How do you tackle a programming project that's way beyond what you know?

Yes, a small chat program sounds like a useful detour. It gives you a reason to practice connections and message handling without also dealing with boards, posts and replies.

Start with one client sending a message to a server. Then try two clients and work out how a message gets from one to the other. Once you can explain that flow, you'll have a better starting point for deciding how the forum should work.

Before adding more code, sketch the design. What does a client send? What does the server keep track of? What happens when a connection arrives, sends data or disconnects?

Walk through that sketch with your friend and ask them to point out gaps rather than rewrite the program. The goal isn't a perfect design. It's to have a model you can test and revise when the code surprises you.

If you want to experiment with select(), think in terms of one loop watching multiple sockets, not a separate select call for every connection. The listening socket becomes readable when a connection is ready to accept. A client socket becomes readable when there's data to read or the connection has closed.

That can be a way to explore multiple clients without introducing threads at the same time. Keep the example tiny first, because you'll still need to decide how to recognize complete messages and handle disconnected clients.

Check the documentation for the tools you're considering before committing to a low-level design. A networking guide can help you understand the moving parts, but a library may handle some of those details for you. Look at what its examples expect your code to do, then choose the level you actually want to learn.

Keep a copy of the working forum and change one thing at a time. You don't need to add every multiplayer feature in one leap. A small change you can run and inspect is much easier to debug than a pile of new concepts added all at once.