Kodak had prototypes of digital cameras in the 1960's but they were afraid it would cannibalize film sales so it died in meetings. The United States Post Office had online bill pay and digital certificate PKI email system in the 1990's but never releases it beyond a few of their vendors due in-fighting and the fear of losing first-class mail volume.
Typical of management to shut down engineers with superior cognition, integrity and foresight.
If you have the ability to do so, don't let business types control your products. Engineers should stick together and build in their own. The investor and executive class, that doesn't do actual problem solving but just focuses on using other people to collect more money regardless of all else, should not be rewarded.
Who are you kidding? Most engineers don't know what and who they are designing and building for. Engineers sticking together? These are the same folks that fork over every microaggression to release a command-line tool that did the same thing the last three releases did except it’s not rewritten in Rust.
In other news, Boeing internally developed warp FTL travel years ago and IBM developed a working multi-billion-qubit quantum computer internally years ago.
Because Google's actual mandate is to keep everything the same because they profit massively from the status quo.
They aren't innovators or change agents, they are a calcified heard of pandas that can't hunt. Yes they can do interesting research, but they can't successfully productize that research and take it market. The graveyard of google products is the evidence.
That’s actually not a problem, as long a non-compete agreements are not enforced (they are not) and as long as Google takes in a lot of interns (they do). Plenty of knowledge transfer with very little delay…
Companies don't want to support internal products with in-house resources that cost too much when they can get it cheaper from contractor or vendor or opensource.
2022 LaMDA: https://arxiv.org/abs/2201.08239
2020 chat bot: https://arxiv.org/abs/2001.09977
2018 bert: https://arxiv.org/abs/1810.04805
2017 LLM: https://arxiv.org/abs/1706.03762