
Adrian de Wynter successfully got OpenAI's GPT-4 to play Doom, a development that has been met with mixed reactions. While some find it amusing to see the AI model navigate the game, albeit poorly, others express concern over its readiness to engage in virtual violence. This marks a significant step in exploring the capabilities of large language models in gaming, with GPT-4's performance demonstrating both the potential and limitations of AI in understanding and interacting within complex environments.





GPT-4V can play Doom, badly — doesn't hesitate to shoot humans and demons https://t.co/HyTya0peuf https://t.co/uvuOR1Cutg
OpenAI's GPT-4 Can Play Doom, But Only Like a Noob https://t.co/7WTMDOjMsX
On the one hand it's fun seeing OpenAI's GPT-4 play Doom (badly). Turning what's on screen into descriptions to then process into movements and actions On the other, little bit worrying how happy this model is to pick up a gun and blast away (virtually) https://t.co/RNNBrW7PJu