Why One Tech Writer Avoids X — And How AI Could Fix It

A tech writer who goes by the name jefftk published a short essay explaining why he avoids X (formerly Twitter). He compares the platform to drugs: he thinks it would change him for the worse, and that once he started, he wouldn't be able to stop. Recently he cross-posted a few AI-related tweets because he had something important to say — and it went fine. The problem came afterward, when he kept checking notifications.

The bait, he says, is in the replies. He kept seeing posts that were confidently wrong about things he happens to know well, and the urge to jump in and correct strangers was strong. That pull doesn't just cost time — he already spends time on Reddit and Hacker News without regret. The real issue is the culture. X rewards people who raise the temperature, dunk on each other, and take sides, because fights drive engagement. He argues the algorithm pushes the crowd that way, and users internalize the drive to make their numbers go up. Many people cope by blocking aggressively and using a filtered feed, but he thinks the meanness seeps in through replies and quote-tweets anyway.

His most interesting suggestion is about AI. Since modern AI systems are good at judging text, he proposes using them as ranking signals: write down rules for good conversation — the kind of guidelines Hacker News or thoughtful online communities use — then have AI give more visibility to posts that actually follow them. The catch is obvious. Calmer feeds mean fewer clicks, and fewer clicks mean less advertising money. So the platform has to want that outcome, which it currently has little reason to.

For the rest of us, the practical lesson is that AI is already helping people stay informed without wading into the muck. He praises AI newsletters and roundups that quote important tweets and add context, letting him follow AI news — which traditional media covers poorly — without opening the app. Read the summaries, skip the food fight.