Artificial Intelligence (AI) firms such as OpenAI and Anthropic have typically invested heavily in hardware to support their AI algorithms, but an innovator named Jean Michel Sellier has managed to develop a memory-light machine learning platform using an Atari 800 XL and a genetic algorithm.
Sellier’s platform evolves 32 generations of mathematical predictions within the Atari’s limited memory constraints and is written in BASIC.
The experiment proves that AI innovation need not be dependent on huge investment in hardware, and that efficient processes are key to advancing machine learning on limited resources.
The 8-bit Atari may not be a game-changer in the AI space, but Sellier’s experimentation echoes the untapped potential of genetic algorithms in AI, which have fallen out of fashion in recent years.
