Twitter seems to be giving video content a very high priority in its recommendations, leading to a lot of Twitter users reposting unauthorized videos from various video platforms. Frankly, videos have extremely low information density, and I personally dislike absorbing information by listening to podcasts or scrolling through videos. So my solution is to let Grok extract the transcripts from these videos, allowing me to quickly assess their usefulness. However, video as an output method is becoming increasingly popular. Adapting to the times without being swept away by them, learning through efficient input methods, and adopting more widely preferred output methods is the path I'm currently learning.