Exploring the Interesting Aspects of Chemical Compounds and Theorems with AI Potential

The study of alkanes reveals that their interest lies in the properties they exhibit, which can lead to advancements in knowledge and technology. AI can determine the likelihood of certain alkanes being considered interesting by analyzing their production rates and academic mentions. Similarly, in mathematics, theorems are deemed interesting if they offer new information and are not easily derivable from existing ones. AI could potentially predict interesting theorems by training neural nets on existing mathematical literature. However, finding genuinely new scientific concepts poses a challenge as neural nets often prioritize mainstream ideas. Human choice plays a pivotal role in determining which new facts or directions to pursue, highlighting the arbitrariness of knowledge expansion.