John Carmack, renowned game developer, has come to the defense of Microsoft’s AI-generated Quake II demo, which has faced substantial criticism from the gaming community. The demo, created using Microsoft’s World and Human Action Model (WHAMM) AI game engine, sparked controversy after its release due to its poor performance, graphics, and gameplay. Critics argued that the demo disrespected real developers, pointing to issues such as NPCs disappearing, texture glitches, and disorienting environments that rendered the game virtually unplayable.
Despite the backlash, Carmack emphasized the potential of AI as a transformative tool in game development. He argued that AI can significantly enhance the capabilities of programmers, artists, and designers, fostering progress within the industry. While acknowledging that the AI-generated Quake II demo does not replicate the original experience of the iconic game, he and other industry leaders believe that AI’s role in game development will only grow, despite ongoing concerns about potential job losses.
Carmack’s defense underscores a broader industry perspective where AI is viewed as a catalyst for innovation rather than a threat. As AI technology continues to evolve, its integration into game development processes could pave the way for new creative possibilities and efficiencies, fundamentally altering how games are created and experienced.

Collections

John Carmack defended the use of Microsoft's WHAMM AI game engine for the Quake II demo amidst substantial criticism for its subpar performance and graphics. He emphasized AI's potential to enhance game development by improving the capabilities of programmers, artists, and designers. Carmack and industry leaders believe AI's role in game development will grow, despite concerns about job losses, and see it as a catalyst for innovation in the industry.

John Carmack Defends Microsoft's AI-Generated Quake II Demo Amid Criticism

I’d wager detractors saw the output of WHAMM less as a threat to developers and more of a catalyst for garbage-- generative AI applied to the area it is weakest in.
EDIT: I notice “World and Human Action Models towards gameplay ideation” amongst WHAMM’s references. That gives me an idea of what their end goal is, although I doubt it will enable any productive ideation.

If I had a tool which was going to put me ahead of the competition, I wouldn’t be out here trying to convince anyone of it’s power, let alone be telling anyone about the tool in the first place.

And now show us an AI generated game that no-one has ever played before.
After all this AI hype is over it’ll be mostly used by developers anyway. But not for gerenating whole codebase but as a mentor and QA.
Others will use it for faking everything and porn.