's out there because it has to be to support and help people but then other people do fashion or cooking or art or anything like that and it's nothing about anyone else it's their own little recluse in the busy busy world and it's about them finding their happiness and their expression and their outlet and so if you're doing it To serve others, it's got to be online.

37:07If you're doing it to serve you, it doesn't have to be online.