The world is full of problems to be solved.
Each problem (e.g. reading a CXR, interpreting bloods) can have many variables.
We call any single combination of these variables an "instance" of the problem — a single pathology image is an instance of the problem of reading pathology images.
An algorithm is a generalised procedure that can solve all instances of a given problem.
They have a massive impact on our lives, taking in huge amounts of data and churning out predictions.
Which begs the question: what is data?