This article is based on Jack J. McCauley’s brilliant talk at the Computer Vision Summit in Boston. As an AIAI member, you can enjoy the complete recording here. For more exclusive content, head to your membership dashboard.
Today, I’ll be talking about the history of video games, the evolution of AI, and how the two are linked. But don’t worry – it won’t be too technical!
You might be wondering what qualifies me to talk about this topic, so let me tell you a little about my career so far. I've worked in the entertainment industry for over 35 years, mainly in video games but also in feature films like Terminator 2, Dick Tracy, and Edward Scissorhands.
I'm also an inventor. I invented the scrolling wheel mouse and designed most of the hardware for Guitar Hero. Plus, I was the chief engineer and co-founder at Oculus VR, where I worked on their VR headsets.
These days, I've moved into more executive roles, so, sadly, I do less of the hands-on engineering work. However, I stay involved in cutting-edge tech projects when I can. For instance, I teach mobility engineering at UC Berkeley and fund opioid research at the RAND Corporation.
But enough about me. Let’s get to the matter at hand – the evolution of video games and how it’s fueling the explosion of AI.
How video game development is powering the rise of AI
To give you a sense of how graphics have evolved, let me share a clip from 1991's Terminator 2. Fair warning: if you’ve got kids around, you might want to save this one for later.
That part where the sword melts through the door? That’s rendered frame by frame. The part where the bad guy’s head splits in half and comes back together is partially rendered. All in all, it took six weeks to render just that scene on 26 workstations running in parallel! Today, you can render a far more realistic-looking scene in real time, so long as you have the budget for it.
It’s a similar deal with games. In 1993, Doom came out, with those distinctly 90s graphics – compare that to the eerily realistic Unrecord from 2023. The obvious difference in quality is due to the evolution of the GPU (graphics processing unit).
While GPUs aren’t used directly for AI computing, the GPU and AI are closely related – they use the same core hardware. Let’s explore how that hardware has evolved.
