You probably saw "OpenAI proved ETH is false" research. Breathe out. As funny as it sounds, it is not that ETH, and has nothing to do with Ethereum. ETH here is the Exponential Time Hypothesis, a 2001 assumption about how hard the hardest problems in computer science are. In plain words. There is a textbook hard problem called SAT: a long logical formula, find values that make it true. Everyone assumed the only way is to try all 2^n combinations, give or take. Hundreds of "this task is as hard as SAT" results rest on that assumption. An OpenAI model wrote a 67-page paper with an algorithm that does it meaningfully faster than 2^n. If it holds, the assumption is dead and a lot of theory gets rewritten. Lean formalization is released, three top complexity theorists are checking it now.