Using AI To Build A Game From Scratch (NO Experience)

Using AI To Build A Game From Scratch (NO Experience)

by Matt Wolfe

Applied Sciences & Engineering AI — Models & Research UYQArtificial intelligence (AI)UYQFGenerative artificial intelligence
🎙 Matt Wolfe 👥 1.0M 📅 April 2, 2023 ⏱ 26 min 👁 554K 📄 documentary 🧭 2026-08-28
Available in: English (current) Français

Keywords

GPT-4game developmentno-codeAI-assisted codingJavaScript

Summary

Matt Wolfe, a non-programmer, documents his attempt to create a video game from scratch using GPT-4. He starts with Visual Studio Code and asks ChatGPT for a simple game concept, settling on a platformer. He follows the AI’s step-by-step instructions, copying and pasting code into HTML, CSS, and JavaScript files. He encounters numerous issues, such as the player not jumping or falling through platforms, and repeatedly asks the AI to debug and provide corrections. After about four hours of iterative troubleshooting, he achieves a basic platformer with movement, jumping, and collision detection. He then enhances the game by adding a side-scrolling camera, collectible coins, a score, a reset button, and a parallax background, all through natural language prompts. He also uses Midjourney and Leonardo.ai to generate visual assets. The final game features an animated character, a lava pit, and an explosion effect on death. The process cost about $20 in GPT-4 API usage. Matt reflects that while the task took longer than expected, it was a valuable learning experience, giving him a basic understanding of JavaScript structure and problem-solving.

179 words

Critical Evaluation

Value of the Information & Strength of the Argument

The video provides a genuine, unscripted demonstration of AI-assisted coding for a complete beginner. Its value lies in showing the real-world process, including the frustrations and iterative nature of debugging with an AI. The argumentation is based on direct experience, not theory, which lends credibility. The creator clearly explains each step, the problems encountered, and the solutions provided by the AI, making the process transparent and educational. The video effectively argues that while AI can significantly lower the barrier to entry for coding, it still requires patience, clear communication, and a willingness to iterate.

Scientific Rigor, Source Quality, Title Accuracy

The video is a personal account, not a scientific study, so its rigor is limited. However, the creator provides links to the final game and the GitHub repository, allowing viewers to verify the outcome. The sources cited are primarily the AI tools used (ChatGPT, GPT-4 playground, Midjourney, Leonardo.ai) and the creator’s own resources. The title accurately reflects the content, as the creator has no coding experience and uses AI to build a game from scratch. The video does not claim to be a comprehensive tutorial but rather a documentation of a personal experiment, which is consistent with its content.

208 words

Title / Content Match

The title accurately reflects the content: the creator, with no coding experience, uses AI to build a game from scratch, documenting the entire process.

Quality & Reliability

7/10

The video is a transparent, first-person account of using GPT-4 to build a game without coding experience. It shows real successes and failures, and provides links to the final game and code. However, it is anecdotal and lacks rigorous methodology or independent verification.

Key Moments

Markers derived by PSI from the transcript: the creator did not define chapters.

Cited Sources

Concurring Sources

  • OpenAI GPT-4 — The AI model used in the video, known for its advanced coding capabilities.
  • GitHub — Platform hosting the game's source code, demonstrating the practical output.

Contribution & Novelties

The video provides a unique, first-hand account of a non-programmer using GPT-4 to build a functional game, highlighting the practical capabilities and limitations of AI-assisted coding. It demonstrates that with clear communication and iterative prompting, even a novice can create a working application, though the process is time-consuming and requires troubleshooting. The creator also shares tips on using the OpenAI Playground and providing context to improve AI responses.

Pour aller plus loin :

  • OpenAI GPT-4 — The model used in the video, known for its advanced reasoning and coding capabilities.
  • JavaScript — The programming language used for the game, with documentation on MDN.
  • Canvas API — The web API used to render the game graphics.
  • Midjourney — AI image generation tool used to create the game’s background and assets.
  • Leonardo.ai — Another AI image generation tool used for the character sprite.

141 words

Radar Profile

The radar chart shows high scores in quantity of information and reliability, reflecting the detailed documentation and transparent process. The technical level is moderate, as the video is accessible to beginners but still involves coding concepts. The overall quality is strong, with a slight dip in technical depth due to the creator's lack of coding expertise.

Reliability 7/10

💬 Très positif. Sur les 30 commentaires analysés, la grande majorité exprime admiration et enthousiasme pour la démonstration, soulignant l'accessibilité de la programmation assistée par IA et la valeur éducative du processus.