I’m currently reading “Algorithms to Live By” by Brian Christian & Tom Griffiths. I just finished the chapter on caching, and the author’s discussed the concept of caching as an analogy for how human memory works.

Extrapolating from the author’s words, it seems that, in some circumstances, our memory could be thought of as a linked list. For me, this calls to mind the methods by which folks who compete in memory competitions are able to memorize large sequences of information very effectively (e.g., the random ordering of a full deck of cards).

The thing that really stood out to me and I will hold onto for a while is how this analogy of “memory as cache” applies to people as they age. In computing, we use caches for faster access to important or relevant data. In human memory, this can also be incredibly useful (i.e., associating a physical experience with learned information). However, as we age, the “cache” is continually filling up with more and more experiences, relationships, etc. And thus, the problem as we grow older is not necessarily that our brains are “in decline," but that the cache is becoming more full.

If we think of memory as a linked list, then of course it would be harder to pull data from an ever-lengthening cache while using the “wetware” of our brains as opposed to computer hardware. In computing, we would, of course, solve this somewhat simply: we’d utilize a cache eviction policy.

So the key takeaway I have is this:

human memory does not come equipped with a cache eviction policy