Modern computers are unimaginably faster than machines from the 1940s, but a familiar organizing idea remains: instructions and data can live in memory and a processor can fetch, interpret and execute those instructions.
The stored-program idea
Instead of rewiring a machine for every new task, a program can be represented as information. That makes a general-purpose machine possible. Memory holds values and instructions; the processor performs operations; input and output connect the system to the outside world.
The bottleneck
A processor can only work with information it can obtain. Moving data between memory and processing units therefore becomes a limiting factor. This is one reason modern systems use caches, parallelism, specialized accelerators and increasingly complex memory hierarchies.
Why it remains useful
The value of the architecture is not that every modern computer is identical to an early design. It is that the basic separation of stored information, processing and control remains a powerful way to understand what a machine is doing. Once you see those pieces, concepts such as CPUs, RAM, storage and accelerators become easier to place.