Show HN: A competition for small neural networks that play strategy games

The discussion centers on tinybrains.dev, a platform to run a competition for very small neural networks that play strategy games. The creator, codetiger, framed the project as a revival of the fun and learning from contests like the Google Ants AI Challenge and said the goal is to teach network design and tuning. Commenters responded with enthusiasm and personal anecdotes: AnotherGoodName recalled placing 72nd and now doing AI research, atmanactive mentioned a similarly named Steam game, and lokar and cheschire pointed to related projects (Core War and Screeps). adityamishra241 asked about the intended network sizes, and sitzkrieg supplied the size brackets (nano to large with specific KiB/MiB limits).

Opinions split between supportive community interest and a few usability critiques. Most commenters encouraged the idea as a fun learning platform and offered comparisons to past competitions, while lostdog urged the author to rewrite the site copy without AI and remove confusing text. codetiger acknowledged feedback, prioritized improving presentation, and reiterated that learning small neural networks and tuning is the primary motivation, highlighting that competitive ranking and technique sharing motivated his original participation.