People prompt every AI model the same way and get worse output for it.
Claude wants structure. GPT wants role + context.
I’ve seen $300 courses that skip what’s on page one of this.
How to rewrite 1 prompt for 27 models 👇 free.
99% of people are running the same prompt across every AI model and wondering why the output is mid.
Each model wants to be prompted
differently. Claude ≠ GPT ≠ Gemini.
I made a sheet on how to rewrite ONE prompt for 27 models.
Comment “HONE” and I’ll DM it to you.
@shushant_l Midjourney wants native syntax — not ChatGPT prose.
Rough: "Most people are using 10 AI tools to do what 1 workflow can do."
→ Most people are using 10 AI tools to do what 1 workflow can do., cinematic lighting, shallow dep…
https://t.co/WZ3ROvz7L1
@ElsaSofia__AI ChatGPT wants native syntax — not ChatGPT prose.
Rough: "BREAKING: If you're not using ChatGPT for travel, you're making …"
→ You are an expert assistant. Task: BREAKING: If you're not using ChatGPT for travel, you're maki…
https://t.co/ELKzOvhum8
AI models change every month.
The syntax changes faster than the threads.
Plain English → Midjourney-native on the card.
Drop your rough line — wrong syntax is expensive.
@hrswatigupta Rough: "BREAKING: Claude can now build a full Instagram business that runs on au…"
Hone for Claude:
<task>BREAKING: Claude can now build a full Instagram business that runs on autopilot.</task> <constraints>Be …
Hone does this for you →
https://t.co/iPKhoTVJrU…
@Alexie_Ai Rough: "BREAKING: AI can now design at Apple-level creative standards — for free…"
Hone for Claude:
Create a comprehensive analysis of the claim that AI can now design at "Apple-level creative standards" for fr…
Hone does this for you →
https://t.co/iPKhoTVJrU…
@Rixhabh__ Rough: "BREAKING: Claude can analyze a stock better than most retail traders."
Hone for Claude:
<task>BREAKING: Claude can analyze a stock better than most retail traders.</task> <constraints>Be explicit. N…
Hone does this for you →
https://t.co/TbimPBvb16…
@Raul_IA_Prod Rough: "My wife sent 85 applications. Zero responses."
Hone for Claude:
<task>My wife sent 85 applications. Zero responses.</task> <constraints>Be explicit. No filler.</constraints> …
Hone does this for you →
https://t.co/BxySfxRxk4
AI models change every month.
The syntax changes faster than the threads.
Plain English → Grok 4-native on the card.
Drop your rough prompt — I'll show the honed version on the card.
@AnishJaitwar Rough: "BREAKING: AI can now design like Apple-level creative directors (for fre…"
Hone for Claude:
<task>BREAKING: AI can now design like Apple-level creative directors (for free).</task> <constraints>Be expli…
Hone does this for you →
https://t.co/iPKhoTVJrU…
@piyascode9 Rough: "BREAKING: AI can now design like Apple-level creative directors (for fre…"
Hone for Claude:
<task>BREAKING: AI can now design like Apple-level creative directors (for free).</task> <constraints>Be expli…
Hone does this for you →
https://t.co/iPKhoTVJrU…
@TechAI_X Rough: "BREAKING: I asked Claude to upgrade my LinkedIn profile."
Hone for Claude:
<task>BREAKING: I asked Claude to upgrade my LinkedIn profile.</task> <constraints>Be explicit. No filler.</co…
Hone does this for you →
https://t.co/22nFCJoANT
🚨 9 Grok prompts went viral this week.
You only need one rough question.
Type: "what are people saying about AI prompts on X right now"
Hone → Grok-native mega prompt in 3 sec ↓
https://t.co/E7EkdD3ABI