#98: practicality vs. play
[companion machines module 05] what if the fastest way to solve a problem is to make something fun?
Welcome to Month 5 of Companion Machinesâa monthly essay series within Time Spent on emotionally grounded relationships with AI, for educators, artists, parents, and anyone curious about working with intelligent systems.
Not here for AI? â Unsubscribe from just this series
Very here for AI? â Upgrade to unlock a paired module with key skills, risks and readings for intentional AI use
Good morning,
Earlier this year, I watched my friend Kawandeep Virdee give a talk at the MIT Media Lab called Thinking with Sand, about the role of creative technology in a fast-moving AI landscape. I couldnât stop thinking about it because everything I consume about AI is so serious, and Kawanâs talk⊠was not.
Heâs one of the most lighthearted, playful and thoughtful people I know and through a series of examples of things heâd done with AI, he reminded me that play can be a legitimate (sometimes the best?) interface for meaning-making, especially when tools are unfamiliar.
As someone who tends to reach for tools only when I have a clear use case or actual problem to solve, I found myself rewatching the talk a few times. It made me want to start experimenting again. What if the way to learn wasnât through analysis or use-case optimization, but through joyful, small-scale, emotionally resonant experimentation?
So, I called him up and we had the loveliest chat about practicality vs. play and what it means to build tools (and use tools) in a way that lets us feel human.
The full conversation is below, lightly cleaned for readability, but left mostly intact. Below it, youâll find this monthâs module on how to stay rooted in your own logic while using machine tools.
(The interview is free, the module is paywalled.)
Enjoy!
Jihii
On Practicality vs. Play: An Interview with Kawandeep Virdee
Jihii: Why donât we start with context so I can explain who you are. Tell me what you do and how did you get into doing it?
Kawandeep: What I work on right now is AI prototyping at Google Labs. More specifically, I joined as part of their creative technology residency program. Itâs a team of CTs that are embedded within Labs UX. So Iâm within a design team, but I build interactive prototypes.
Google Labs sees itself as a collection of futuresâthinking of what kinds of products and experiences you can build that will be available in the future, things like six months ahead to a few years ahead. So what it means for me as a creative technologist is that Iâll have conversations with people around the company and other designers, come up with ideas and more specifically come up with prototypes for them.
The reason why they were interested in starting the role was because often when new ideas would come up, most people are working on their day-to-day tasks, and there isnât that much time to think ahead or think weird, think different. The thought was if you had an idea that was experienceableâlike you could play with it or interact with itâthat it would give the idea more momentum.
Weâve known each other for over a decade. Since then, my path has been a little bit meandering, but there always has been a common thread of creative technology in my work. For the most part in my career, itâs always been on the side, outside of my full-time work. And then this role was like, whoa, first full-time role really doing it, but after having done it for 15 years.
Jihii: In my mind youâve always done this work, but in reality, youâve been sustaining a practice for a long time thatâs now your actual job.
Kawandeep: Yeah. When I look at it on paper or LinkedIn, Iâm like, man, this can feel very meandering and wandering. But there is this thread that for me feels really meaningful.
Jihii: How does one become a creative technologist in the first place? What does that mean in terms of education or experience?
Kawandeep: I think the role itself can be kind of amorphous and depending on context, it can shift. Generally, my sense is itâs somebody who uses emerging technology in a way thatâs more exploratory or requires some artistic talent as well. Having the ability to think more creatively or broadly or make unexpected connections that you would do in artwork can become useful.
There are certain programs devoted to thisâParsons has a design and technology program, NYUâs Interactive Telecommunications Program. More and more programs are combining design and technologies.
Jihii: What got you interested in LLMs and genAI?
Kawandeep: At an AI hackathon I organized in 2022 (with your husband!) one person I talked to was like, âthe last time I felt this kind of excitement in the Bay was when the iPhone came out.â And to hear that and then see where weâre at nowâitâs like, okay, this thingâs really important.
For me, when I looked at it, I saw what people were making and I saw a space that was just an open field ready for exploration, where people were hungry to see different ideas. So I thought, let me put this creative technology ability I have and focus it on this emerging technology of LLMs and see what happens.
Jihii: Thatâs awesome. What I am curious about is the raw skill and posture that you need to develop in order to be able to play and build things like that. When I watched your talk, it felt like this approach to tech is something you might be able to take for granted because itâs just how you live, but most of us donât.
Kawandeep: When I read this in your earlier note, I felt like you were seeing things and articulating things that for me are like the waterâI donât notice it.
But yeah. With AI, a little over a year ago, there was just a ton of criticism I feel like was peaking and it felt like to work in the space and to be interested in it felt like a moral failure or ethical failure. It just felt like something was off about it.
Even though I was just making stuff independently, I was like, âIâm using these tools to make things, and I have this sense that maybe the tools themselves are inherently harmful. So am I making art with guns? You know, am I making them fun and playful and cute by making art with them, but the tools themselves are dangerous? And in one of my office hours conversations, Katy Peters, who I met years ago at a Media Lab event, joins and I shared this question because I feel, âthis is what Iâm wrestling with right now.â And she says, âwhat youâre doing is youâre showing another way to imagine the future and how we think about and work with these tools. And we need more playful ways to think about them and playful things that we make with them. The issues that youâre mentioning here arenât based on the tools themselves. Theyâre about how the tools are governed and the organizations that use them and how theyâre governed.â (Iâve written more about it here.)
And that was just kind of illuminating to me because she had spentâsheâs in the civic tech space, so sheâs spent much more time thinking about this. And so it was a bit relieving and I was like, âokay, Iâm not gonna stop. Iâll keep exploring this and now I see a kind of purpose or a vision around it that I didnât have before and around play itself as well.â
Jihii: I feel like thereâs this unprecedented opportunity weâre seeing right now where people can finally develop some of the literacy skills that we couldnât do with just the version of the internet we had for the past 15 years. Weâre now moving into a system where people have the chance to actually get how to live interacting with technology with a lot of intention because thereâs so much more you can do. So you kind of have to get your hands dirty.
Kawandeep: It makes me think how hard and weird it was at first. Two years ago when I was prompting an LLM and just being like, âI actually donât really need this. Why are people talking about this so much?â And then all the little steps that went from there to now Iâm using it to practice Punjabi or share stories in Punjabi with family, to prepare for health and medical things, to navigate complex forms. Once youâve used it a little bit, then youâre doing something and youâre like, âwait a minute, lemme just use this tool for it.â
Jihii: So hereâs where it gets interesting for me. Where is the intersection point between practicality and play? By practicality, I mean, I kind of have a problem that needs solving. It could be the smallest, most personal problem in the world. In my case, I have a dog that needs training and no trainer quite gets the situation and it would cost me too much money to give them the amount of information I would have to in a one hour session.
So I was like, Iâll just put all the information into ChatGPT, and then talk to it and doing that changed everything. I decided, screw it, Iâm gonna just do the training myself and have this ChatGPT coach that knows the history of my dog very well. Iâve been doing training every day with a daily training planâhereâs what we need to work on, hereâs how it went, and then it gives me back advice. Now I just have a dog trainer that is an AI.
Kawandeep: This is just an amazing use case. I love hearing this because you donât need to build an app or something. You have this tool, then you figure out a way to leverage it to be extremely effective.
Jihii: Yeah, but for me, I mostly am using AI these days to offload cognitive planning that is too taxing for me. So itâs like we know what we need to do, but we need to make a very meticulous plan, taking into account my limitations of time. Itâs just like I needed a bigger, faster brain.
But I want to be able to be more comfortable with use cases that arenât just practical. I donât wanna just solve the problem and get out. Every time you take one action to solve one problem using AI, youâre rolling out the carpet in front of you and you could go in a million directions, which feels very new. So Iâm curious if youâve thought about this intersection point between practicality and play.
Kawandeep: I love this question. Part of what comes to mind is the way in which the two can sit alongside each other. You can have something thatâs practical, useful, and then the playful parts of it will make it delightful, make it feel meaningful and enjoyable. And then youâll want to use it more.
I think right now, if I give a presentation, thereâs a way I can give the presentation thatâs probably the best for communicating ideas. But may not be the most enjoyable for me to make. And so I still wanna get the ideas across, but if I use bright colors and have fun with it, I just enjoy the experience more. The ride is better.
When it comes to making a tool or a prototype, sometimes it can be as simple as thinking about the aesthetic of it. I can have a little bit of fun with this and do things that I normally wouldnâtâuse bright colors and gradients or weird effects or emojis everywhere. Things that make the experience of putting it together more enjoyable for me.
Jihii: Letâs talk about examples. Thereâs two I liked the most from the talkââGas Me Up,â and the âEnthusiasm Bot.â I feel like theyâre very easy to understand because thereâs a super clear purpose to them, but theyâre incredibly playful and fun. How on earth did you arrive at that?
Kawandeep: All of these were driven by what technology was available. When I started playing with LLMs, people were making different chat bots and imagining different ways to pre-prompt them so you could create these different personalities.
So I started making these different personalities that would be helpful for myself. Thatâs where Enthusiasm Bot came in. I was living at home and we were coming out of the pandemic and I needed to be social and get out of the suburbs because I was getting restless. So I would have trouble just doing thingsâsomething would come to mind and I would have so many things pushing against that.
I made it to talk to me. To convince me. The prompt was something like, âyouâre my best friend and you help motivate me.â So I would share things that I really wanted to believe, but I had trouble with. So it would give me the language that I could read to myself and then feel okay about whatever decision I was making.
I found that thereâs an affordance to these chat bots where we may not always be able to remember or recall things in the moment when we need to. But these can help with that. Iâll be like, âI want to go into the city, but I donât know what Iâm gonna do, but I feel like itâll be good for me.â And then itâs like, âjust think about the last time you were in DC and how fun it was.â And Iâm like, âyou donât even know what I did, but youâre right.â
For âGas Me Up,â OpenAI launched their GPT Vision model so you could submit an image. I thought back to these wellness things. What if I went the direction of affirmations? The first few runs I was like, âoh my God, this is so sweet.â But then as I ran it again and again, I was like, âoh, this is really repetitive.â
So then my interest moved towards seeing other people with it. When Iâd see other people use it, Iâd see them light up and Iâm like, âwow, this is so nice.â It felt so good to see the happiness it brought for other people.
Jihii: So you were just doing them on your own, no reason just âcause you were curious?
Kawandeep: Yeah. And that overall question of like, âHey, thereâs this emerging technology. A lot of people are paying attention to it. I think I could make something valuable by bringing my expertise of creativity.â The hope was Iâd make it and share it and then it could start conversations.
Jihii: Iâm thinking from the perspective of someone whoâs at the exact moment before you made Enthusiasm Bot. How doable is it for someone whoâs not a technologist?
Kawandeep: For that one, I couldnât have done it had there not already been that demo and code available. But now with code generation, you kind of ask for certain things that you want and then it generates code and then you put it in.
One thing I realized as I was putting together the talk was the theme of it was tools for thinking and creativity. Software is one tool for it, but a pen and a paper or certain processes of facilitation are just as potent. And even the LLMs themselvesâyou can build things like Enthusiasm Bot, you can do that right now with it. You donât have to build an app for that. You could just prompt it in a certain way, kind of like the way youâre doing the dog training.
[Note: Kawan and I talked a few months ago and since then it has become even easier to do, see his latest newsletter for some methods.]
Jihii: For the actual designing and play process, what are kind of things that you think people should keep in mind to be able to play?
Kawandeep: This started to become more clear to me about a year ago. Itâs this feeling ofâwhat do you get when you play and whatâs happening? It feels like this state where whatever youâre doing is being recognized and something is responding to you and youâre responding to something else. And thereâs this general light tone or attitude towards it. So feels like maybe you feel like youâre in the edge of giggling.
I find moments like that, if I can get to that in a conversation, it just feels like, âoh wow, this is connecting. This feels really nice.â And so how do I get to that reliably?
Thereâs this feeling of the unexpected, the surprise and turning towards something that is pleasurable. Weâre living in this world and we can decorate the spaces we live in to look stunning, or to just make ourselves happy. You could just put up something because itâs silly.
So I think about that quality of the unexpectedness or the spontaneity, but also this tone of delight. Sometimes people can see humor asâif someoneâs really silly, itâs like theyâre not serious. But they donât have to be at odds. You can have deep connection with humans.
Jihii: You made me think of this exampleâone of the reasons I married my partner is because he is play 100% of the time. I remember in the pandemic, I had a really late night editing a podcast and he went to bed before me and was gonna leave early for work. When I woke up in the morning, I found, on the couch, that he had laid out his full outfit, his jeans, his sweatshirt, and instead of his face, there was just a drawing of his face saying good morning. He could have texted me, âgood morning, have a good day.â But because he was playful with it, the outcome was wildly different. It absolutely made my morning.
Kawandeep: And itâs so memorable. We live in this world where this is possible. How giddy he must have been setting it up. And then he knows youâre gonna wake up and he is anticipatingâwhatâs she gonna say? Itâs just so out of the ordinary, and so potent. I really feel like this is the stuff we live for.
Jihii: I feel like it just changes the quality of your life to try to solve your problems with that spirit. Thatâs what I feel drawn to move toward in my professional and personal life these days. Anyone can play in these practical little ways but you have to have an open channel to that playful, childlike part of yourself to even go there.
(Those are just some highlights from our chat but you can go watch Kawanâs talk, Thinking with Sand, to learn more!)
What struck me most about Kawandeepâs examplesâEnthusiasm Bot, Gas Me Upâwas both the playful presentation aesthetics but also why they worked.
He wasnât optimizing for efficiency or scale, but feeling and surprise and the specific texture of his own mind and needs. Which got me thinking about something Iâve been wrestling with: when we engage with AI tools, whose logic are we actually using?
Are we adapting our thinking to get better outputs based on the design of the tool at hand or are we teaching the tools to think more like us?
The difference matters more than I initially realized, and itâs what I want to dig into below.âJihii
đ Module 4: Logic vs. Self-Logic
Note for free readersâthe full paid module below includes:
The posture shift from productivity to play
The risk of letting machine logic override human knowing
Self-logic as a design practice
Suggested exercises for playing your way into clarity