In short:
Image/video related to your interest = dopamine hit. Constant dopamine hits = addiction. Addiction = depression, less braincells, mental illness.
Text only = none of the above because no image/video = makes you think. And thinking is (obviously) good.
This is why the Gemini protocol is better than HTTP for me. All those media formats still work, but they’re usually embedded as links and you see primarily text. The spec design means that it can’t become the unhealthy thing that the internet has.