step two: profit from people willing to code for enterprises for free
There has to be some reason that Google released PyTorch under BSD-3 by Meta and TensorFlow being released under Apache by Google.
step two: profit from people willing to code for enterprises for free
There has to be some reason that Google released PyTorch under BSD-3 by Meta and TensorFlow being released under Apache by Google.
Even LGPL is super challenging here due to requiring dynamic linking. This is fine in Java (unless you use Graal AOT compilation) and Python, impossible in Go, and at best discouraged in other compiled languages like Rust and C++.
I think there's likely a copyleft middle ground - something like the Mozilla public license seems about right, where it copylefts improvements without infecting users. I don't know why Google chose Apache 2.0 over MPL or the like - for something like Android it makes sense, for something like Guava it's not clearcut. But, when people talk copyleft, they're rarely, rarely talking MPL.
I think it's unlikely that Google opensourced Tensorflow with a view to getting some free dev hours, they're doing it to push their OSS agenda, to wield influence over how everyone else does work.
Google saw what happened with other of their technologies like map reduce (not open sourced, the world reinvented in Hadoop and the progress started happening there instead of in MR) and wanted to avoid becoming the bit player in a space they pioneered in. It's a reasonable strategy. (The risk is that you're giving away something that could be a competitive advantage.)
(worked on tensorflow at Google for four years.)
Google really hates copyleft and the GPL.
Full disclosure: I got my current job before I had contributed anything to open source but I was asked a lot about it while applying to companies like it's expected that you work in your free time
No wonder Microsoft loves it.
US government policy is that software developed with public funding is a public good. Current policy encourages government software to released under the most permissive license possible (ideally Public Domain, but practically permissive licenses like MIT/BSD). GPL is farther from the Public Domain than licenses like MIT/BSD. To enter the rabbit hole of material on this topic start at https://code.gov/about/overview/introduction/.