That may be the situation for this particular use case. Other use cases may have different requirements.
A system is only as fast as the slowest component a specific work load depends on. If your use case requires a specific amount of memory, say 16GB, and you're using an 8GB system then your use case is limited to the speed of moving information to and from RAM and the SSD.
While SSDs have decreased the impact of low RAM situations they are no substitute for RAM. I mentioned earlier that the SSD in my M1 Mini is capable of reading data at 2.1GB/sec and writing it at 2.5GB/sec. RAM in the M1 systems is rated at approximately 60GB/sec.
One of the speed enhancing steps Apple made with the M1 was to create Unified Memory. The idea behind Unified Memory is to ensure all the different components that needed access to data had access to data without the need to shuffle around. Each component, such as the GPU, had direct access to the same memory as the CPU. By doing this performance could be increased.
Now you have a number of people who are effectively recommending negating this performance benefit by recommending using even slower SSDs. Instead of keeping everything the shared memory the recommendation is to go out and retrieve data from SSD through the virtual memory paging system. Now you have a high performance CPU which cannot perform to its peak because it's waiting on data from a, relatively speaking, slow SSD. That doesn't make sense.