Playing a Neural Network's version of GTA V: GAN Theft Auto
What it is
Sentdex demonstrates GAN Theft Auto, a GameGAN-based neural network that recreates GTA V so the network itself is the playable environment. He shows the results and releases the code.
Why Hacker News recommends it
Cited as an early, striking demo of neural networks acting as game engines and world models.
- 3 people recommended it
- 4 different months
- 0 strong recommendations
- first 2021, last 2025
What HN commenters said
“This video really amazed me in terms of automated game development.”
Shows promise of automated game generation.
ALittleLight, 2021-08 · Recommendation · in “Where are the robotic bricklayers?”
“You might like Gan Theft Auto”
Neural net encodes game world rather than bitmaps.
eru, 2022-12 · Recommendation · in “Netflix: 30% bitrate saving using Film Grain Synthesis”
“reminded me of this cool project”
Cool related project of neural network GTA.
7734128, 2024-08 · Recommendation · in “Diffusion models are real-time game engines”
Excerpts from Hacker News comments, linked to the originals. Summaries and labels were written by a language model from those comments.
Topics: gan, gamegan, gta v, world models, generative ai
Also recommended by the same people
explainerThe Antique Toaster that's Better than YoursTechnology Connections · 12 HN recommenders · 2019–2025
explainerBatteries aren't the only way to store energy. Here's another.Technology Connections · 1 HN recommender · 2021–2023
explainerThe Epidemic of Fake DiseaseMedlife Crisis · 7 HN recommenders · 2020–2026
explainerTransformers, the tech behind LLMs | Deep Learning Chapter 53Blue1Brown · 13 HN recommenders · 2024–2026
lectureDeep Dive into LLMs like ChatGPTAndrej Karpathy · 10 HN recommenders · 2025–2026
tutorialLet's build GPT: from scratch, in code, spelled out.Andrej Karpathy · 14 HN recommenders · 2023–2025
tutorialA Hackers' Guide to Language ModelsJeremy Howard · 8 HN recommenders · 2023–2024
talk[1hr Talk] Intro to Large Language ModelsAndrej Karpathy · 9 HN recommenders · 2023–2026