A Future for Software Writers

Stan Ulam was a Polish mathematician working at Los Alamos on the Manhattan Project. After a serious illness, he spent long days recovering in bed, passing the time with endless games of solitaire. He began to wonder: instead of calculating the odds of winning by hand, what if he just played thousands of games and counted how often they worked out? That idle experiment became the Monte Carlo method.

Software is entering a similar moment. The act of typing code is no longer scarce. Machines can produce it endlessly. What matters now isn't writing code but choosing between the countless ways it could be written.

This is the new craft. The best engineers won't be the ones who crank out commits. They'll be the ones who see the options clearly, interpret the tradeoffs, and decide which path actually fits the system they're building. Code is a commodity. Judgment is not.

The future of programming is leverage through architectural decisions. The skill is not in typing, but in framing the problem, exposing the forks, and guiding development toward the right outcome.

Good engineers have always thought hardest before they typed.

Soon, that will be the whole job.