Before Agile was taught in school, people were doing Agile. There just wasn't a single name for it.
The start to what we call Agile was in February 2001, when seventeen people met in Utah to talk about the things that they found were working well in software development in a world where there were a lot of projects that were behind schedule, over budget, and maybe never even delivered. This meeting led to the Manifesto for Agile Software Development. But these seventeen people were doing these things for a while by this point - they had developed or were actively using methods such as Scrum, Extreme Programming, DSDM, Adaptive Software Development, and Crystal on real-world projects for real companies and delivering value.
Although not represented, I'm pretty sure that there were also others. A good candidate example would possibly be Lean Software Development, where the first book about it was published in 2003, but it seems unlikely that the Poppendiecks developed it, vetted it, and wrote a book about it in 2 years. I'm sure if you were to ask the seventeen people who were at Snowbird, they would be able to mention other people who were not only doing, but writing and talking about their approaches to building software.
However, waterfall was and still is used as a methodology for developing software projects, even though it is not only flawed (and identified as such in the 1970 Winston Royce paper titled "Managing the Development of Large Software Systems" that identified it). People as early as the 1960s were identifying that managing software projects isn't the same as managing other complex engineering projects and wanted to work toward iterative and incremental development and tight feedback loops with customers, and you can see this in the 1968 NATO Software Engineering conference.