The success of neural networks on problems like Go and protein folding, long considered intractable NP-hard problems, is profound. It suggests our formal understanding of computational hardness, which focuses on worst-case scenarios, may be an incomplete model for how to find useful, approximate solutions in practice.

P represents problems we can solve, while NP represents problems where a solution can be easily verified. If P=NP, any problem with a verifiable solution could be efficiently solved, implying we can know everything we want to know. It's a question about the ultimate limits of discovery.

Unlike traditional engineering, breakthroughs in foundational AI research often feel binary. A model can be completely broken until a handful of key insights are discovered, at which point it suddenly works. This "all or nothing" dynamic makes it impossible to predict timelines, as you don't know if a solution is a week or two years away.

History shows that major breakthroughs are often preceded by someone who meticulously defines a problem, attracting solvers to it. However, society celebrates the solver, not the definer. Spending more time on precise problem definition is a powerful, yet under-appreciated, path to innovation.

Andrej Karpathy's 'Software 2.0' framework posits that AI automates tasks that are easily *verifiable*. This explains the 'jagged frontier' of AI progress: fields like math and code, where correctness is verifiable, advance rapidly. In contrast, creative and strategic tasks, where success is subjective and hard to verify, lag significantly behind.

AI excels at solving problems with clear, verifiable answers, like advanced math, allowing for effective training. It struggles with complex societal issues like unemployment because there is no single, universally agreed-upon "correct" solution to train against, making it difficult to evaluate the AI's path.

While any NP-complete problem can be reduced to another, SAT solvers are the practical choice because of the immense effort poured into developing heuristics that efficiently handle the structured instances arising in real-world applications. Their advantage lies in engineering, not pure theory.

While problems like protein folding are NP-hard in theory, the instances found in nature have structural properties that allow for efficient solutions. Real-world cases of NP-hard problems aren't the adversarial, worst-case scenarios used in complexity proofs, explaining the gap between theory and practice.

For some NP-hard problems, like 3-SAT, a random guess satisfies 7/8ths of clauses. The PCP theorem proves that finding a solution satisfying just an epsilon more (7/8 + ε) is as computationally hard as finding a perfect 100% solution. This places severe limits on approximation algorithms.