Well, most of the IDEs still perform all of the work on the remote machine and just stream IO to the laptop, so you're not actually editing files on the laptop, you're sending file editing commands to the remote machine, all the compilation and execution stuff happens on the remote machine, etc. For example, VS Code has client/server components which make the experience feel seamless and handle network disruptions to hide latency and make it feel like it's all happening locally. This setup works pretty great if you're mostly editing text files, i.e. source code.
Actually mounting a remote disk solves a different set of problems, such as if you need to edit something that doesn't support doing the work on the remote machine or if you want to explore the remote disk in Finder or some other GUI.
Most remote dev machine setups I've seen involve some combination of SSH, port forwarding, mounting remote drives, etc., not actually running remote desktop. So you don't RDP and open your IDE on the remote machine, you open the IDE on your cheapo laptop and use whatever SSH or remote filesystem mechanism it provides to connect it to your dev machine. Services like Tailscale make it easy to establish the connectivity between devices. Doesn't work for all dev workflows though, sometimes you're forced to use the GUI on the remote host.