Groups are not difficult to unbundle from Facebook. What's difficult is financially supporting the cost of running such a network when Facebook's model is 'free' for people to use. That means you're going to struggle to charge for the service, which wipes out a big category of potential competition. Sure, you
could charge, and you'll struggle to gain any traction. It leaves advertising, thin (one person, low costs), and non-profit avenues.
Thin either requires charging or running advertising. If you run ads on people's groups, it's going to annoy them inevitably, devaluing your product. You can charge as a thin product, and it'll be a very slow growth model. The non-profit approach requires donations or the equivalent, which is difficult as well.
Ning was an example of the rapid growth the segment makes possible. It proved financially unviable in the manner in which they approached it, and they proceeded to ruin their own product's appeal during the business model conversion.
It's certainly feasible to run a thin model (avoiding the bleed-to-death financial problems of a Ning), in the style of Pinboard, however it'll be a very hard slog. It would take someone insanely dedicated, with a willingness to invest many years of struggle (five years minimum, plan for ten). That's a difficult pitch in the age when (in the US) an average software engineer can earn $140,000 per year with good benefits, and an average web developer can earn $80,000. Someone with the ability to do it, gives up a lot to pursue that (an average sofware engineer at five years of struggle, making barely-getting-by money pursuing this idea, loses upwards of a million dollars in total compensation vs working for a decent non-FAANG tech company).