Overview
Jeremy Gilbertson and Mark Fielding from Thinking on Paper interview Harry Kumar, co-founder, director, and Chief Creative Officer of Moth Quantum. Moth is the company behind Quantum Backrooms, a browser game whose maze is generated on real quantum hardware. Each tile is a qubit, and the walls come from the correlations between them. Harry is clear the game runs no faster than a classical one.
He explains why Moth built Atlas, a platform for musicians, artists and game makers with no quantum background, why he thinks quantum today is where generative AI was in 2016, and why he tries not to let AI replace his thinking, starting with the first draft.
This podcast is divided into 14 chapters as follows:
- Meet Harry Kumar of Moth
- Why Moth built a Backrooms game
- Every tile is a qubit
- No quantum advantage, and why that is the point
- Atlas: a platform, not a games studio
- Who is quantum creativity for?
- From an entanglement shader to a reverb plugin
- Can you hear quantum randomness?
- How the quantum industry sees Moth
- Which modality makes the best beats?
- Pseudo-randomness and computing with nature
- Is creativity inherently quantum?
- The handover question: AI and the first draft
- Where to try Atlas
Editor’s Note: In the podcast Harry uses the phrase “Dylan Goes Electric”. This refers to an incident at the 1965 Newport Folk Festival where performing artist Bob Dylan shocks the audience by unexpectedly using an electric guitar instead of the traditional acoustic guitar normally used at folk festivals. Harry sees this as an analogy to what can happen when creative people take a hold of a new technology and run with it.
Transcript
Meet Harry Kumar of Moth
Jeremy Gilbertson: Disruptors and curious minds, welcome to another episode of Thinking on Paper. My name’s Jeremy. This is Mark. We’re talking around two of our favorite subjects, I think, today. Who do we have on the show? Kick it off and tell us what we’re getting into.
Mark Fielding: Consumer quantum computing, quantum computing and gaming, quantum computing and music. Where, what is happening? Why does it link? Why are we speaking about quantum mechanics, quantum computing and gaming and music? We’re gonna find out with our very special guest, Harry Kumar, the co-founder of Moth Quantum. You might have seen their Backrooms quantum game. You might have been wondering, how does quantum fit into the creative and arts world? We’re gonna find out. Harry, welcome to Thinking on Paper. Thank you for Thinking on Paper with us today.
Harry Kumar: It’s a pleasure to be here and excited to share more about Moth.
Why Moth built a Backrooms game
Jeremy Gilbertson: Mark, before we start, have you seen backrooms?
Mark Fielding: I’ve played the game, I haven’t seen the film, no. I’m aware of what it is, but I haven’t seen it, no.
Jeremy Gilbertson: But yeah, you haven’t seen the film? You haven’t seen… so funny story. I went, just super quick story. I went with my kids. They wanted to watch it, don’t know if you’ve ever, you’ve probably been in movies that make your body and your brain feel weird. That one made both my body and my brain feel weird, almost to the point where I was like, I might want to go, but I can’t leave at the same time. So anyway, Backrooms is very interesting. Let’s start there, Harry, if we can, because you released something really right when that movie came out on the gaming side. And what brought you to kind of grab that as a potential inflection point into the market?
Harry Kumar: Yeah, you know, it’s a great question because I had a lot of friends tapping us on the shoulder at that time saying, “How did you guys know Backrooms was going to be such a big thing?” And I think it comes from two things. One, you know, we’re a quantum computing company, but we also have a team of creative technologists, artists, people really, you know, fascinated by contemporary culture and an even broader sort of community of people who are really part of that internet culture, contemporary creative practice sort of world. And we were starting to hear murmurings about backrooms and the whole internet lore around it for quite a while before we thought about making that game. But there was also a sprinkling of luck and serendipity in all of it. We had one of our researchers working on a new paper, and he decided that he wanted to build a demo that could work around this paper and really showcase what he’d been working on, what the team had been working on. And the original idea was actually a mansion-style game, where, you know, it’s a bit like a haunted house mansion. And just at that moment where the kind of proof of concept for this game started to be written down on paper, one of the members of our team, I think it was actually our CTO, brought up backrooms and said, “Hold on, this feels like the perfect game mechanic for a backrooms game.” You know, a world that shifts around, the walls never stay in the same place, they’re almost in superposition. This feels like a kind of spooky backrooms environment. So it was serendipitous that those two things happened at the same time. And then we of course decided to launch at a time that was around the launch of the film. We thought that would attract a bit of excitement and interest.
Every tile is a qubit
Jeremy Gilbertson: So we’re talking backrooms. We’re talking about the game that you guys built to demonstrate something that one of your team members was trying to prove in a paper. What were they trying to prove in the paper and what did you want to demonstrate with this game?
Harry Kumar: Yeah, so the paper was about something called quantum imaginary time evolution, and really broadly about quantum dynamics and leveraging QPUs, quantum processing units, to harness their quantum dynamics and apply that to a game mechanic. That was the essence of what we were trying to do, and then we wanted to demonstrate that in a game. So the quantum dynamics, the inherent randomness, indeterminacy, quantum phenomena emerging from running things on quantum computers today, utilizing that for something that’s engaging in the form of a game mechanic. In this case, you know, the shifting of walls as you move through a space. And the effect that we really achieved in that game was that you’re almost exploring the quantum processor itself. You know, each tile of the grid, each tile of the map that you navigate through to try and escape the backrooms, each tile is essentially a qubit. And whether there’s a wall or not between one tile and the next tile is determined by the correlations between those two qubits, the relationship they have to one another. And then what we do is we set up an ideal state for that end map, that you know, the map, the maze, the configuration of walls. And then we allow a quantum process to kind of evolve towards that state, and so you’re really kind of exploring the QPU as it, you know, evolves through this process.
Jeremy Gilbertson: So the end state, the ideal end state, that was specifically mapped out to accomplish something or just a creative endpoint to land on?
Harry Kumar: Yeah, more the latter.
Mark Fielding: Finish the game.
Harry Kumar: More the latter. I mean, you guys might be familiar with this kind of quantum algorithm from something like quantum chemistry, you know, an optimization algorithm trying to optimize towards a grand energy state. It’s in this case reappropriated for, you know, the building of a game and game dynamics. But yeah, the end state we’d set was one that is creatively defined. And in fact, we just a few weeks ago had a booth at Gamescom, which is a big gaming conference in Germany. And we had everyone from kids to people who are part of the quantum industry to game devs coming up to the booth and setting their own ideal map state, running that on, in this case it was IBM Fez or IBM Miami, and then actually seeing their own Quantum Backrooms level created. So yeah, anyone can set the end state, anyone can run the process, anyone can make a quantum backrooms level.
No quantum advantage, and why that is the point
Jeremy Gilbertson: So, Mark, if this sounds confusing, we’re gonna point back to a couple of words that you bring up a lot whenever we talk about quantum is like path integral. So you have everything’s interactions, everything is this, having some sort of coherence with this and some kind of interaction with that, but infinite interactions and infinite possibilities. Harry, would you say that this is like being able to, you know, move in a direction of the room and somehow understand all of the potential options and or interactions with all of the pieces and then figuring out or not figuring out but snapping into the wave collapse of an optimal version of that interaction or am I getting too weird? Like what’s happening here?
Harry Kumar: No, no, not getting too weird at all. I love the way you’re thinking. And actually, if I take a step back for a moment, precisely what you’ve just described there and the way you’ve articulated that and the way you’ve turned something you know about in the quantum industry or the quantum sector into some kind of creative idea is precisely what Moth’s trying to do, which is make it possible for people who are not developers at a quantum company, engineers, researchers, make it possible for those people to start thinking about interesting ways that we can apply quantum at a level that…
Mark Fielding: That means you’re wrong, Jeremy.
Harry Kumar: I mean, I’m not saying you’re wrong. What I’d say is that, you know, quantum computing does unlock that sort of space, or it’s something we are excited about, is being able to unlock and process things in a completely different way, akin to what you just described. What I’m hesitant to say, and what Moth as a company is, you know, hesitant to claim, is that there’s any kind of quantum advantage present in this. Or that backrooms presents any kind of performance upgrade to a classically developed game. It’s not doing that because the underlying hardware today does not allow for those kinds of speed ups or the solving of intractable computational problems. We know that those will be solved in the future with new generations of hardware, and we’re pointed, at least a great number of us at Moth on the research team are pointed towards unlocking and discovering what those new generations of hardware, you know, and fault-tolerant quantum computers, what will be possible then. But the game today is not solving for an optimization problem and providing a performance speed up or anything of the sort. It is, however, and I think this is where you weren’t wrong in how you described it, it is allowing you to situate yourself within a gaming world that is governed by quantum dynamics, by the things you described. And those, you know, things like indeterminacy, randomness, the dynamics of a quantum system are present in the game that you’re exploring.
Atlas: a platform, not a games studio
Mark Fielding: So are you using the game higher level? Are you using the game to explore quantum mechanics as much as you as the player would use quantum mechanics to explore the game of backrooms? Is this a tool?
Harry Kumar: I’d say that’s a good way of looking at it. I mean, Moth is not a company that is focused on building games or developing games. We made quantum backrooms, but what we’ve really been focused on over the last couple of years and what we’ve actually just this week started releasing to our alpha user group is a platform, a platform that was used to make quantum backrooms. But that anyone could use to make their own quantum game. So that is a set of tools and providing the capability for someone to play with quantum at an abstracted level. And that’s what we’re focused on building as opposed to games. Now, quantum backrooms was a really great test for us to try out the platform, see how it integrates with other existing tools on the classical side as well, and really show the world what’s possible to make with our platform. It’s kind of, I mean, Mark, you might be more familiar with the idea of Blue Peter, you know, the TV show that was on when I was a child. And something that always came up in that show was here’s one I made earlier. And backrooms is kind of one of those. I can’t believe I’ve just referenced Blue Peter on a podcast.
Jeremy Gilbertson: All right, I gotta stop you. What the hell’s Blue Peter? Talk me through it.
Mark Fielding: Yeah. No, don’t, I’ll explain after the show, Jeremy.
Jeremy Gilbertson: Okay, fair.
Mark Fielding: Don’t go there.
Harry Kumar: Yeah, I can’t believe I’ve gone there. But the idea really being that like we’ve got a platform that can be used to make games, it can be used to make music production tools, it can be used to make almost anything you’d like that uses a QPU and makes that easy for someone to do even without quantum experience. And Quantum Backrooms is just an example of one of the things you could make.
Who is quantum creativity for?
Mark Fielding: I’ve played Quantum Backrooms. I haven’t played GTA six. It’s not out yet. But who’s this for? Because a teenager, an adult, anybody who is interested in making games, this doesn’t feel to me like I’m gonna make a quantum game. I could go on, I could use AI, I could go on an app store and build it brick by brick in a classical sense very easily, almost plug and play. Who’s this for?
Harry Kumar: So the platform that we’ve built or quantum backrooms?
Mark Fielding: Both.
Harry Kumar: Yeah. So I mean quantum backrooms is for anyone who likes games, who likes the Backrooms genre, and is interested in quantum. So, you know, we’ve had kids coming up to our booth at Gamescom, we’ve had people from across the creative technology scene being really excited by it. And it could even be for the general public. It’s a game that’s playable on any browser anywhere, and it’s super accessible in that way. The platform we’ve built that made it possible to make quantum backrooms is really pointed towards in the first place creative people. We’ve got people who make music plugins, but we’ve also got musicians who don’t know how to code but would like to process music using a quantum process. We’ve got game devs as well as gamers who are part of our Discord community and trying out the games that other people are building using our platform. And in theory, like one of the easiest ways that you can interact with our platform, which is called Atlas, by the way, is you could log onto the web platform, you could go to one of our engines, which is an image-specific engine, you could drop a JPEG or a PNG into that engine, move a couple of sliders, press run, and you can get an output that has run either on an emulator or on real quantum hardware without doing any coding. So in that sense, it could be used by anyone who wants to see what an image looks like when they run it on a quantum computer and run a certain set of processes. So in that sense, we’re trying to open up the aperture to as many people as possible. And really, you know, when you look at the quantum industry, like you guys have had plenty of guests on talking about quantum before, you know, it’s really been confined to a very small subsection of people, a very narrow demographic. And those people are concerned with problems like quantum chemistry, cryptography, cybersecurity, pharma problems. And there’s a broad range, and there’s no doubt that that work will lead to real-world impact. But at the same time, the quantum industry and quantum computers themselves are sometimes described as a solution looking for a problem, right? We know we’ve got these computers. We can see that the timelines are lining up and that they will reach a point in the next few years where they can do things that cannot be simulated classically. And yet at the same time, we only have this narrow set of problems that we know they’ll be useful for. From my perspective, it’s like we will only unlock that full imagination space for new applications of quantum computing when we have a broader set of people engaged. And I don’t think this is a wild hypothesis. I mean, if you’d asked a classical computer scientist in the nineteen forties or fifties, what will a computer be used for in 10, 15, 50 years’ time, you know, none of them would have said we’re gonna have something like a video call technology, or maybe some of them would have done, but probably very few would have predicted Twitch or generative video. You know, the broad application space only really unlocks when you have the public engaged, when you have a broader spectrum of people ideating about what this technology could be used for. And so for us at Moth, for me, I think it’s let’s break down that barrier and start a new phase of discovery.
Jeremy Gilbertson: This is almost like a Roblox or a Unity engine using quantum in the mix to kind of play around with. Is that close? Is that a bad analogy?
Harry Kumar: No, it’s not a bad analogy. I think we aren’t confined to a single vertical or creative technology space. So, I mean, we often make the comparison to like Epic games. So the Unreal Engine comparison is good. You know, Epic games make Fortnite, but they also deliver and ship to the world Unreal Engine, which is a tool that can be used to make games. So on the one hand, they build the games. On the other hand, they ship their own games, but also many, many more people ship games that used Unreal Engine. So in that sense, there’s a comparison to be made. What I’d say is with Atlas, this platform we’re delivering and the API that comes along with it, we built it in a way that can be so generalized to different domains that you could take something that we’ve built, in fact, let’s talk about the specific thing we built used for Quantum Backrooms, quantum graph. Now this process could be applied to a music plugin. You could re-sequence music using this quantum process. You could also apply it to images or video or you could even do something really wild and say take a sequence of DNA and run that through, set that as your end parameter and run the process on that. We want to allow people to be as creative as possible. And so we’re trying not to constrain people to specific media domains, specific applications.
From an entanglement shader to a reverb plugin
Jeremy Gilbertson: Perfect. Perfect. That’s it. So my mind’s kind of rolling now. So I think about like you mentioned music. I think about DAWs and Logic and Pro Tools and all of that. And you have plugins in those that affect the audio in a certain way. So what does quantum do? What could quantum do to make a reverb plugin more fun and more engaging and more, I don’t know, what does quantum do to a reverb plugin?
Harry Kumar: I mean, you know there’s so many different ways you could attack that. And in a way, my call to action to the, you know, to the prolific vibe coders in the world, as well as the plugin creators, would be go tell us how we can make a reverb plugin that uses quantum dynamics. I mean, there was I think someone on our science team recently was looking at one of these engines we’ve built for 3D design. Now, this is a kind of shader plugin. I don’t know if you’re familiar with Blender, for example, you know, 3D software. Now, a shader is used to kind of render the texture of a surface in 3D environments. Now, one of the things we’d observed is that actually it’s really hard to render iridescence in a way that is like kind of physics real or accurate to that iridescence in the real world. You know, think, you know, oil in a puddle. You know, when you’re at a gas station or petrol station for us Mark and you’re seeing that shimmer in the puddle that’s coming from that strange combination of oil mixed in with water. It’s actually quite hard to produce. And so we built this shader application, we call it the entanglement shader, that calculates the coefficients of light as they hit this complex material and produces this iridescence effect. Now we built that as an app that you can use in Blender, but on our platform, we’re also gonna provide it as what we’re calling a core engine, which means that we leave the processing, the pre-processing, the post-processing to you, the user. And someone suggested the other day that why don’t we try and turn this into a reverb plugin? What about using a shader, a tool that’s been designed for a shader and trying to turn that into a reverb plugin? Now that is the exact kind of experimentation, I think, that leads to the invention of the 808, that leads to the invention of the MOOG. I mean, if we’re talking about the history of music and the history of technology and culture, those kinds of moments where someone goes, but what if we unplugged that and plugged it into here? I guess what I’m waiting for right now, we spoke about music in the preamble, I’m waiting to see Dylan goes electric. You know, that moment where…
Mark Fielding: Yeah.
Harry Kumar: …someone super creative takes a quantum computer, doesn’t have any experience or background, and starts plugging things in like that, using our platform, runs it on whichever hardware provider they want. We’re trying to be as hardware agnostic as possible to enable that, and starts making really cool things happen.
Jeremy Gilbertson: Harry, have you ever run across the name Richard Devine before?
Harry Kumar: I have indeed. In fact, it’s really funny you bring him up. He actually came up in conversation yesterday, super talented guy. And our CTO, Spencer, is based out of New Lab in New York. Spencer previously founded a company called Physical Synthesis, and they built acoustic synthesizers, one in particular called Cicada, and I think, if I’m not mistaken, Richard Devine worked with them on that project.
Jeremy Gilbertson: Yeah, so Mark, he’s a mod synth genius, like well-known creator. And there’s also a guy, Sam Cape, who I need to get you introduced to. So Sam, this is you can cut this out later, Mark, if you want, but I gotta say it. So Sam created the cryptographic math that in twenty dollar bills in the States, there’s that little stripe that makes the twenty dollar bill valid and non-counterfeitable, but he’s also a synth guy and just a wizard. Like these dudes should know who you are and they should be playing in your sandbox, man, for sure.
Harry Kumar: I mean that would be amazing.
Mark Fielding: Yeah, I mean.
Harry Kumar: It’s exactly the sort of engagement we want. I mean, and I’ll make the point again. I think the quantum industry needs this. You know? I think the quantum industry, yeah.
Can you hear quantum randomness?
Mark Fielding: Dylan goes quantum. Dylan goes quantum. Is this, so listening to you speak, a new 808 or 303? So this is a new musical instrument in a way because you can get a reverb pedal and you can alter a reverb pedal in other ways, but this is giving creative musicians the opportunity to discover something new. A new way to look to play. Would it be right to say harnessing the randomness, the serendipitous randomness of nature? I mean, is that something that they could do? I mean, what are they getting that they couldn’t get elsewhere?
Harry Kumar: No, this is one of the questions we get most often, and it is, you know, there’s a few ways to answer. So the first thing is we will have like an entropy engine and a QRNG engine on our platform. And I actually think that will be attractive to a lot of creative people. Musicians, for example, and producers that we engage with, one of them is a long time collaborator called Illar. And I know that Illar will reach for the QRNG sequencer that can be built using our platform and will see, you know, will input a sequence of notes and see as those get reshuffled and randomized using real quantum computation. Now, the thing is, I think that will be really exciting to creative people, but we have to remember that if you presented all three of us with a QRNG versus a classical random number generator, we wouldn’t be able to tell with the naked eye the difference between those two things. And you could make the same argument for using quantum in that way in music production. You’re not going to be able to tell the difference unless you’re told. Now, you could argue that that means like, why should we do things with quantum computers today? Because we know that the quantum advantages will come with later iterations of hardware, with later generations of hardware. But actually, we believe that you can do already much more than just randomness or leveraging the non-deterministic nature of quantum today. We think there’s a lot more there than just that. And that’s the first thing. But the second thing is that getting creative engagement now, even if it’s doing things that you could replicate classically, is probably arguably one of the only things you can do with quantum computing today, running on real quantum hardware today, where someone finds it valuable. I mean, this is I think a really key part of the Moth thesis. If you look back at 2016 and AI development at that point, you know, we’ve just seen AlphaGo from DeepMind, Google Deep Dream had come out. I don’t know if you remember the kind of images that were being produced between 2016 and 2020. They were like seriously cursed images, you know, and you’d prompt a generative AI model, you see the Google Deep Dream stuff, they look like something out of a bad trip or some kind of nightmare. And yet that was what attracted creative people to them. You know, a close member of our sort of closed alpha group is a guy called Terence Broad. And in 2016, he made a film called Blade Runner Auto Encoded, where he took, I think it was frame by frame, the film Blade Runner, and using machine learning, re-rendered those frames. And it created this really spooky film that looks like a machine’s nightmare of Blade Runner. Now, at that moment in time, AI was not commercially useful. It wasn’t, you know, no one was thinking I’m gonna drop my Photoshop subscription, gonna drop the Adobe Suite or whatever their creative tool production was. But, you know, OpenAI’s DALL-E did the same thing in 2020 or 2021. They released these images out to the world. They were super janky and weird. And yet a million people signed up to the waitlist to pay for that tool, right?
Mark Fielding: Yeah.
Harry Kumar: So we’re talking about a completely different kind of value and value creation. Now, some people could describe that as sort of novelty, and that’s one way of looking at it. I see it as the step that comes before the real commercial value of quantum computers. And I also see it as one of the most valuable things we could do today. And I guess personally, as someone who comes from a creative background, I’m really excited to see the kind of emergent aesthetics, the kind of emergent cultural artifacts that come from people playing with quantum today.
How the quantum industry sees Moth
Mark Fielding: It’s like pop art and punk and rock and grunge and jazz…
Harry Kumar: That’s it.
Mark Fielding: …and it emerged from these new forms that at the beginning were seen, were labelled alien. You said the word I can’t remember, what’s the quantum industry saying about you? How are you being received when you travel the world and you’re faced with the PhDs from IBM and IonQ and their peers?
Harry Kumar: Yeah. So I mean we’ve got really good relationships with the quantum hardware partners. And, you know, I think increasingly as we start to boast the biggest user base, I don’t think it will be long from now that Moth has the biggest engaged user base in the quantum industry, running probably the most number of jobs on a QPU out of anyone out there. I think that’s possible. I think that will come.
Mark Fielding: IBM has a lot, don’t they? On their they have like six and a half a million, a million or…
Harry Kumar: Engaged users is where I’d draw that caveat, but maybe we won’t reach that number. But the point is I think we will have at least one of, if not the most engaged active user base in the quantum sector. And I think the hardware partners will find that beneficial. I mean we’re running the same circuits, different circuits on quantum devices, on a plurality of quantum devices across modalities, we’re actually able to collect a lot of data and information that is valuable to the development of these machines. We’re also kind of testing and pushing them to their limits. And there’s a kind of, you know, I won’t call it an unintended consequence because it’s something we’re interested in, but it’s certainly not the priority of what we’re doing. But Quantum Backrooms almost served as like a benchmarking test. We ran this on devices from a number of different hardware providers. And by the end of that process, our team was able to identify the hardware used based on the levels that they produced. I think that’s a really cool, unintended consequence.
Mark Fielding: That’s very cool.
Harry Kumar: And I think over time, especially as we advance to the next generation of devices, you know, we’ll see that value appreciated even more so by those hardware partners. I mean, it’s really important to Moth that we’re as close to the metal as possible in that way. We need to be able to anticipate the upcoming generations of hardware, prepare for those. And like I said, you know, a lot of our team is focused not on making cool and creative applications today, but on uncovering and preparing for the most sort of impactful and valuable applications of quantum computing in the future. So being close to the metal is important for that.
Which modality makes the best beats?
Mark Fielding: I’m gonna ask you a technical question in a non-technical way. Which modality makes the best beats? Which modality is the best for any aspiring musician wanting to get creative with quantum?
Jeremy Gilbertson: Spin qubits, of course, just like the record player, man.
Harry Kumar: Do you know what? I even…
Mark Fielding: That has to be your… spin qubits has to be. Come on, that would be such a good advert.
Harry Kumar: Yeah, do you know what? I’m not gonna give an answer to that. And the reason is that I kind of want people to go and test that. I think my intuition is not valuable here. And the second thing I’d say is that we’re just starting to see this new generation of quantum applications with our first alpha users and, you know, playing with our platform. In the last week, there have been like I think six or seven new quantum music tools created that ping our API every time they run, that run either on emulation or on real quantum hardware. And I’m kind of keen to see where, you know, ask me that question in a month, and I might have another sense.
Jeremy Gilbertson: Can we ask…
Mark Fielding: Okay.
Jeremy Gilbertson: …you what type of qubits you’re using in your systems, or is that more of something you want to hold tight? That’s okay too.
Harry Kumar: Do you know what, I think I will not, I won’t respond to that one now, but again, maybe in a future conversation.
Pseudo-randomness and computing with nature
Jeremy Gilbertson: No, all good. All good. So let’s dig into quantum a little bit more. Help people understand randomness a little bit and how a classical computer treats randomness because it’s actually really not random. Like all classical is largely deterministic, right? But it can appear random. And explain what advantage the quantum version of randomness provides what you guys are building.
Mark Fielding: When can I have it on my Spotify playlist?
Harry Kumar: Okay, I love the follow-up question. No, look, I should say that my background is not as a, you know, a physicist or a computer scientist, quantum information scientist. So I’m not the best person to answer questions about explaining the underlying mechanics of these systems. We’ve got an incredible science team, an amazing engineering team led by our CTO James, who was mentioned earlier, our CSO has an amazing team as well. So they would be best positioned to give you good answers on that.
Jeremy Gilbertson: Yeah, so we don’t have to go into the hyper weeds of it, but isn’t there something there to demonstrate the value of what you’re doing versus what classical computing can do? Because the way quantum addresses randomness is different than the way classical addresses randomness. That’s what I was kind of getting at with like…
Harry Kumar: No, no, but I know I’m very happy to respond to that. Classical systems are essentially built, classical computers are built on human logic, on binary systems, on zeros and ones. And quantum computers are completely fundamentally different in that they are leveraging quantum mechanics, quantum phenomena, and a different system of logic. And you know, this underlying difference in the way that quantum computers are built and computed with gives rise to these different computational forms. And one of those is, you know, non-pseudo-randomness, real randomness. You know, classical computers can only produce pseudo-randomness. And until now, you know, there were other ways of producing real randomness. I remember hearing that Cloudflare has, you know, a wall of lava lamps that is used as their source of entropy. I thought that was cool. I don’t know if it’s true, but there are other ways of essentially leveraging and harnessing natural systems to produce randomness and indeterminacy. But a quantum computer has that baked in because you’re computing with nature. And I mean, this was what Richard Feynman in his lecture in the nineteen eighties described as the kind of inception of quantum computers. You know, if you want to simulate nature, if you want to compute like nature, you’d have to use a computer that operates on the logic of nature, and that’s what a quantum computer is. And I again, I think for the creative user, the mere fact that you are doing something that interfaces with a natural system that is not confined to the realm of human logic and binary systems, but is actually leveraging quantum phenomena, I think that alone, I mean I’ve seen it, that alone is enough to excite people and entice them into the quantum realm.
Mark Fielding: Love that. It’s such a huge thing to think about for us at Thinking on Paper and our audience. We speak a lot about creativity and where does that inspiration come from? And everyone is aware of Rick Rubin now. And creativity is kind of channeling the whispers of the universe with your antennas. And this is such an interesting way to think about what quantum computing can give to the creative act, when we’re just bombarded with the heavy duty science of quantum and what it can do for material science and what it can do for protein folding, to think about it in this way, is a breath of fresh air for me personally.
Is creativity inherently quantum?
Jeremy Gilbertson: All right, here’s my hot take. I totally agree. So we’re gonna get back together in like five or six years. And you’re gonna have Harry, you’re gonna have done so much with this and enabled quantum in such a way that you’re actually gonna write a paper and it’s gonna point back that the creative process is inherently quantum. And here’s the proof.
Harry Kumar: I love that. I mean, you know, there’s at some point we were interfacing with a researcher who was exploring the idea that music is inherently quantum. Now, rather than try and paraphrase his work, I’ll point you to Rakhat-Bi Abdyssagin who is a fascinating researcher. Whether or not that ends up to be true, I think it’s a brilliant question to ask. You know, what things out there are inherently quantum? Which things are, what are the quantum native problems to solve with quantum computing? But also what are the quantum native problems to start thinking about with a quantum perspective? I mean, quantum computing and quantum mechanics forces us to entirely reorient the way that we perceive the world, understand the world, and think. And if that process of unlearning, I like to call it unlearning, if that…
Jeremy Gilbertson: It’s a great word. It’s a great word.
Harry Kumar: …if that doesn’t give rise to new creativity and new forms of expression, then I don’t know what will. And there is something somewhat intuitive about the idea that we can encode creative meaning in some kind of quantum state. That’s one for sci fi, and that’s a comment from me and not the Moth research team.
The handover question: AI and the first draft
Mark Fielding: Creativity is at the forefront of everyone’s mind…
Harry Kumar: Yeah.
Mark Fielding: …as we get further and further into this AI world. And maybe this is the counter-argument, maybe this is the counter strike to that giving up of creativity. And maybe this could prove or this could be used to think about can AI really be creative and what maybe quantum could help with that creativity, maybe give it to the machines or make sure that the machines never get it. Which is my way of leading into the crossover question, Harry, because I’m aware that we are limited by our time, that entropy is approaching the end of Thinking on Paper, and we must close this out. But we will be back and we’ll take you up on your, and we’d love to take this conversation further because we’ve barely scratched the surface, I think, of, like it takes 40 minutes for our brains to start working. And then so the crossover question, the follow-over question, the carry-over question, sorry, left by our last guest for you. So our last guest was Nathan Silvernail, CEO of Planted. And his question for you was: which activity in your day-to-day life have you replaced with AI? And the follow-up to that, which is perhaps more interesting, what would you turn back or turn off that AI is being used for today?
Harry Kumar: Yeah, I mean that second part is such an interesting question, one that I would oscillate in my feelings towards. But to the first question, you know, I try not to let AI replace my thinking. That’s really vital. I think one of the, I forget who to attribute this to, but I read a blog post early doors, I think ChatGPT. And someone wrote a post about how they were really concerned about the death of the first draft. And it really struck a chord with me because I studied mathematics and philosophy at university. I remember staying up late at night, right up until the deadlines, writing my philosophy papers. And the first draft was always like 60% of the effort, if not more. And then you get that first draft and you start going through and you start then culling, you know, the bits that you can get rid of. You start rethinking aspects of it. But that first draft is where so much of the thinking goes on. And the concern at that time was that AI would replace the first draft, that you go, here’s the idea, or maybe you even outsource the original idea. Come up with a cool idea for an essay about AI and culture. And then that process of drafting that original idea, that human developed idea, is gone. And then it also renders the mind less capable of being creative, of being interrogative, of challenging ideas, and of thinking original thoughts. So I know that’s not an answer to the question. That’s what I don’t use AI for. What I do use AI for is like the more administrative stuff, you know, I don’t want to spend hours and hours and hours going through expense receipts, you know, the laborious things. And then I will admit I have been dabbling with vibe coding tools. I think it’s hard to not see Claude Code as the most significant development or the most significant app since Google search. I mean we may have opinions about whether it’s a good thing for the world or not. What I will say is that it’s impossible to admit that Claude Code in particular for coding things is an extraordinary development in technology. The ability for a whole new part of the public, you know, a whole new and you know, the majority probably of society to go from not being able to code, not being able to build new computational technologies to being able to. Now obviously there’s all those caveats of you know, someone generates an app that is a hundred thousand lines of code, more than it should be, because Claude Code’s not there, but I’m sure it will get there. And of course other AI LLMs are available. But that…
Mark Fielding: Yeah.
Harry Kumar: …to me is something I find exciting. It’s something I’ve started using, but otherwise my answer is the more administrative stuff.
Mark Fielding: I couldn’t agree with you more on the first draft and every time somebody posts about ideation and using AI for ideation, a little piece of my inner creative drive dies.
Harry Kumar: Mm-hmm.
Jeremy Gilbertson: I’ll echo that too. We talk about this a lot, Harry, on the show. And in fact, Thinking on Paper comes from a quote from William Zinsser that if you can write about anything, you can think. It’s thinking, you’re able to think using your hands on a piece of paper, the friction of stuff coming out of your head. And sometimes you gotta grind through the BS. Like sometimes part of the brilliance of the 60% of the first draft is like running up against this roadblock of like, man, this doesn’t feel right. And all of a sudden, boom, there it goes. And it’s brilliant. And I’m worried about that too. I’m worried about people losing that ability as well.
Harry Kumar: If I can pose a question to the next…
Mark Fielding: Please.
Harry Kumar: …the next guest. Related to this topic, I’d ask, do you believe you could enjoy a piece of media, music, art, film, that was generated solely by an autonomous system? I have my own answer to that, but yeah.
Jeremy Gilbertson: So clarification on that.
Harry Kumar: Yeah.
Jeremy Gilbertson: Fully autonomous, no one prompting it, just it somehow getting good at this from the loops, start to finish?
Harry Kumar: Mm. Auto culture. Auto culture is what I’ll describe it as. Could you be…
Mark Fielding: Take hold chat.
Jeremy Gilbertson: Okay, great question.
Harry Kumar: …a consumer of auto culture?
Jeremy Gilbertson: Great question. Harry, this has been a pleasure. Like Mark said, I think we’re just scratching the surface on some of this stuff. You know, next time Mark will tell you why he thinks consciousness is the wave collapse. But we’ll leave that for another time.
Mark Fielding: I didn’t say it is. That was just speculation.
Jeremy Gilbertson: But this has been a lot of fun. It’s been a…
Harry Kumar: We must speculate. It’s been so much…
Mark Fielding: Yeah.
Harry Kumar: …fun. Thank you so much for having me. And I’d love to come back another time and scratch even deeper into the sort of dive into the quantum realm if I can.
Where to try Atlas
Mark Fielding: Where can people play with your quantum machines?
Harry Kumar: So platform.mothquantum.com. That is our platform. It’s in an alpha user phase right now. It’s called Atlas. You’ll find it on our website as well, mothquantum.com. We’ve opened it up to a wider group of alpha users. You can right now, if you install it, run things on an emulator, bring your own key. And to the people who are using it the most, we’re offering free compute credits and we’re also on the brink of announcing a few things like an open call for creative projects and we also have a hackathon online next week, the Moth Hack, with some challenges and prizes up for grabs. Make your best image with a quantum computer, make a quantum game, you know across the spectrum, make it weird. Please sign up to the platform, sign up to our Discord and, you know, fire away what you’ve been making. We want to see all the creativity coming from the community.
Jeremy Gilbertson: Outstanding.
Harry Kumar: Amazing. Well thank you so much for having me.
Mark Fielding: Thank you, Harry. It’s been a pleasure. Thank you. And until next time, be disruptive. Stay curious.
Jeremy Gilbertson: For Pete’s sake, keep thinking on paper, the first draft is the most important.
October 9, 2026