the problem I have with the article is that FOSS/OSS used internally and modified for that use and not distributed would mean under normal copyright and work rules that yes GS did own the changes to OSS/FOSS used internally but never distributed.