That is actually overly broad. There is nothing technically wrong with VLANs being used over a large network area as long as the network is designed properly. In modern networks, you can use more advanced Spanning-Tree (Per VLAN RSTP, etc.) to overcome a lot of issues in traditional spanning-tree implementations and you can use things like Overlay networks to provide a VLAN that can be implemented in many different locations via being overlayed on top of a layer 3 routed WAN. This allows the use of a single VLAN or multiple VLANs that cover much larger physical and logical areas without having to configure the VLAN on everything in between the locations and pass it as a traditional layer 2 broadcast domain.
Lightweight access points and WLC do not address layer 2 design issues such as large implementations of spanning-tree, etc. In fact, they depends on the layer 2 network and layer 3 networks being stable and consistent and require that those networks provide the access required by the client devices on the access point as any standalone access point would.
The defining feature of lightweight access points is that they lack the integrated full management capability of standalone access points and depend on a centralized management platform or controller (WLC) to implement things like basic configuration (channel config, transmit levels, etc.) as well as the more advanced features like authentication and encryption config, VLAN tagging, and client and traffic optimization (802.11r and 802.11k etc.).
Lightweight access points do not relieve the network of the need to be designed properly to provide transport for the client devices to the desired network services. Wired networks have the same issues: any given wired network switch needs to be able to provide the required access that the client devices in that location will need.
