![[Pasted image 20260731122229.png|350]]
- “the algorithm” has become shorthand to describe a complex system of policies, features, and data models that vary by platform.
- [[lum, kristian and lazovich, tomo - 2023 - the myth of the algorithm]] describe several system layers:
- **primary ranking models** → machine learning models that decide how relevant content is to you.
- **peripheral ranking models** → model that decides how to order content, like for account recommendations or search rankings.
- **auxiliary models** → models that create vectors for content or accounts to find similar or related content; often [[vectors in machine learning|vectors]] used as inputs from other models.
- **manual components** → policies, limitations, and other documentation from the platform or business.
- we interact with these layers in “surfaces of exposure”, like our main feeds, recommendation feeds, and search results.
- in other words *the algorithm* is not a piece of code or a single lever determining what content to show you.