Hater Mode — stress-test your launch against 12 internet personas before the internet does it for you

Every launch gets roasted. The only variable is whether the roast happens in your editor a week before launch, or in a Hacker News thread an hour after — permanently attached to your product's name in search results.

Hater Mode is a prompt technique that moves the roast earlier. You ask an AI to critique your product in character as the specific audiences who will actually see it. Generic "give me feedback" prompts produce polite mush; persona-anchored criticism produces the exact sentences that will appear under your launch post.

Why personas beat "be critical"

A model told to "be brutally honest" gives you one voice: vaguely negative, abstract, easy to dismiss. But internet audiences have distinct failure-finding styles, and each one finds different bugs in your story:

• Hacker News attacks the technical claims and the business model ("so it's a CRUD wrapper around an API you don't control?")
• Product Hunt skeptics compare you to the three lookalikes that launched last month
• Reddit sniffs out astroturfing and pricing dark patterns
• Designers catch the inconsistent spacing and the AI-generated logo
• Software testers find the empty-state and error-path holes
• Laypersons reveal that your headline is incomprehensible to anyone outside your bubble
• Late adopters ask the question that matters most: "why would I switch from the thing I already use?"

One critique pass per persona = seven different audits for free.

The prompt

You are running HATER MODE on my product. Below is my [landing page copy /
README / pitch / demo description].

Critique it in character as each of these 12 personas, one at a time:
1. Product Hunt commenter (seen 1,000 launches, bored, comparison-happy)
2. Hacker News commenter (technical, pedantic, allergic to hype)
3. Reddit user in the relevant subreddit (suspicious of self-promotion)
4. X/Twitter quote-tweeter (dunks for engagement, takes things out of context)
5. Facebook group member (non-technical, practical, price-sensitive)
6. LinkedIn thought-leader (asks about the business, the moat, the market)
7. IndieHackers builder (supportive but has seen this exact idea fail twice)
8. Professional software tester (probes edge cases, errors, empty states)
9. Complete layperson (doesn't understand any jargon)
10. Professional designer (typography, spacing, visual hierarchy, taste)
11. Late adopter (happy with current tool, needs a switching reason)
12. Pure hater (bad faith, will twist anything — what's the worst-faith read?)

Rules:
- Each persona speaks in their own voice, 2-4 sentences, specific to MY
  product — no generic criticism that could apply to anything.
- No softening, no "but overall it's great."
- After all 12: list the 5 objections most likely to actually appear on
  launch day, and for each, whether the right response is (a) fix the
  product, (b) fix the copy, or (c) prepare a reply.

[PASTE YOUR MATERIAL HERE]

How to use the output

The 12 critiques are entertainment. The closing list is the deliverable — and the (a) fix product / (b) fix copy / (c) prepare reply triage is what makes it actionable:

• Fix the product items are real gaps the personas found (the tester persona is disturbingly good at finding actual missing features).
• Fix the copy items are misunderstandings — if three personas misread what your product does, the layperson didn't fail, your headline did.
• Prepare a reply items are objections that are wrong but inevitable. Draft the comment response now, calmly, instead of at 1 a.m. on launch day with your heart rate up.

Two tips from running this on real launches

Run it twice: before and after your fixes. The second pass tells you whether your changes actually closed the objections or just moved them.

Take persona 12 seriously. The bad-faith read feels unfair — that's the point. If the worst-faith interpretation of your product is "it's a privacy nightmare" or "it's a wrapper with a markup," someone will say it, and screenshots last forever. Knowing your worst-faith read before launch is the cheapest reputation insurance available.