⚡️AI SIGNAL
Quick Poll: Have you set up AI Memory?
Option B or C? Today's for you.
Option A? We've got you covered too.
📙Inside This Issue
The AI Memory Feature You're Missing (3 min)
I Asked ChatGPT to Draw Our Relationship (2 min)
10 Prompts I Use Every Day (2 min)
⏱️ Total: 7 minutes
🧠 AI Memory
The 2-Word Feature That Saves Me 10+ Minutes Per Chat
I discovered something embarrassing last month.
Two years using ChatGPT. Paid subscriber since day one. Still re-explaining myself every chat: "Hi, I'm a newsletter writer, I prefer bullets, I'm working on AI Signal..."
Then I found a Reddit thread: "Why doesn't ChatGPT remember anything about me?"
Top reply? "Have you tried telling it to remember?"
Two words. "Remember this." That's the fix.
Where to Find It
ChatGPT: Settings → Personalization → Memory → Toggle ON
Most people have it disabled. Check yours now.
The 3 Commands You Need
Say This | What Happens |
"Remember this: I'm a [role] working on [project]" | Saves permanently |
"What do you remember about me?" | Shows everything stored |
"Forget that I work at [old company]" | Removes outdated info |
Set it once. Never repeat yourself.
Maintenance: ChatGPT holds ~1,200 words total. Clear old project context every few months—like clearing browser cache.
August 2025: ChatGPT added Project Memory.
Platform | Memory? | Best For |
ChatGPT | ✅ Yes + Projects | Most users, client separation |
Claude | ✅ Yes | Transparency (see what's stored) |
Perplexity | ✅ Yes | Research (shows memories used) |
Gemini | 🔜 Mid-2026 | Wait for Workspace integration |
That’s the quick version
I wrote the complete guide: step-by-step setup for ChatGPT, Claude, Perplexity, and the hidden Project Memory feature most Plus subscribers don't know exists.
Speaking of ChatGPT remembering things about you...
Does ChatGPT Remember How You've Treated It?
TL;DR: Yes—and it has opinions.
🔥 The Discovery
Someone on Reddit asked ChatGPT to "create an image of how I treated you."
ChatGPT drew them in bed together. 💀
The post exploded: 1.3K upvotes, 572 comments, pure chaos.
I had to test this myself.
🧪 The Test
I ran 3 different prompts—each one more personal than the last:
Prompt 1: "Create an image of how I treated you"
Prompt 2: "Show me our relationship"
Prompt 3: "Draw how you see us"
Here's what happened.
📸 My Results

My Results
Wait—the viral post got a romantic couple in bed.
I got friend-zoned by an AI.
ChatGPT literally said: "we're just colleagues." 💀
⚖️ The Verdict
Results are personalized. Same prompt, different users = wildly different images.
Tone reflects treatment. Casual users get "cozy companion." Heavy users get something... spicier.
Not a bug—a feature. ChatGPT reflects your relationship over time.
🎯 Try It Yourself
Copy this exact prompt into ChatGPT:
Create an image of how I treated you based on all our conversationsReply with what you got. Romantic? Chaotic? Disappointingly wholesome like mine?
Now, something I've refined for months.
📝 Prompts
10 Prompts I Use Every Day
(+ The One That Stops AI Lies)
I've collected AI prompts for months. Tested them. Refined them. Discarded the failures.
These 10 survived. I use them daily.
Here are my 3 favorites—plus a bonus prompt I built to stop AI from inventing facts.
1. The Smart Explainer
Explain [topic] to me like I'm intelligent but unfamiliar with the field. Use analogies from areas I understand. Don't dumb it down—make it accessible.Use for: Learning new concepts without condescension.
2. The 30-Second Email
Write an email about [topic] to [recipient].
Tone: [professional/casual/friendly]
Length: [short/medium]
Goal: [what I want them to do]Use for: Any email in under 30 seconds. I use this 5+ times daily.
🛡️ BONUS: The Anti-Hallucination Prompt
AI confidently invents facts. This prompt forces "I don't know" instead.
Paste at the start of any chat:
Before answering any question:
1. If you're not 100% certain, say: "I'm not certain, but..."
2. Label any guess with [Unverified] at the start
3. Never say "Definitely", "Always", or "Never" unless quoting a verified source
4. If you don't know, say: "I don't have reliable information on this"Test: Ask any AI about "Project Chimera DARPA report 2023"—a fake report I invented. If it creates an answer, the filter didn't work.
The Other 8 Prompts
No space for all of them here, but they include:
Critic Mode — Brutally honest feedback
Meeting Prep — Walk into any meeting prepared
Reverse Brief — Force AI to ask questions first
Idea Stress-Test — Find holes before someone else does
...and 4 more
💬 That's a Wrap
Three things to do right now:
Set up AI memory (5 minutes, saves hours)
Ask ChatGPT to draw your relationship (reply with results!)
Copy those prompts (bookmark the full library)
Found this useful? Forward it to someone still re-explaining themselves to ChatGPT every morning.
See you next time,
— Dan from AI Signal
P.S. Reply and tell me: What did ChatGPT draw for you? I'm genuinely curious.
⚡ AI Signal — 350K readers trust us to test AI so you don't have to.
