

In a sane and healthy software company, they would contribute their changes back to BSD just because it makes easier to apply latest BSD changes amd security fixes.
Most companies wont. They fork, modify the BSD code for their specific hardware, sell it, and never ever update thier software.


Each investor will get one CD of ChatGPT training data per share as a bankruptcy settlement. Every CD need to be inserted into a single user-provided server to train the model anew, because it’s saved as a multipart archive.