Maybe for Reddit and Apollo case, this is alright & makes sense. But at the risk of sounding pedantic, don't generalize it.
Most of our commodity software is built & deployed by packages, APIs and frameworks we have very little control on. We just hope things don't break/change as drastically and we can modularize our projects as much as we could, to bear some shock or disruption. Unlikely anyone can build & maintain consumer grade softwares ab initio
The closer you are to the final customer the better is your chance to bill for the value you deliver (standing on a fragile cobbled together set of third party components). The short term rewards strongly favor relying on others. The business dynamic compounds this as it places a premium on first movers. Going against this dynamic must be carefully considered and only makes sense in isolated cases (where it is the winning move).
Disagree. (In the generalized sense) You cannot extricate yourself entirely and call it a winning move, unless your business is at a scale as large as Dropbox moving out from AWS ecosystem.
I still feel no matter how natively one tries to build products - they cannot build everything. You cannot create CI/CD, monitoring, frontend, containerization, and cloud services just for your software or service. Those depend on some platform API which you won't create just for your product. Short or long business value - unless one becomes a major player with several software engineering teams building a product ecosystem - other people's APIs and frameworks will be used. And that is perfectly fine. That is how good products should be -using nice building blocks. No need to reinvent the wheel everytime.
Most of our commodity software is built & deployed by packages, APIs and frameworks we have very little control on. We just hope things don't break/change as drastically and we can modularize our projects as much as we could, to bear some shock or disruption. Unlikely anyone can build & maintain consumer grade softwares ab initio
reply