You, Me, and the Machine

Shawn Foust
9 min readJan 17, 2023

--

(This is an aside from the main thread on Fortis. On occasion, I meander into foreign pastures.)

Let’s talk Generative AI.

When I pictured the rise of our robot overlords, I generally imagined more deathbots and fewer chatbots. Of course, eradicating Humanity is probably easier if we stop interacting with one another entirely, and ever more sophisticated chatbots seems like a reasonable start down that path. I’m having vivid visions of millions of bots talking to each other on Tinder and no one ever going on dates until Humanity dwindles into nonexistence.

Skynet has gotten real weird.

Wild times.

Prompt: Chatbot Wars. Stable diffusion is making me unstable. Where do the images come from? What do they mean? Time to get some red string and connect the dots.

I’ll leave off the existential questions for another time and focus on the relationship between AI and the video game industry.

The games industry is an odd bird. It’s this awkward marriage of technology, creativity, and community. Technology tends to be the dominant partner in that trinity, with innovations in distribution, storage, rendering, etc. having an outsize impact on the nature of creativity and community in games. Figuring out a better/cheaper/faster/clever way of doing things can lead to massive opportunities. Add in the fact that games folk tend to be early adopters and you’ve got a recipe for constant shifts in how things are done.

It’s part of the reason games tends to spin out so many valuable companies so consistently. In games, the way things have been done is a weak indicator of how they will be done on any timeline greater than a few years. Particularly recently, where the pace of disruption has increased at an exponential rate. Scale incumbency is an advantage, but it’s not an unassailable one.

And things are just gonna get more wild from here.

The Era of Compounding Tech

Why?

We’re in the Era of Compounding tech. At this point, the rate of advancement is increasingly unbounded. Technology is more interconnected, with advancements in one having a broad, rippling effect across the ecosystem. That ripple runs into ripples of other advancements and creates all sorts of unpredictable outcomes. It’s B-A-N-A-N-A-N-A-N-A-N-A-S.

Think about it this way: For the vast majority of Human history, a parent’s experience with technology wasn’t all that different than their children’s. Perhaps some trader rolled into town and had a better set of awls for leather-working or something, and people were all: “Sweet, this is a better way to punch holes through leather. +3.5% to village leatherworking productivity. Maybe I’ll make an extra belt this week.” Exciting, but it wasn’t like the Human condition was constantly and fundamentally changing.

Compare that to today.

I have to spend ten hours a week just trying to understand WTF is going on. My childhood was defined by the shift from OFFLINE to ONLINE, which was massive. My ten year old is being raised in a different world. One defined by technological accessibility. Tech sits in the middle of all interactions with all things. She is immersed in a technosoup that permeates the reality of her being. For her, everything in the world is easy and immediate (on a relative basis to me).

Desires have instant gratification.

Want to know something? No problem. You can get an education on any topic within minutes — along with some helpful suggestions of ten related rabbitholes organized in an ever-growing, ever-streaming playlist. Much like the spice, the information flows. Some poor wizened scholar got cast out of society as a heretic back in the 16th century for whispering about a heliocentric solar system while my daughter is lecturing me about how Pluto isn’t a planet (it is, and I refuse to engage with planet-denialism).

Want to buy something? No problem. If it’s online, it’ll be here instantly. If it’s physical, it will be here in a day courtesy of our Amazon Prime driver (who should be paying rent for the amount of time he spends here). The idea that something might take more than a day is STRANGE and CONCERNING.

Want to play something? No problem. My kid can download and then GORGE upon a GREAT FEAST of free games (assuming she’s willing to be absolutely obliterated by ads, which she totally is). Back in my day, I had to beg my mom for a month to take me to Toy ‘R Us so I could buy a game I spent four months saving for and then wait another two months for Nintendo Power to explain to me how to beat it. And yes, the car ride was uphill both ways.

Everything is fast. Everything is now. Everything and everyone is a click away. And at any given moment, whatever the current thing is will likely be replaced by the next, better thing.

As a parent, I’m expected to know and understand this situation and manage it. I’ve got an edge over most given the nature of my work, but the cognitive load of staying on top of things is extreme. My TikTok is NOT taking off. I’m poorly equipped for this. The problem is that I merely adopted technosoup existence while my daughter is being raised in it. Her life is different. Radically so.

She also has the MVR (minimum viable rhythm) for TikTok.

Rough paraphrasing of conversations with my kid. I really don’t know where she gets it from.

She’s a denizen of the Era of Compounding Tech. Everything is churning and roiling and evolving. Faster and faster. Broader and broader.

Increasingly, I’m worried that the rate of technological gain is outpacing the rate of Humanity’s ability to adapt to it. But that isn’t going to stop it. In a fully interconnected world, technological optimization will occur, spread, and render anyone who doesn’t adopt it inefficient. The magic of integrated globalization.

This is a polite way of saying the moral and ethical debates about technology tend to fade in the backgrounds to the reality of inevitable progression.

Which brings us back to AI and its consequences.

ChatGPT. Stable Diffusion. Midjourney. The four hundred other companies VCs are going to put capital into. It’s all coming.

The Value of Human Work

I think a lot about the value of an hour of Human work. On an absolute, subjective, and relative basis. The relationship between type of work, location of work, and supply of work produces all sorts of strange equilibriums, which technology has a habit of upsetting with some regularity.

Some tech advancements might have little to no impact on an existing equilibriums (the better leather awls example), while others provoke a wholesale reshaping of society — Agriculture, Industrialization, and Internet, I’m looking at you. When a tech-shock to the system occurs, there’s typically some delay as the market fiddles with capital allocation and deployment until a new system can settle in. Longer periods of reordering tend to allow for a calmer procession, as people shift skills and locations to adjust.

As far as I can tell, those periods are growing shorter. Early efforts at industrialization took far longer than the more recent entrants. The internet managed to spread globally in a few decades. Mobile reached global penetration in about a decade. All of this has had impacts. All of it has had displacements. The oft discussed Gig Economy is a good, recent example.

Certain fields have been largely protected from these shifts. Creative fields are front and center here. The cultural aspects to being professionally creative are a barrier to entry that extends beyond standard capital cost considerations. While a widget may be turned in many locations, a piece of artwork that resonates with a person tends to be a more geocentric ordeal. Sure, there’s been some shifts, tax incentives in Georgia or Canada means LA might a less attractive to place to film something, but by and large, the value of an hour of work in creative fields has been somewhat protected.

It takes a long time to train your hand to move in a certain way to produce a certain visual outcome that will be appreciated by the certain group of people that are culturally aligned to appreciate that thing. These things are distinctive and difficult to duplicate. They require training, dedication, and nuanced understanding. Producing a work of quality is an endeavour.

Enter Generative AI.

Oof.

It turns the system on its head. The millions of hours of work that have been produced by generations of talented artists (singers, and writers) is now weaponized against the artists of today and those to come. All of that talent, skill, and sweat is distilled into a training set for a voracious, never-tiring cybernetic art wizard. It consumes. Compiles. Evolves. Every interaction gives it more knowledge, more understanding.

This was good.

That was bad.

ArtWizard V1 understands.

ArtWizard V2 will do better.

An artist might devote a lifetime to developing a style and craft. An AI can consume that output and begin duplicating it with reasonable effectiveness instantly (and getting better all the time). The bounding agents of experience, cultural knowledge, and talent that have historically been the gatekeepers of creative professionals tend to fall apart in this context. The accumulated history of all available works are sufficient to train an AI to compete.

And (soon) compete well.

I suppose this was inevitable, considering the direction of things. Muscle memory has long been a depreciating asset in the modern era. Knowledge may be joining that group — why learn a thing when you can google or ChatGPT it? AI has the potential to usher in a post-skill, post-knowledge…nirvana?

That feels like the wrong word. Particularly given the impacts upon people that are currently skilled and currently knowledgeable. And Human expression generally.

So, assuming we’re heading down this path and we can’t move off of it, what changes? Probably a lot, over some timeline that’s measured in a low number of years. As Generative AI continues to evolve, it will continuously reduce skill and craft knowledge as contributors to outcomes relative to judgment, capital availability, and risk friendliness. This re-weighting magnifies the existing trend that was already moving in that direction.

Good time to be an entrepreneur with a dream.

Bad time to be a highly invested incumbent.

Generative AI & the Games Industry

A lot of people are making the argument that Generative AI is just going to become some sort of production tool pocket companion rather than a wholesale replacement for Human crafters. That’s probably true, but even at that level it’s likely to radically shift the cost-per-work on a given art team. That could mean that the size of art teams is about to get considerably smaller …

…OR…

…what is possible within a game for a given cost structure is about to get considerably higher. It will take some time to adapt, but the modern game industry’s standard designs are all predicated on velocity of content consumption being a primary constraint on how games are developed and operated. Level design, meta design, and economy design are all keyed to the idea that content is limited and costly to come by.

We may be entering a world where the same art team (or a team with the same number of people with compatible skillsets) are capable of producing an order of magnitude more content.

If true, the entire foundation of modern system design begins to show cracks. Design will need to adapt to a content-rich environment.

Branching content.

Responsive content.

Endless content.

The implications are profound. The possibilities are exciting. For those of us preoccupied with the Design Lane, it’s a moment infused with potential.

Potential for progress, but at a cost. Displacement. Rearrangement. A wholesale refactoring of the creative pipeline. All possible, all increasingly probable. How that will impact each of us is unclear, but we should remember that the industry has rarely favored the unresponsive. Teams will be a first mover in adopting these practices and tools, and, should they prove to be the future, the rewards will follow.

Thoughtful leaders should be considering not just the potential of these tools, but how best to build a stable bridge to the sort of team that could make use of them. This isn’t a simple affair. The march of technology can be brutal, but the Human element of making games is its best element. Regardless of the tools involved, creative potential is magnified when different people of different skills are brought into collaboration with one another. Protecting that is a worthwhile endeavor.

And I don’t just say that as an aspiring BLOG THOUGHT-VISIONARY nervously eyeing ChatGPT.

Stay safe, friends.

--

--