Stop Multicasting Slop

The essay argues that AI can be useful for reviewing work, documentation, reporting, and some visual design, but relying on it to think or write can weaken the result. In communication aimed at many people, the creator has a responsibility to respect their time: a keynote, publication, or company-wide message should make its point clearly and usefully. Audiences judge the finished document or presentation, not how it was made, so AI assistance neither excuses poor work nor removes the creator’s ownership. Presenters still need to shape complex material into something understandable, and writers need to recognize what good writing looks like.

The author evaluates communication by whether it conveys its purpose efficiently, builds understanding through sound structure and appropriate information density, and uses tone to keep audiences engaged. AI makes careless work easy to produce at scale: slide decks can have weak titles, confusing terms, awkward layouts, and too much or too little information; prose can be repetitive, oddly structured, verbose, or choppy. Such flaws burden readers and can cost a speaker their audience’s respect. Because language models tend toward probable, familiar output, they may also miss the unusual ideas that make writing worth reading. The practical advice is to discard bad drafts rather than sink time into polishing them, and to refine the craft of communication instead of using AI as a substitute for judgment.