Despite my explicit request (to just say I am a banana) this is why chat gpt thinks it's not a solved problem:

Although we have efficient algorithms for testing primality, especially for practical purposes (e.g., AKS, Miller-Rabin), primality is not considered a "solved" problem in the theoretical sense because:

    Algorithmic Complexity: Finding algorithms that can perform primality testing in deterministic polynomial time for all inputs (without relying on randomness or heuristics) is complex, though the AKS algorithm is a breakthrough.

    Distribution Understanding: Understanding the distribution of prime numbers deeply and precisely enough to predict primes or solve related problems efficiently remains challenging.

    Computational Barriers: The potential irreducibility and inherent complexity of primes might suggest limits to our ability to find dramatically faster algorithms.