Services only makes sense from the perspective that you can make your software "stickier" by adding consulting to help with implementation and customer success.
This is reflected in the multiples that VCs and acquirers give software vs. services companies - high-growth software companies often command a 10-14x trailing 12 month revenue, while services are lucky to get 4-5x. It's just a less attractive business.
That makes no sense at all. No LP would be in favor of paying an "investment tax" which is not broadly paid yet is broadly used. LPs put money into venture firms for higher than average returns. What you describe would make the investor's job harder.
In the Valley (and on HN) the VCs are the BSDs of the money scene. But they need their customers (LPs) and if you go to any LP meetings you'll see that most of the money comes comes from the actual BSDs of PE, hedge funds and pension funds. And since VC is such a tiny part of the financial world, you'll see that the folks that represent the big LPs at those LP meetings are just kids on their first job after an MBA, or even between college and MBA: all they really care about is reporting good returns to their bosses.
Overall, GitHub should go the services route. Right now enterprises are footing the bill with private repositories and organizations. If they want to become positive by just jacking up the price, buyers are going to start shopping around, and GitLab looks better everyday. Services like setting up CI/CD pipelines for enterprises could be highly lucrative, especially if they can be templated with internal tools. Most software (and SaaS) companies eventually arrive at the conclusion that they are better off selling the expertise on how to use their software than selling the software itself.