Instagram has expanded its algorithm personalization tools, allowing users to adjust the types of recommended content they see in their main feed. Users can now request more or fewer posts on specific interest-based topics like 'rescue dogs' or 'parenting humor,' giving them greater control over their experience. The update, driven by advances in large language models, aims to make algorithmic feeds feel more transparent and user-directed.

However, this control does not extend to one of the most common user requests: seeing more posts from accounts they follow. When users attempt to prioritize content from followed creators, the system returns a 'no results found' error. Instagram head Adam Mosseri acknowledged this limitation, explaining that the platform has shifted away from chronological, follow-based feeds in favor of algorithmic discovery.

Mosseri noted that personal updates have largely moved to Stories and DMs, making polished, infrequent posts from friends less engaging in a main feed. As a result, recommendations from unfamiliar accounts now dominate. While a separate 'Following' tab still exists, it remains hidden from the default experience. Instagram says it’s exploring deeper personalization, including mood-based feeds and content types, but for now, user agency has clear boundaries shaped by platform priorities.