Personally I wouldn't want to work for a company actually detaining people, but call me terrible but I'm not sure I'd feel the same about letting them pay to host some code...
1. You provide tools to a group. 2. You believe (in a informed way) that the group intends to act immorally. 3. Your tools will make the group more effective at acting immorally.
Do you have any responsibility for what happens?
1. I make tools (wedding planning software)
2. I believe (in an informed way) that a group intends to act immorally (use my tools to plan a gay wedding)
3. My tools will make the group more effective at acting immorally
Do I have any responsibility for what happens?
Or (alternatively):
1. I make tools (highly specialized chemicals)
2. I believe (in an informed way) that a group intends to act immorally (use my chemicals to improve abortions)
3. My tools will make the group more effective at acting immorally
Do I have any responsibility for what happens?
=====
IMO, it's better not to attempt to be morality police. Focus on making great tools.
So yes I think everyone would agree you have some responsibility (otherwise no one would care about the outcome of the case, why request the right to discriminate if you don't feel that you have any responsibility for the resulting acts?)
I will agree that it's dubious that that nuance is reflected in the rulings on the Masterpiece case. I'm trying to follow the same lines set there, though, in my argumentation.
Would the added features be useful in another context? Would they be useful regardless of the type of work being done? Are they merely enhancements of the platform, or are they specifically tailored to the client's domain?
So if ICE asks github to prioritize better native CI support for Windows, or something, they're doing it because it is of maximum benefit to ICE.