Half of data scientists or ML engineers I work with can implement and train models but do not necessarily understand the underlying principles or math.

Most people just care that it works and not why.

Really dangerous depending on the problem and field tbh