I've read many articles like this... While trying to show how some hyped up tech is better than some existing proven tech, they discover that the benefits are insignificant or non-existent... But then the article ends grasping at straws trying to spin the facts to make it seem like the minuscule differences are actually significant and worth the massive tradeoffs (which aren't mentioned in the article)...

Let me spell it out; ProtoBuffers are a failed technology. It was only ever popular because it advanced some developer's agenda at some big tech firm. Its benefits of tiny performance improvements in very specific scenarios are not worth the huge, huge, huge drawbacks such as not being able to easily inspect request or response payloads during debugging.

ProtoBuffers has almost certainly cost billions of dollars in lost productivity as developers have struggled to debug API endpoints due to the reduced visibility of binary compared to JSON strings and also due to the additional time required to integrate with APIs.

ProtoBuffers is the result of irrational exuberance in big tech and would have no place in a free market where developers are trying to optimize for a balance of software efficiency, security, reliability, compatibility and maintenance costs.

It's worse on all fronts except for insignificant performance gains which are even more insignificant once you consider that serializing/deserializing typically represents a TINY fraction of the total workload of a system. This format should never have existed and it speaks volumes about how far many big tech companies have fallen since their golden days.