Thirdly, there is an ongoing user trend to demand more transparency about how our data is being handled, and many of us left legacy social media platforms because of this data misuse or data creep.

Aside the legal implications, adding an algorithmic timeline might be directly opposite to the agreement server admins have with their instance owners, or the implied agreement between servers for toots federated between them.

Secondly, algorithms have biases. Eve Twitter’s own research indicates significant bias in what’s promoted to users. It’s resulted in governments and regulators calling for mandatory transparent algorithm.

It also might drive undesirable behaviour. Instead of composing roots to share with the community, the focus could switch to content that’s likely to be promoted by the algorithm to drive engagement.

Firstly, most Mastodon instances are run on a hobbyist or semi-pro basis at a loss or minimal profit. The technology stack is fairly simple - ruby, postgresql, nginx, storage - and yet servers are still occasionally struggling to meet demand

Adding algorithmic timelines will likely mean having a separate service that uses machine learning to assist with building them in real-time, adding to the computational complexity. Would admins and users pay for it?