I'm trying to educate the Thunderbolt 4 is everything Thunderbolt 3 should have been! Sure they're the same speed and all but 4 has to have it's own chip so every port has 40G bandwidth to 3 one thunderbolt controller for two ports!
You're talking about an implementation detail of Thunderbolt 4 that is not always true. For example, PC's that use Intel Maple Ridge Thunderbolt 4 host controller will have two Thunderbolt 4 ports per Thunderbolt bus that have similar shared bandwidth as a Thunderbolt 3 host controller. The two Thunderbolt 4 ports will be limited to PCIe 3.0 x4 speed (<31.5 Gbps) of PCIe data (but more likely ≈24 Gbps which is what I've measured for Titan Ridge Thunderbolt 3 host controller using two NVMe Thunderbolt enclosures and software RAID - one per port).
The Thunderbolt 3 buses of the Ice Lake based MacBook Pros are integrated in the CPU so they don't have the 31.5 Gbps PCIe limit (the limit is more like 41 Gbps for any combination of up to 4 ports). It still has two ports per bus though so you can only connect two displays - one per port of a bus, or two for one port of a bus. PCs with Tiger Lake are similar except they support Thunderbolt 4.
The Thunderbolt 4 buses of M1 Macs are also integrated in the CPU. Apple made each port its own bus so each port can connect up to two displays.
Anyway, the benefits you see with the Thunderbolt 4 implementation on M1 Macs is not because they are Thunderbolt 4. It's because Apple made them that way.
I suppose if you want equal capability with the discrete Thunderbolt 3 and Thunderbolt 4 controllers (Titan Ridge, Maple. Ridge), you could just expose one port per Thunderbolt controller, and connect multiple Thunderbolt controllers. It would use up 4 PCIe lanes per Thunderbolt controller though.