Applied math progress tied to human trial-and-error, not just AI solutions

An opinion piece argues that the process of human trial and error in applied mathematics, not merely final solutions, drives progress, implicitly cautioning against over-reliance on AI. The claim is an editorial perspective rather than a new empirical finding, so its analytical weight is limited. It matters as a contribution to the ongoing debate on AI's role in scientific discovery.