TL;DR :: Artificial intelligence promises to completely transform every creative workflow. What does the need for creative talent look like in the game industry?

If we look to the past, we see three potential archetypes.

Notice the pixel art game level above? This image took 26 seconds for Stable Diffusion XL to imagine from my written prompt. Considering how fast AI is moving, it won't be long before I can write and manifest feature film quality shots. Game engine AI’s will produce, code, and make art for me to assemble.

Hard to accept, I understand, but AI has arrived.

Jeffery Katzenberg has indicated that 90% of movie production can be replaced by AI. Barring the IP and ethical data concerns with these models, that disruption figure will apply to every creative production endeavor.

In considering this rapidly evolving toolset, I give much thought to what the realities will be for "the generative game developer.”

For this essay, let's focus on the archetypes of talent.

“By talent, which is a word I hate, I mean … the ability to be productive.”

--- Gabe Newell, Cofounder Valve Software

"Armed with generative AI tools and services, we will start to see more viable commercial games produced by tiny “micro studios” of just 1 or 2 employees."

Source :

The Generative AI Revolution in Games / James Gwertzman and Jack Soslow : https://a16z.com/the-generative-ai-revolution-in-games/

With so much getting abstracted by AI, what core talent will propel "microstudios" - tiny teams of 1-2 people creating commercially viable games, as predicted by Andreesen Horowitz above?

Looking back over decades, large specialized teams made rigid hierarchies. But startups build small, agile crews wearing many hats. As AI accelerates change exponentially, adaptability and skill blending will become essential.

Like the early days of film production, video games emerged from large studios. But formats like mobile gaming enable tiny groups fusing technical and creative strengths. A 100 person visual effects crew flattened into 10 versatile creators - programmers, writers and 3D artists in fluid roles. Looking ahead, digital workflows, seemingly engraved in stone for decades, must now demonstrate flexibility almost overnight.

If we think of AI not as new, but a rapid accelerant of what may has happened over 20 years, we can see a change in productive capacity--- a change that could continue to accelerate.

I think about three archetypal skillsets rising to prominence, especially from entrepreneurs and smaller companies. These may be individuals, or it might be that AI will require qualities of all three for every person.

For each archetype, it’s about problem solving for a specific area in developing a real time interactive experience.

Someone (not an AI yet) has to have a killer idea.

This person will stop at nothing to will their vision into reality. AI will do the calendars and bug tracking, but vision must come from human passion.

The designer is not just the kid in the back drawing Mario maps, but someone who has learned to pitch and excite. Most likely, as an introvert, they had to pitch and excite regardless of pain. A builder, who is coordinating technically and socially.

These rare types bleed for their creative projects. As a result, the person who defies the horrors of game development to create an idea against the odds, is a hero for any team.

This will always be valuable, regardless of AI.

The second type of archetype is the engineering artist.

For me, it's artists who understand not the art, but the workflow of assembling it. We often have artists who know one core set, but the ability to string together workflows across multiple software platforms is hugely valuable. The more technical the art, the more it tends to be because it is still developing, and the tooling is unwelcoming to use.

Narrow AI will exist as different software packages.

Software may give us some solutions, but the technical artist hacks the “unthought of.” There is a new problem to solve every day. Particle explosions today, tomorrow procedural bricks, NPC pathfinding the next. Lot's of things are unknown when you are imagining a virtual world, and the person who knows how to prototype it rapidly lays the tracks for the rest of the shop.

The technical artist will always focus on how to string it all together. Every day, they will need to reinvent and reimagine.

Frustratingly, I find most students pretend playtesting isn’t part of development. Games are crafted through play – testing with yourself and others is how we understand the operational ability of the game we created.

We get better at building games by playing them.

The game player plays endlessly, providing feedback to the team on bugs but also suggestions. They build mods and assets for others to use. They run tournaments and build active discord channels. They become your viral marketing engine not just through word of mouth, but actively recruiting new users.

This player archetype will evolve into a prosumer, taking literal ownership for maintenance of the game. They will document and teach new users. They theory-craft and debate online. Quality Assurance and Inspirational show-off will be combined.

Complicated workflows in game development, disrupted by AI, will likely focus less on a technical, but rather, human archetype skill:

• Producing Designer

• Technical Artist

• Community Player

I believe that AI will drastically reduce costs for all other production tasks.

Procedural worldbuilding tools, outsourced AI coding, project management, marketing, and infinite amounts of expansive content libraries will all be magically available. Platforms like Unreal, Niantic, and Google Maps may soon digitally replicate our world for us. Or maybe it will be something yet to be imagined? New techniques arise daily across Twitter, GitHub and YouTube. It’s near impossible to follow.

While job losses loom, smaller generative studios could rapidly propagate and flourish. (At least I hope so.) Students who look to the generative future, should double down on adaptability, problem solving, and human passion.

Monetization models remain uncertain. But if we understand workflows for generative creation like we do for writers or animators, incredible opportunities may emerge.

… Here’s hoping, anyway.

This essay was researched, transcribed, and refined through a combination of improvisational takes using Otter.ai, and then with Claude v2 from Anthropic. Large portions were re-written by human, other sections are generative. Imagery created in Leonardo.ai's version of Stable Diffusion.

This was an exercise that spanned over a few days at random improvisational intervals.

Nye Warburton is a creative technologist and educator. He lives in Savannah, Georgia.