He opened the textbook to problem 8.12—a steady-state heat transfer problem with a 4x4 matrix. No manual. No shortcuts. Just paper, a pen, and the cold war between his brain and the universe.

He closed his laptop. “No,” he said gently. “But sit down. Let me show you how to solve problem 6.11 the real way.”

Three years later, Leo was a grad student. He was teaching his own section of numerical methods. A student stayed after class one day, eyes red, pencil chewed.

“Yes,” Leo said, trying to sound confident.

He started the Gaussian elimination by hand. At midnight, he made an arithmetic error and had to restart. At 1 a.m., he realized the matrix was diagonally dominant, so he tried Gauss-Seidel. By 2 a.m., he was writing a basic Python script on his laptop because doing it by hand was like digging a trench with a spoon.

She sat. He picked up a pencil. And for the first time, the ghost of Chapra smiled.

Leo leaned back in his chair. He wasn’t relieved. He was something stranger: he was competent .

Leo looked at her. He saw his old desperation. He remembered the false prophet of easy answers.

The next day in class, Dr. Varma collected the homework. He flipped through Leo’s submission. His eyes narrowed. “Leo,” he said, loud enough for the room to hear. “Your error analysis for problem 6.11 shows a relative error of 0.0001% after three iterations.”