Nothing a well-trained model won't be able to instantly solve. It's literally just grunt work, not rocket science.
F DRMs though. Good news is those AAA games are rarely worth anyone's time anyways. Better spin up indies or classic games - a good SNES game is worth a hundred of those garbo AAA license rehashes.
It’s hard to overstate how poor the technical quality of modern games is. Pretty much every DX12/Raytracing title is just a stuttering mess with absurd shader compilation and traversal stutter. Nevermind ridiculous ghosting artifacts and the like produced by Lumen. Modern games are optimized to look good on screenshots and not-realtime in-engine renders for trailers. Hardware and APIs have never been more powerful, and engine developers turn out the worst-running games since 30 fps hardlocked Xbox ports from the 2000s.
I can't stress enough how much this is not true for any meaningful production threat detection software. Models regularly fail reversing even basic obfuscation tasks. Try it yourself. No, ChatGPT is not going to break Denuvo. Probably not even in OpenAI's wildest dreams.
Generating an image or sound is seemingly child's play compared to actual complex software tasks. There's not 1,000 different open source DRM codebases you can train against. It's not a diverse field.
Generating code to do script kiddie hooking? Sure. Reversing a complex multi-tiered obfuscation and trust platform? Yeah, right.