When I studied compsci at university, convincing someone to investigate compsci in freshman or sophomore year is way too late — You had to embark on the 2 full years of ‘premajor’ classes starting with your first quarter of freshman year or you would definitely be on the 5 year plan for graduation. You weren’t allowed to declare a compsci major without a B+ average in 2 full years of ‘premajor’ classes, and the major required the balance of 2 full years to complete.
But for what it is worth, I had some exposure to computers in grade school (grades 4–5 — sinclair 1000 and apple 2+, basic and logo), but I didn’t do one bit of computer science again until I was in college, taking classes in computer music, which piqued my interest. The idea that you can’t be a strong compsci candidate without programming continuously from a young age is totally false. I’m as strong a developer as anyone I know, with the resume to back that up, and I didn’t write my first real code until my second year of university. But I did have an extended education as a result of the late start, since the shift to compsci did require a full 4 years of study after I discovered my interest in my 2nd year.
It really would be much more effective to begin that exposure in high school, however, not to mention useful. I hate the emphasis, in the popular press, on ‘coding’ as equivalent to software engineering, but it does have its utility for anyone in an academic context, since being able to automate certain tasks, create simple databases, and perform analysis in spreadsheets is very useful.