Ask HN: Is it inefficient to execute programs while working?

tldr; I execute my programs without dry running them in my head first, is that inefficient?

I have a habit of executing the entire program, after I implement a feature to check if it's working correctly.

I usually isolate my functions, in a small test environment, so the feedback loop is fast enough to highlight issues, if any.

But since I started working with a larger code bases, this approach feels inefficient. So my question is, would it be more efficient to read through the code and find out any typos or logical issues before actually executing the code?