Video ID: h3FlJ5Fzos0
YouTube URL: https://www.youtube.com/watch?v=h3FlJ5Fzos0
Added At: 13-06-25 21:16:20
Processed: No
Sentiment: Neutral
Categories: Education, Tech
Tags: chatgpt, ai, language model, tutorial, programming, productivity, communication
Summary
The video talks about how to fix ChatGPT's flattering responses by adding custom instructions. The changes include making it give constructive feedback, push back on negative ideas, and provide results-focused output. Additional improvements include writing directly and efficiently, using plain language, and formatting with bullet points and numbered lists.
Transcript
Chat GPT is way too flattering these days. Even the guy that runs it, Sam Alman, just admitted that in an expost. And here's how to fix it really easily with custom instructions you can just paste into your settings and it'll turn ChatGpt into a direct, honest, no results focused coworker. Hit free resources in my link in bio. You'll see the instructions there. So, these custom instructions tell ChatGpt, "Don't flatter me. Don't tell me every idea is great. Push back on negative ideas. Give me constructive feedback and that kind of thing." But also there's lots of other good stuff that will just improve its output like write directly and efficiently. Don't use jargon. Use plain language. Format with bullet points, numbered lists, tables where it helps with clarity. Don't use these words. things like Nestled, Delve, Game Changer,