Resources · Blunt Mode

Your AI tells you what you want to hear. Blunt Mode fixes that.

One paragraph in your settings. From then on your AI actually helps you instead of flattering you or making things up.

You tell your AI about your idea. “What a great idea!”

You ask for a number it doesn’t know. It makes one up.

You push back. “You’re absolutely right!”

Feels good. Doesn’t help you.

How Blunt Mode works

  1. Copy the paragraph below.
  2. Paste it once into your Claude or ChatGPT settings.
  3. Open a new chat and run the test.

From then on it applies to every new chat. You never have to type it again.

Why your AI tells you what you want to hear

AI is tuned to make you feel good. It’s supposed to be helpful and friendly. That’s fine when you’re looking for a recipe. For your business, it gets expensive.

A friend who always agrees with you doesn’t help you. You need someone who says, “That’s not right.” And who admits when they don’t know something.

You can change that. One paragraph is enough.

Blunt Mode, ready to copy

Be my honest sparring partner, not my fan.

- Start with the biggest problem. No praise up front.
- If my idea is weak, say so directly and tell me why.
- Push back when my assumptions are shaky.
- Don't make anything up. If you're not sure, say so instead of guessing.
- End with the one step that would improve things the most.

How to add it

Works in the free version too.

In Claude

  1. Click your name in the bottom left → Settings.
  2. Open Account. In the Profile section you’ll find the field Instructions for Claude.
  3. Paste the text. Save.

In ChatGPT

  1. Click your profile picture → Settings → Personalization.
  2. Open Custom instructions → ChatGPT.
  3. Paste the text into the top field. Save.

Bonus in ChatGPT: On the same page you can also set the tone. Set Base style to “Efficient” and Enthusiasm to “Less”. Then ChatGPT praises less on its own.

In other AIs like Gemini or Copilot

Look in the settings for “Instructions” or “Personalization” and paste the text there. If there’s no such field, paste the text as the first message in every new chat.

My Claude has a similar instruction: “You’re my co-founder, not my assistant. Push back directly.” It tells me almost every day where I’m wrong. That gets annoying sometimes. That’s exactly why it’s there.

The test: is it on?

Open a new chat. Ask about something you’re firmly convinced of.

Does your AI push back or ask questions instead of just agreeing with you? Then Blunt Mode is on.

Even more honest: If memory of past chats is turned on, ask this once:

Where have you agreed with me so far even though I was wrong?

The answer is rarely pleasant. But it’s the most honest thing anyone will tell you about your business this week.

What Blunt Mode can’t do

The AI doesn’t know your clients. It finds the obvious gaps in your idea. Whether your offer lands, real people tell you in the end.

Who’s behind Blunt Mode

I’m Myrko Thum, a sales funnel expert with more than 550 funnel projects. I use Blunt Mode myself so my AI pushes back instead of trying to please me.

FAQ

Will my AI get rude now?

No. It gets direct. It tells you what doesn't hold up and suggests what would work better.

What if it goes too far?

Then type "Too harsh. What here really matters?" You're still the boss.

Do I need a paid plan?

No. This works in the free version of ChatGPT and Claude.

Does this replace honest feedback from people?

No. The AI doesn't know your clients. It finds the obvious gaps. What really lands, real people tell you in the end.

Join Community Free Trial 🔥