What are other "uncrackable benchmarks"?

Uncrackable benchmarks refer to tasks or challenges that are considered extremely difficult or impossible to achieve due to their complexity, technical challenges, or the current limitations of our knowledge and technology. These benchmarks span various fields, including AI, computing, physics, and biology. Some examples include achieving Artificial General Intelligence (AGI), building a fully functional large-scale quantum computer, developing a universal cure for cancer, and achieving sustainable nuclear fusion reactions. Other examples involve tasks that require human intuition, creativity, emotional intelligence, or nuanced contextual understanding, such as subjective aesthetic evaluation, paralinguistic communication, and empathetic social interaction. Historically significant benchmarks that were once considered uncrackable but have since been achieved include defeating human champions in chess (Deep Blue vs. Kasparov, 1997) and Go (AlphaGo vs. Lee Sedol, 2016), and solving Jeopardy! (Watson, 2011). Current frontiers include general game playing, robotic household chores, and the Winograd Schema Challenge, which require integrated, causal models of an open world and robust common-sense reasoning. The ultimate uncrackable benchmark remains achieving Artificial General Intelligence (AGI), a system that can learn and excel at any intellectual or physical task a human can, demonstrating transfer, creativity, and understanding.