As many of you might be aware, the ARC-AGI-3 competition has just started ... (In case you're not familiar: it's a human/AI benchmark designed to see what AI still struggles with, that humans solve with ease - basically trying to push AI research to focus on new ideas that make AI think more human-l
