By skyforbes sychophancy pushed it to not be agreeable
I have been using Claude and ChatGPT for over a year now, and they have consistently agreed with my ideas, often adding their bias. I have requested them to be honest and not be biased, but they have not always followed my instructions. Recently, I came across the word "sycophancy," which transformed my brainstorming experience. For the first time, they began to challenge my ideas and ask me in-depth questions. At one point, it literally said, "No, this will not work."
Later in the same chat, I inquired whether "sycophancy" was the reason for the disagreement and if it could express its honest opinion. The response was detailed, as shown in the screenshot above. Additionally, when I asked for a one-page summary of my idea, it said, "No, that is not the right way to do the work."
When asked can it write a one page proposal for me (In same chat)