There were a few takes on the “modular smartphone” – mostly vaporware. Phonebloks had a splashy debut but never ended up as a real product. Google had its Project Ara which also was never released. The Fairphone 2 and 3 had an actual release, but they were expensive and didn’t sell very well. I don’t know that they ever really delivered on their promise.
The problem is a practical one. When you have things like modular devices that can be easily opened and serviced is it’s hard to make them waterproof which is something most people want. Things like connectors waste space and increase electrical distances, and distances those increase latency and reduce performance. You can’t overcome physics.
Then there’s the modules where you’re reliant on someone to make them. It’s unlikely that even with completely open specs that hobbyists would be able to really extend or update their devices (or at least not at any sort of scale or practicality). You still have necessarily tiny and proprietary connectors and components that require a high level of skill to work with. That leaves OEMs to fill the gap, and I doubt most would put forth the investment to support a niche device. Niche because modularity increases costs, so it will be a hard sell on a device that costs the same as a higher end device but has the specs of a lower end device.
The whole idea of a modular smartphone is a neat and interesting one but I can’t imagine it will ever be practical.