Episode
How I AI: Stop writing like a robot - ban these AI words and phrases
- Podcast
- How I Work
- Published
- Mar 29, 2026
- Duration seconds
- 1145
- Processing state
not_requested
Actions
POST https://stenobird.com/v1/public/podcasts/how-i-work-960132/episodes/how-i-ai-stop-writing-like-a-robot-ban-these-ai-words-and-phrases/transcription-requests
Idempotently request low-priority transcript generation for this episode.GET https://stenobird.com/podcast/how-i-work-960132/how-i-ai-stop-writing-like-a-robot-ban-these-ai-words-and-phrases.md
Read the agent-friendly Markdown representation of this episode resource.
Summary
Download Inventium.ai 's anti-AI slop prompt here: https://amantha-imber.kit.com/d5f746bd3e You can always tell when something's been written by AI. The LinkedIn post that opens with "in today's fast-paced world." The article that ends with "it's not about X, it's about Y." Technically fine, but somehow hollow. Like it was written by nobody in particular. The frustrating part is that this can happen even when you're doing the real thinking. You write the ideas, hand it to AI for a cleanup, and it flattens your voice, packs it with em-dashes, and strips out everything that sounded like you. In this How I AI episode, Neo and I get into the specific instructions you can give AI to keep your writing sounding human, and share some of the most overused AI words and phrases worth banning from your prompts entirely. Neo and I cover: Why sounding like AI is a credibility problem, even when you've done the actual thinking and are only using AI to refine and edit your work. The single most important instruction to give AI when asking it to clean up your writing. Why em-dashes are so hard to eliminate from AI output, and the specific phrasing (including pasting in the actual dash character) that gives you the best chance of getting rid of them. Using a ninth-grade reading level as a brief to keep language plain, accessible, and free of jargon. Why it's fine to let AI start a sentence with "and" or "but" occasionally, and which stiff transition words (like "additionally" and "moreover") to ban outright. What "corrective antithesis" is, why AI overuses it, and how to cut it down with a simple instruction and example in your prompt. Amantha's process for building a mega anti-AI slop prompt. How to build your own personal banned-words list, so AI stops using words that don't sound lik…