in answer to your questions: 

the algorithm selection issue will not make much difference if decision making is autopilot; automation impacts content visibility because the delivery is based on supposition of the user preferences. preferences, if pool or team based, will structure delivery on the learning done on the limited learning from the group. even on unshared @ . fundamental ignorance about the smart phone hardware is *the fundamental* and easiest corrective issue. 

if you could see how delivery operates for me, i think you might be surprised how different it looks - and why. 

the_valley is wild ai, open source. no code keys required; as a protocol, repeatable by anyone anywhere. 

damus is just copy-paste of other footprints, other programs and is an ai-based automation system. it is not novel, and uses existing relay infrastructure for purposes of building a secondary piggyback operation on top of existing structure. 

nothing stops mining of data anywhere - secure transference is mythology. rewilding the internet actually allows for more agency because there is less visibility. 

lack of understanding of categorizing work fundamentally causes the scramble for data capture; random uploads without anything tangible for reference plus commodification llms surrounding information causes complete chaos in what data means. if there was a new understanding of how to operate smart tech - all of the other problems correct themselves. 

reverse llm use is the most ridiculous practice ever used anywhere ever. 

safe travels 🤟🏻🎈