0 comments
Gamers have started mashing together different video games using AI tools. So far, we’ve seen Mario in Halo, Minecraft in Skyrim, and much more. Mind-bending results aside, the tech required to pull this off has some people worried about online security and digital infrastructure.
Ever wondered what Super Mario 64 would look like in Halo: Combat Evolved? How about Wii Sports in Silent Hill? Fancy a Minecraft session in the world of Skyrim?
These ideas used to be mere fantasies for noughties gamers, with most franchises kept strictly to their own titles and mainline series. Thanks to AI, however, merging entire video games is now a reality, leading to some truly astonishing mashups you have to see to believe.
This extraordinary achievement has only been possible thanks to Claude AI agents that are building mods from the ground up via vibe coding, whereby users simply implement what they want to be made using prompts and real-time feedback. Without AI involvement, it would be near-impossible for human coders to manually stitch together multiple games, let alone have them run correctly.
A ‘passthrough mod’ has been developed by Claude AI to run two games at the same time, with one providing the map, level, or world, and the other implementing its unique mechanics like movement and combat. This is how Mario can leap about in Halo and Spider-Man can swing across Gotham City (just to name a few examples).
you all, mind telling me what spiderman is doing on batman? https://t.co/M0cdJlU5Um
— MotionViz (@Motion_Viz) October 2, 2026
It’s worth mentioning that this is not the same as a traditional mod. We’ve seen manual costume changes on characters in games before, across decades at this point, but nothing of this scale and complexity. Cosmetics are not the only thing being altered; it is entire game systems being intertwined with one another, as each one alters the foundations of the other in real time. It’s trippy stuff.
However, amidst all the social media spectacle and joyous mashup clips, some are raising alarm bells about the use of AI coding in this way and what it might mean for the future.
Critics warn that humans being able to complete complex tasks in mere hours is an insane, new precedent that could fundamentally alter how we view software engineering and cybersecurity.
AI coding assistants (such as Claude AI) can generate software and connect different systems with relative ease, which could eventually leave traditional infrastructure like transport grids, electricity networks, and national security at risk of hacking and cyberattacks. Sure, it might seem a little hyperbolic today, but with a lowered barrier to entry, radically reduced labour times, and exceptionally fast machine learning, it’s entirely possible that major security risks could emerge in the coming years.
