Volos Rakic likes to experiment with microcontrollers and small displays. The maker shares his knowledge on the YouTube channel Volos Projects and explains there how you can implement ads, games and other projects on tiny screens. His latest project is a simplified poker game on a T-Display S3 board from LILYGO, which houses an ESP32 microcontroller and an integrated display. Rakic usually programs everything by hand, but this time he had help from ChatGPT for once.
In his project video, the maker explains how he got the AI to program the basic game logic, solving what he found to be the hardest part of the project. In fact, ChatGPT quickly leads to initial results in the usual dialogue form and creates a functioning text variant of the poker game that can shuffle cards, remember up to five of them and evaluate the result at the end of the round.
However, when Rakic wants to add additional jokers to the virtual deck of cards, ChatGPT gets in the way and insurmountable communication problems arise. In a comment to the video, Rakic writes that sometimes it is simply impossible to explain to the AI what you mean. Ultimately, he gives up on the feature, foregoes the additional joker cards and asks the AI to convert the program to Arduino-C. After a successful functional test in the online simulator Wokwi, he inserts the required graphics into the program code by hand.
Navigating ChatGPT to your desired destination is not always easy.
Despite these additional steps, Rakic is overall happy with the result because he saved a lot of time with ChatGPT’s help. However, he sees AI more as an auxiliary tool and not as a replacement for a programmer. He leaves it open whether he will continue to experiment with AI. If you want to look at the source code, you can find it in this GitHub repository.
To the home page
#Microcontroller #Poker #game #programmed #ESP32 #ChatGPT