# Pasquale D'Silva and Maximizing Creativity using AI

by [kirupa](https://www.kirupa.com/me/index.htm) | filed under [Interviews with Creative People](https://www.kirupa.com/podcast/index.htm)

[Pasquale D'Silva](https://twitter.com/pasquale) treats AI less like a trend and more like a new kind of leverage. In this conversation, he explains how AI tools helped him move deeper into engineering, rethink what it means to be a creator, and imagine richer pipelines for animation, products, and storytelling. The mood is curious and optimistic without losing sight of the real cultural shock these tools are creating.

Watch the interview: https://www.youtube.com/watch?v=v3dQPn3_CQM

A conversation with **Pasquale D'Silva** | 46m

## About this conversation

What makes this episode compelling is how personal the AI discussion feels. Pasquale is not speaking in abstractions about disruption. He is describing a concrete shift in his own workflow, where tools like Copilot and language models helped him move from limited coding fluency to building far more of his ideas directly. That matters because it turns AI into something more specific than a headline. It becomes a bridge between taste and execution.

That bridge changes identity as much as output. Pasquale talks about how the old lines between artist, designer, and engineer are getting softer, especially for people willing to learn across domains. Instead of pledging allegiance to one specialty, creators can now move more fluidly between visual thinking, prototyping, implementation, and iteration. The excitement in the episode comes from that sense of unlocked range rather than from automation for its own sake.

The middle of the conversation gets practical about research, prompting, and the kinds of projects that become possible when the setup cost drops. Pasquale makes a convincing case that AI can help people start work that previously lived in an awkward middle zone: too small to staff properly, too technical to do casually, but still worth building. He also emphasizes that language itself becomes a core skill. Clear context, precise requests, and good judgment are what make these systems useful.

From there, the horizon expands into real-time production, personalized media, character design, and simulation-heavy storytelling. Pasquale sounds energized by the possibility that game engines, generative tools, and conversational systems could reshape how animated experiences get made. Even so, the conversation never treats change as frictionless. It recognizes the artistic backlash, the social uncertainty, and the strange speed of the moment. That tension is exactly what gives the episode its spark.

## What you'll hear about

- AI becomes most powerful when it shortens the gap between imagination and execution.
- Creative identity is getting broader than artist versus engineer.
- Prompting is really a communication skill under pressure.
- Lower setup costs make more niche creative projects worth attempting.
- The cultural response to AI may matter as much as the tools themselves.

## Jump to a topic

- [0:00](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=0s) **AI as a personal engineering tutor**: Pasquale explains how AI tools helped him become far more fluent in building software directly.
- [8:00](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=480s) **From artist versus engineer to creator**: The conversation shifts into how these tools blur older labels and open a broader creative identity.
- [12:00](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=720s) **High-leverage tools and blurred roles**: He argues that stronger tools will let more people move fluidly across design and implementation work.
- [16:00](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=960s) **Projects that can finally start**: Pasquale describes the kinds of ideas that become realistic once the technical barrier drops.
- [20:00](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=1200s) **Prompting as a real skill**: They dig into why language, specificity, and context are becoming central parts of creative and technical work.
- [24:00](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=1440s) **Real-time animation and game engines**: The discussion explores how AI could combine with tools like Unity and Unreal to reshape production pipelines.
- [28:00](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=1680s) **Personalized entertainment**: They imagine a future where shows and experiences adapt far more closely to each viewer’s taste.
- [32:00](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=1920s) **Building believable characters**: Pasquale shares experiments using language models to prototype the rules and personality of fictional characters.
- [36:00](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=2160s) **Everyday magic arrives**: The tone turns toward optimism as both hosts reflect on how quickly impossible-feeling capabilities are becoming normal.
- [44:00](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=2640s) **Why art opened the door first**: The closing stretch reflects on why generative art triggered such a strong early cultural reaction.

## To learn more

- [Pasquale's Twitter](https://twitter.com/pasquale)
- [Pasquale's TikTok](https://www.tiktok.com/@psql)
- [Pasquale's Website](https://pasquale.cool/)

## Transcript

The automatically generated captions have been organized by speaker, lightly edited for clarity, and broken into paragraphs for readability. Names and wording may still contain errors.

### [0:00](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=0s) - Kirupa

The rise of AI- and ML-based assistance has had a major impact across a lot of sectors and the kinds of content people create, but one of the biggest areas has been visual design, visual art, and content creation that lives purely in the realm of pixels and looking really nice. In this interview, I talked to Pasquale D'Silva, a close online friend I've known for over 20 years—one of the most talented, hard-working, and probably the best animating designers I've ever run across.

It's going to be pretty exciting to get his take on how all the changes we're seeing thanks to the rise of artificial intelligence will change your life, my life, and lots of others. It's going to be fun. Let's get started. All right, Pascal, great to meet you again after having known you for almost 20 years now—online a few times, offline as well. Dude, probably two decades is nuts.

### [0:57](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=57s) - Pasquale D'Silva

Yeah, it's been a fun ride.

### [1:00](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=60s) - Kirupa

But, you know, glad to catch up with you again and hopefully share a lot of the fun stuff that makes you an interesting person, both as a designer and as an online individual. So glad—great to have you on board. Tell me a little bit more about yourself. What do you do, where are you, and what makes you tick?

### [1:18](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=78s) - Pasquale D'Silva

Oh man, I feel like the last 20 years of knowing you has been such a roller coaster of personal, creative, and professional experiences. Where I am right now is in such an almost manically exciting state that I think might be shared by quite a number of peers in our industry. For me right now, the exciting keywords are artificial intelligence, real-time graphics, characters, animation, and just kind of everything that touches that. I've been going through an almost personal Renaissance moment of being able to move between so many different disciplines.

I think we briefly touched on this when we had a little catch-up when I was in Australia last month or the month before, but I've been chasing a white rabbit for pretty much my entire hobbyist and professional career. It's like the last Infinity Stone, which is engineering. I've always found that to be the most difficult of all the available disciplines. A lot of other things have come easy, like learning animation and graphics and illustration, and maybe even business stuff and math and all the other subjects that sit around it or could plug into it. I'd be able to get through that.

But as long as I can remember—even coming onto your forum back in the day and wanting to learn to engineer, wanting to learn ActionScript, and over the years starting and stopping on so many different programming languages and so many different environments and configurations—I would start up the ramp, then get burnt out and give up almost, or just throw in the towel and be like, bro, I can't engineer. I'm really good at all this other stuff. I should just thank my lucky stars and compromise in a way—hire engineers to do this work that I wish I could zero to one.

But the big breakthrough that happened when I had a lot of time to myself in Australia was discovering that the best possible teacher for engineering, for me, was artificial intelligence. I've been surfing this wave certainly long before it hit the more mainstream news cycle, almost feeling like a crazy person because I was getting so much value from using Codex and GitHub Copilot. Just having another brain underneath my fingers has allowed me to become fluent enough, at least inside one environment—Unity with C-sharp—to build things all the way through end to end.

That's something I've wanted to be able to do forever, and it came together from pretty close to zero C-sharp fluency to being able to build a full-stack thing in two and a half or three weeks. It's like in The Matrix, where Neo gets the cord plugged into the back of his head and he's like, 'Oh, I know kung fu.' I feel like I finally had that with engineering, where by no means would I consider myself a good engineer, but definitely good enough to work with myself now and just do the thing all the way through.

And I gotta say, man, that feeling of being able to do every part of the thing is maybe one of the most satisfying feelings. I feel like I could die right now and I'd be happy. Everything moving on from here is bonus round. It's been such a utopian mental space to be in, I guess.

### [6:30](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=390s) - Kirupa

Yeah, I know, and I definitely want to dive deeper into your work in the AI/ML space. But before we even get there, do you consider yourself more of a designer or an engineer? It seems like designer based on what you mentioned, but I just want to get your take on it.

### [6:45](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=405s) - Pasquale D'Silva

Dude, this has been such a funny one too. I was explaining to my mom and dad, being back home in Australia, the feeling of the breakthrough that was happening. You know how there's that whole classic story of coming out to your mom and dad about your sexuality, and they have to accept it? I was like, Mom, I got something to tell you—and it's not to do with that, but it would feel like that conversation. It's like a parallel version of that conversation. I've always identified as just an artist and struggled with the engineering piece. I wanted them to understand how big a deal this is for me.

I was finally able to get comfortable with saying, “I'm cool calling myself an engineer,” even if I'm not a great engineer or fully there. And if I can call myself an engineer, where is the line between being an artist and an engineer? Is that even a thing anymore? At this point, I don't know where that line is. Code finally feels like picking up a pen, a pencil, a brush, or a guitar, where I can noodle with it and jam in a pretty fluid way. So, how old am I? Almost 34.

I don't know, for 25 years or more—I would have said I'm an artist, or committed to just being an artist. And there used to feel like more of a dichotomy: artist, engineer; designer, engineer. But in discovering this new part, it feels like a real part of me that was waiting to be unlocked. So I don't know what the word is anymore. It's getting weird. What would you call yourself?

### [9:17](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=557s) - Kirupa

I call myself a creator.

### [9:19](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=559s) - Pasquale D'Silva

Yeah, creator is probably the word. Not to toot our own horns, but there was almost this Greek or Italian Renaissance vibe back in the day where there were these homies who were like, I'm a mathematician, I draw, I paint, I do this, I do that, I engineer the things I create. Those people came along once in a blue moon. But I think we're about to enter an era where more of those types of people emerge and it becomes pretty normal. Some of us dorks just happen to be getting there a little earlier, but I sense that those labels and definitions are going to start to blur even more.

### [10:22](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=622s) - Kirupa

No, I completely agree with you. Especially when you look at just in our lifetime—the time you and I have known each other—back then, if you were a designer, you had very specific tools that were very siloed in what they did. And if you were a developer and your output was something that worked in the browser, worked on desktops, maybe installed somewhere, that was a very different set of tools and a very different way of approaching the problems.

I think when we connected on Flash, the tool that allowed you to do both visual design and programmatic interaction, that was one of the first apps, at least to me, that blurred the lines between what fell very strictly into the bucket of a designer and what fell into the bucket of a developer. The tool was what bridged those worlds together. But as you're alluding to, the whole world of AI and ML is bridging even more of those gaps where it's going to be about the tools. It's about what you're actually interested in and focused on.

If you care about building great apps and you're a developer, great—you can now rely on and delegate a lot of the design side to some computer system. Or if you see yourself as a designer and care about the visuals, the environment, the tone, and the feeling of what you're building, you can delegate all the technical things to some AI- or ML-based assistant and be almost as successful as if you had spent years learning it or hiring someone to fill that role for you.

### [11:53](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=713s) - Pasquale D'Silva

Yeah, I think that's spot on. There was a generation of building software, or of working inside the software scene, where you really had to commit to a side. You're like, I'm going full send on being a designer and owning my craft, because it requires so much work. The grunt work is a lot. It's the same thing in engineering: you have to absorb a lot of domain knowledge, and that leaves you very little time to put those cycles into design.

But now we've got these super high-leverage tools emerging, especially with AI, where you can start to straddle the lines or flip between those domains a lot more fluidly and work on problems more holistically, which is just so freaking exciting. The perspective the overall population working in software is going to get from this is going to be nuts, because you'll have people who are less one-note, or less like they're pledging allegiance to a single party. If you're a full-send engineer, the classic story is that hardcore engineers often don't have the best design sensibility.

And super aesthetic, artistic designers often have very little understanding of how the underlying systems work when you get closer to the metal, so they're operating kind of out of touch with the other end of reality. But now there's such an amazing plane opening up right now.

### [14:18](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=858s) - Kirupa

Yes, very much. In many ways it's almost a case where there are things you care about and things you don't care about. It used to be that to get something out the door, you had to spend time doing both: the things you cared about and the things you didn't care about, and you had to struggle through that to get something out. What we're seeing now is that we've gotten to a point where it's almost a luxury.

You can use and rely on systems, whether it's AI-based or not, where you spend more time doing the things you're uniquely qualified for or think you're very good at, and then delegate all the things you don't really care about to something else while still providing a high-quality output. In the past it would have been something that looked really good but functioned poorly, or something that looked horrible but functioned really well. Now you can kind of get the best of both worlds as a single individual. You don't need a team of people, you don't need to crowdsource, you don't need to rely on a community to make this work.

You have a lot more power at your fingertips, and I think that is going to open a lot more doors for people when we look two, three, or five years down the road at how they're using these things to create even better things.

### [15:26](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=926s) - Pasquale D'Silva

Yeah, I think individuals are now getting access to a conduit that can transform them into like 10x versions of themselves. I've already been running a handful of experiments by having my hands in these tools and experiencing that firsthand. Take visual tools for concept art. In the past, would I have hired a concept artist to do a tiny bit of work? Probably not, because it's in this weird gray area where it's so small that it doesn't make sense to hire someone and do all the contract work and comms work with them, but it's enough work that doing it myself is frustrating. So I just wouldn't do it.

There are a lot of projects that can now start because you have so much more to go with. Or you get a big speed boost from the beginning—on the visual concept art side, on the research side for engineering. No longer do you have to trawl through Stack Overflow or Google and craft the perfect search term. I've been pushing the limits of ChatGPT as my engineering assistant, and before I sit down to build a feature, I do my research with it. I ask the questions I know are going to be hairy—like, is this thing possible? How would I hook this thing up to that?

What does this concept mean?—and it spits the answers out. I can skim them and be like, oh cool, I didn't have to smack my head on a keyboard and torture myself for two or three hours to get those initial answers. It's almost like I had an assistant drop the preliminary research on my desk and I can just go off to the races. It's unbelievable.

### [17:41](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=1061s) - Kirupa

How do you find the accuracy of it? Because we see a lot of things online where people ask it a very good question and get what looks like a very good response, but the accuracy is questionable at best. How do you work through that?

### [17:58](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=1078s) - Pasquale D'Silva

I think it depends on the context these tools are being used for. I tend not to throw total crapshoots at it. For example, I was working on a thing in Unity today. Unity is not great at handling system audio, at least recording. I'm working on a motion-control system for recording animation out of real-time puppets, but I also need to record microphone audio, and I was able to get a solution out of this thing pretty quickly after spending the time to prompt it correctly. I think there's an art to horse-whispering the thing. If you ask it too vague a question, I think the likelihood of getting inaccuracies is pretty high.

But if you're leading it down a path and setting it up for success, I think the chances of it being accurate are pretty high. So, you know, I was asking it like, okay, go for it.

### [19:26](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=1166s) - Kirupa

Which is exactly—it's almost the same as talking to a human being. The quality of the question you ask that person is the quality of the answer you'll get back, independent of whether they know the topic really deeply or not. So from that point of view, I think it's almost par for the course between humans and AIs.

### [19:44](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=1184s) - Pasquale D'Silva

Yeah, I think that's right. There's almost an emotional and communication intelligence one has to have to talk to the AIs you need when you're working with people anyway, except now we've got this incredible brain, this being that will work for you tirelessly. I think there's a new skill to develop, and it really is around this horse-whispering, or knowing how to speak the magic words in a way. It's not rocket surgery here. It's using the tools we've had for thousands of years: words, language, specificity, context, and reasoning.

I think if you can master that in this age we're entering right now, it's probably one of the most valuable skills someone could have. How can you communicate with this new superintelligence?

### [21:04](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=1264s) - Kirupa

Very much so. It's almost like when touch-screen devices first came out. People were using them, and now we're all using them by default. It's almost going to be the same case here. There will be more layers of abstraction built on top of conversational AI, but I think the world where we're going to have to figure everything out on our own, or ask someone for help and wait for someone to respond, may be limited as we look across. But one thing I'm curious about: in my view you're one of the best animators I've ever seen, and one whose work I've encountered for a long time.

What we don't have with AI/ML today, at least as far as I can tell, is a way for me to create really good animations using textual prompts. So I'm curious: when do you think it's going to happen?

### [21:52](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=1312s) - Pasquale D'Silva

It's a good question. I'm already starting to try to break ground with my humble resources and intelligence, and I don't think we're that far off. I was spending some time in LA on my way back from Australia earlier this month and hanging out with my good friend Eric Pringle, who likes to describe himself as a lazy animator. He's fantastic. He works on the show Teen Titans Go! At Warner Brothers as, I believe, the animation director. They're like 10 seasons in, and it's a super efficiently made show, believe it or not, produced in Flash or Animate. It's one of the few remaining popular shows in the world that's made with it.

Before that, he was on this incredible show called Foster's Home for Imaginary Friends. I don't know if you've ever seen that show, but it was super innovative for its time. It still holds up so freaking well, and it leveraged a more puppet-style animation process. That's something that's part of Eric's philosophy, and it's definitely worth taking a look at some of his animation breakdowns he'll post on Twitter. He'll do rig run-throughs, or what we'll call poke-arounds. He tries to draw as little as possible. He told me that when he got his way into the animation scene back in the day, he wasn't a master draftsman or anything.

In fact, drawing was not one of his strongest things. He just worked out how to be really efficient with the tools and how to communicate images with systems he would set up. So we were talking through what happens post Flash-kicking-the-bucket or Animate-kicking-the-bucket. Where are people trying to develop these more high-leverage production-technique types of workflows? A lot of the shows produced today are still done the old way. For the most part they're hand-drawn, and people are generating things close to frame by frame or using even classical tweening.

But when we look at the video game world inside Unity and Unreal Engine, and at real-time production, there is such a ripe space for simulation work to happen. As I've spent more time in these tools and realized how much horsepower is behind them, it feels like these tools are asking to be used for this. Honestly, I think all the Lego blocks are there. I wouldn't say it's trivial, exactly, but the right group of people and resources put behind a thing could definitely start making tools where you could write a sentence or two into an engine and it would spit something out.

One thing we've been interested in trying is looking at the bars of acceptable animation we see today. A classic one people like to measure against is South Park. It's famously crude, but the writing underneath is great. And not only is the writing underneath it great, something that makes South Park hit is the rate that it comes out. Because it comes out so fast, and because they're producing episodes in under a week, they're super plugged into the zeitgeist of meme culture and what's going on in the news. That's a superpower South Park has. So how can you speed South Park up even more?

It was a big deal for them to take the cutout animation they were doing in the first couple years and move it over to Maya. I don't know if you know they use 3D software to make South Park, but Maya is a pretty heavy-duty package used in major 3D studios. It was nuts for them to build all these rigs and implement them in Maya, but the characters are so modular, made of such simple parts, and the animations are crude. As an audience member, no one's really focusing on the aesthetic.

So we were thinking: man, how hard would it actually be to build a South Park scene generator, even if the stuff that came out was crunchy? Could you create a thing where you say, like, Butters and Kenny fight over shrimp and just see what comes out? Could you synthesize the voices, put the puppets into a scene, and have the camera flip between the characters as dialogue is being delivered? Could you hallucinate a background off a bunch of training data and then have a minute- or two-minute scene that is borderline watchable, even if it wasn't excellent? It's a starting point.

Imagine if you, as a director or writer, could test a bunch of ideas out and play them back. Then you're like, you know what, the shrimp scene ain't quite it. But if I tweak two or three things, there's something there. I think we're there now. We just need to put the pieces together.

### [28:12](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=1692s) - Kirupa

No, you're absolutely right. What has opened up as well is a whole world of personalization. Imagine that in the future we can generate in real time some of these episodes, some of these shows, and with game engines and 3D quality getting so much more advanced, maybe even realistic-looking TV shows that are personalized based on your tastes. With the social graph and so many of our interests, hobbies, and dislikes being available online, it would be very easy for someone to create a virtual model of what you like.

Then you could have a TV show, or anything really, in real time optimized for what you care about and available for you to binge-watch, all without any human being, for the most part, being involved.

### [28:57](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=1737s) - Pasquale D'Silva

Yeah, I think we're about to enter a new age which is basically the simulation business. It's moving past linear storytelling. Film is a technology on a reel that plays back linearly because it's arranged linearly out of a projector, and then we eventually moved that over to digital. But video timelines are still sequences of frames. Because we can do stuff in real time, and in some cases mutate it faster than real time, I'm certain we're going to see more of this happen. We already see it in games. There's a reason people have such an affinity for playing games: it's an opportunity to spend more time in simulations.

We're only scratching the surface on console and PC games. That's one way to interact with a simulation where there are goals. But there is also stuff that's a little less interactive and a little more passive. There's stuff that's purely for entertainment, stuff you could watch with your pals or consume with your pals. I think we'll definitely see it in written work first. I'm probably getting the author wrong, but I believe James Patterson is a very prolific author who writes many books a year.

I think the way he does it is that he writes the first paragraph in a chapter and the closing paragraph in a chapter—almost like keyframes an animator might specify—and then hands it off to other writers to fill in the details, the tweening part of it. I can totally imagine that being something ChatGPT or any of the current text-based services could easily solve, and that in many ways is the precursor: personalized books optimized for your mood and your interests, leading eventually to the video or animated version of that. That would be almost one step beyond what Netflix has been experimenting with.

I think it's called Bandersnatch—the interactive TV show where depending on what option you choose, the rest of the movie changes, kind of like those old Choose Your Own Adventure books. I think this is going to be done at scale with AI. I don't know how deeply you've gone into playing with the large language models out there, but ChatGPT is probably the easiest way to experiment with it, and you can already start to feel into that if you craft your prompts well enough.

Before ChatGPT was available to the public, I was playing with another company's large language model that they gave me access to, and I was using it to prototype and test the rules of a personality of a character I'd been working on. I'd already written six or seven paragraphs of lore about the character. You can say this character comes from a lab, it likes joking about Shrek, it feels this way about this. If you create those keyframes of character, you don't have to specify everything, but you're outlining the shape of its mind and its thoughts.

Then you tell these large language models connected to chatbot stuff, okay, I'd like you to role-play as this character. Here are all the things we know about the character. Give it the payload. And it's like, got it, okay, I'm this now. I'm a horse. I'm a horse that loves roller-skating. Then you can start poking at the character, which is so fun, because you can say, hey horsey, do you want to go to the rodeo today and eat ice cream? And it's like, no, I don't have a mouth—or whatever. It checks itself in a way. You start to get these sub-simulations of almost sentient-feeling characters.

Then you can prototype stuff like: okay, now there's another character in the universe that you may bump into. What would happen if you put those two in a room? Write a script about it. Chat to all three of them in a chat room. It's actually pretty crazy that you can virtualize a bunch of this stuff within those models too. Did you see any experiments where people were effectively spinning up virtual machines inside the models? They're like, pretend you're an instance of Linux and I'm inside your terminal. What's in the home directory? Run this command. What would happen? It was role-playing like it had a computer.

### [34:58](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=2098s) - Kirupa

This opens up a whole new interesting world. One of the podcasts I listen to is Lex Fridman's, and he had Todd Howard on. Todd Howard is the creator of many famous games like Skyrim and Fallout, which in many ways are simulations—whole worlds where you have individuals and characters whose own lives are happening in parallel while you're doing your thing.

One of the things he explained is that they had to do a lot of work to make these characters less realistic sometimes, because otherwise you might have these non-playable characters in the background having conversations, picking up really powerful weapons, and just doing whatever they do, which might be counter to what you want to do in the game.

This is all getting to the point where what used to be theoretical and very sci-fi and futuristic—that our computers would have a life of their own and do all these activities—we're starting to see large examples of that in virtual video games specifically, which is almost the precursor to seeing it in more everyday situations.

### [36:07](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=2167s) - Pasquale D'Silva

Yeah, man, I think we're entering an era of magic now too. It's like we just discovered fire for the first time. Dude, I got blazed last night and rewatched the most recent Transformers movie, which is not that good, of course—but you gotta watch it. It's not about whether it's good or not. You just gotta watch it because you've seen the other six or seven already. And honestly, sometimes I like smoothing my brain with those movies and enjoying hundreds of millions of dollars of the best special effects on screen at once.

But I started running through the process of, okay, maybe like five years ago I would have thought this was ridiculous. I'd suspend my disbelief. To watch Transformers successfully, you have to believe that an alien race of robots came in from Cybertron to Earth. That's a big freaking stretch. But fast-forward to now and look how quickly things are evolving. I had this very high moment last night where I was like, wait a moment, something like that almost seems feasible. Like, it could happen.

What if an AI escaped, somehow figured out freaking time travel, or built a race of metal-based organisms using some technology we don't quite understand, and then they crash back into Earth? Things we thought were so ridiculous have started landing in culture as real. Things that seemed like magic are now within reach. If the promise of these things continues to go the way it does, reality is about to get even weirder.

Outside of just the stuff we're seeing computers push back to us, people are going to have these weird higher-order ideas about how to affect the atoms of reality now, because they're going to get all these crazy universe cheat codes out of this. Surely the physical world is going to go through some mutations that we did not see coming.

### [38:55](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=2335s) - Kirupa

You know, we spent a lot of the 1800s and 1900s predicting what the year 2000 or the year 2020 would be like, and we were off by a wild margin. The reality was not quite as grandiose as people imagined it would be 50, 60, or 100 years ago. But the funny thing about what we're seeing now is that I don't think anyone imagined the future would look like this, and what they did imagine was far more conservative than what the reality is, or even what we're seeing today.

The idea that large language models, and all the related follow-ups, can recreate so many things without humans being involved, and do it with reasonable speed—I don't think anyone really expected that to be happening, at least not at this speed.

### [39:39](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=2379s) - Pasquale D'Silva

Yeah, I think it's also forced everyone's hand. It almost slapped them awake in a beautiful way and said, yo, turn on your imagination now, because you've got this imagination engine in front of you. All the stuff people maybe used to be naysayers about is becoming possible. There are definitely people freaking out about it, but I have such an optimistic view of what's going on in a way that, honestly, I didn't even feel a couple years ago when there was a lot more rat-race energy around the crypto boom. That felt like so much smoke and mirrors, like people trying to dunk on each other and get one over on each other.

This just feels like such a gift. We've been stuck in computing for so long, and it felt like we could almost see the edge of that reality. Now it has opened up into a space where no one knows what the hell is going on. Even the best people in it are unsure themselves about what's coming. You know that Two Minute Papers guy? Have you seen him before? The guy who's always like, 'What a time to be alive.' I've been watching that dude's YouTube videos for as long as the channel has been around. That's one of his catchphrases.

He'll show you a crazy piece of technology and be like, what a time to be alive. He's got a great accent and optimism, but it's true. It's crazy that we actually get to be along for the ride of this stuff. We've talked about there being something that would one day touch down in our culture and give us some hope or some answers, and it's just unfolding in front of us right now. Every day is pretty exciting in this space. You just can't keep up with it.

### [42:20](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=2540s) - Kirupa

Yeah, I think what has also happened is that a lot of us have been burnt by upcoming new technologies fizzling out very quickly because they didn't have the longevity they promised. For me, chatbots are always something that comes up. We've been talking about chatbots for decades now. Every few years it's like, chat technology is going to revolutionize how humans communicate, and so on. Every single time, that hype would basically fall back into nothingness because it didn't live up to it, or people didn't actually want to talk to something by typing and so on.

But with this model, the initial hype is continuing to be relevant and is carrying forward as well, so it doesn't feel like it has that same risk of fizzling out. I think you're seeing a lot of skeptics—me included, for the most part—looking at this and going, okay, this isn't just all smoke and mirrors. This isn't just all flash with no substance. There actually is something to it, and I think there's a lot to be said about that.

### [43:19](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=2599s) - Pasquale D'Silva

Yeah, man, I think the way it's unfolded has also taken a lot of people by surprise. I think even Sam Altman was talking about this. The commonly accepted projection of where AI would go was that it would replace blue-collar jobs first, take some white-collar jobs last, and the very last thing at the end of the line that would get its feathers ruffled would be artists, because even though they're starving, it's still kind of one of the ultimate privileges to have.

Then to watch it happen in almost the exact opposite order, where it's like, oh wait a moment, these computers and these models—the generative art backlash aside, which is obviously happening right now—for me, as an artist who appreciates both sides of the argument, what's happening with generative art is so compelling and interesting. Who would have guessed that the first thing to start making waves would be computers making art? That seems like the wedge that cracked things open. Then it was language and poetry and all these other things, and not, like, cars being really good at driving themselves. I don't think we've seen much progress there.

So on the one hand it's like, engineers, you've kind of got a fire lit on your ass now. Artists, you're going to have to evolve. It's been a funny cultural rug pull, but I think it's super healthy because it's shaking everything up. It feels so different from how other technologies have come in as trends, where you can tell the incentive is people want to get rich, or this thing would benefit a particular shareholder, or it's just fashionable right now. This feels more like the genie is out of the lamp. It's got a natural underlying energy to it.

I don't think it feels like there are individuals working in the background to coordinate a scam or run a pyramid scheme. It's more like, no, everyone's tuned in to an energy right now where cognition is evolving as a group. I don't know how it feels for you, but for me I just think it's undeniable.

### [46:23](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=2783s) - Kirupa

Yeah, no, I think there's a lot to unpack there. But I know we're also approaching our time here, Pascal. I think we should have you back on again to talk about the art piece specifically. I think we covered a lot of ground on AI/ML in general and how it's going to transform society and at least our day-to-day lives, but I definitely want to dive deeper with you into the art piece. I think that might be a part two at some point in the future.

### [46:49](https://www.youtube.com/watch?v=v3dQPn3_CQM&t=2809s) - Pasquale D'Silva

Oh, I'm totally down. I'm totally down.


[Browse all Interviews with Creative People](https://www.kirupa.com/podcast/index.htm).