I'm not sure who the appropriate audience is for this article. The writer seems to be genuinely trying to convey that lock-in isn't good, but isn't the be all end all, which everyone knows. Lock in is avoided not by looking at software stacks but by following standards that translate with minimal friction across more than one platform, unless your a highly specific use case that requires lock in due to heavy tuning. Sometimes lock-in can't be avoided, most times it can. Open source has literally nothing to do with lock in unless you're building directly on top one that product. It's just a development methodology that provides resiliency against losing access to the software, as long as you're willing to build any future improvements yourself if it gets locked up, which in a way provides a slight buffer against lock in, but maybe he runs in circles where people all think open source software is magical.