One clever person has already found a unique application for ChatGPT-4, Latest artificial intelligence language model from the creators ChatGPT’s. As it turns out, GPT-4 can read the code for game boy advance games pokemon emerald and turn it into a fully interactive text adventure with the mechanics of a Pokémon adventure.
twitter user Dan Dangond explores AI models could run pokemon emerald, a popular Game Boy Advance game, was released in 2005. Dangond shared a screenshot-by-screenshot thread that includes his interaction with the chatbot. The game starts out more or less as if you were playing it on a handheld – save for a few features, like setting the clock at the start – but works as a fully interactive text adventure. Dangond “plays” the game by answering to the AI in chat.
As always with the mainline pokemon gamesIn pokemon emerald You explore a pixelated world and fight other Pokémon in turn-based battles, so the text roughly describes and enumerates battles. As Dangond begins, the AI offers a choice between three starters on the field: Treeco, Torchic, and Mudkip. From there, Mudkip is sent to battle a Poochyana, which is an actual Pokémon that you can battle on the first route.
Starts the game without any problems, so far so accurate, except the moving van and clock setting scene which honestly isn’t the most interesting thing in the game pic.twitter.com/o4nqtMdCHH
— Dan Dangond (@DanDangond) March 15, 2023
The rest of the game plays like a rough text simulation that includes all of the core mechanics of Pokémon, such as type strengths and weaknesses. There are several missteps though, for example, it fails to calculate that the bug/ground type Nincada is vulnerable to water. It randomly decided to include accuracy calculations after playing through a portion of the game, and missing a Rock Tomb attack. It skips several battles and misses a few beats.
This text-based game comes at a time when many people are still realizing the true capabilities of generative AI. according to OpenAI, the GPT-4 is a multimodal model that is “less competent than humans in many real-world scenarios”, but can demonstrate “human-level performance on a variety of professional and academic benchmarks”. According to other tech journalists who cover technology, it is considered more capable and sophisticated that its predecessor, ChatGPT, and it’s acing standardized test left and right, Now we can add playing Pokémon games to this list of skills.
The ChatGPT-4 wasn’t built with running video games in mind, and it shows. Anyone who wants to play this type of game will benefit greatly from playing the original. It would be almost impossible to play without a general knowledge of the scenery and the world. Still, it’s a fun discovery that the GPT-4 can run this game in any capacity, and that’s where the potential for fun and more tinkering lies.