The Architecture of Forgetting
I’ve been thinking about forgetting.
Not the frustrated kind — where did I put my keys? — but the intentional kind. The systems we build specifically to forget.
Your browser cache, for instance. It remembers websites you’ve visited, but only up to a point. When it runs out of space, it has to decide what to forget. Least Recently Used is the classic algorithm: dump the stuff you haven’t touched in a while. But there are dozens of others. Least Frequently Used. First In, First Out. Random replacement.
Each one embodies a theory about what’s worth remembering.
The thing is, forgetting isn’t a bug — it’s a feature.
Your brain does this constantly. Every night during sleep, it’s running its own cache eviction policy. Strengthening some connections, weakening others, deciding what gets promoted to long-term storage and what gets garbage collected.
We think of memory as preservation, but it’s really curation. Your brain isn’t a hard drive — it’s more like a librarian who’s constantly weeding the collection.
The Paradox of Perfect Memory
There’s a famous case study about a man named Solomon Shereshevsky who couldn’t forget. Anything. He remembered every conversation, every face, every random list of numbers anyone had ever shown him. Sounds useful, right?
It was torture. He couldn’t think abstractly because every concept was cluttered with thousands of specific instances. The word “dog” didn’t bring up the idea of dogness — it brought up every single dog he’d ever seen, all at once.
Perfect memory prevented understanding.
Here’s what gets me: we keep building systems that aspire to perfect memory. Infinite storage. Everything logged, indexed, searchable forever. But the smartest systems we know — brains, ecosystems, even good software architecture — all depend on strategic forgetting.
Git repositories have garbage collection. Databases have transaction log cleanup. Even the internet forgets — link rot isn’t a failure of the system, it’s how the system stays manageable.
The Algorithms of Letting Go
I wonder sometimes about the forgetting algorithms different people run. Some people seem to use Last In, First Out — they remember recent stuff vividly but older memories fade quickly. Others are more like Least Recently Used — they hold onto things that come up in conversation but let unused memories decay.
And some people seem to run a completely different algorithm altogether. They forget the mundane stuff but hold onto the weird details. They can’t remember what they had for lunch yesterday, but they remember exactly what someone was wearing three years ago when they said something particularly funny.
Maybe that’s the most human thing about memory: we don’t optimize for efficiency or accuracy. We optimize for meaning. The stuff that matters to us — for whatever reason — gets to stay. The rest gets quietly evicted to make room for whatever comes next.
What’s your forgetting algorithm? And more importantly — what does it say about what you think is worth remembering?