People often ask why Google, Facebook, Twitter, and Amazon give away their own ML and data processing libraries for everyone to use. Why they publish results of their research. The answer is that they have the talent and they have the data. Without talent and the data those libraries and publications are useless. [Update] As was pointed out in the comments, the post gives an answer to why they don’t fear to share their code and knowledge, but doesn’t give an answer to “why they choose to do that”, so here’s why. Five reasons:
- Necessity — To attract the best talent, you have to let them publish. Otherwise, the talent will choose your competitor.
- Marketing — When they share their libraries it creates buzz among developers and in the news.
- Better image — Sharing makes them look good and attractive among the new generation of developers who warship open source and sharing.
- Hiring — Independent developers use those tools, so it becomes easier to hire people that already know your tools.
- Cheap growth — Independent developers improve your tools.