Charlie Hills

How to make a landing page in 5 minutes.

Using AI and your LinkedIn profile.

Creating landing pages usually takes hours.

Gamma makes it ridiculously fast:

1. Go to LinkedIn, click “More”, then “Save to PDF”
2. Head to Gamma(.)app, and select “Create new“
3. Select from the three options: “Upload a file“
4. Upload your profile and select “Webpage“
5. Edit content and fine-tune preferences.
6. Select from a pre-set Gamma theme.
7. Customise and add components.
8. Publish your landing page.

Optional: Add a custom domain.

You now have an optimised landing page.

This turns your LinkedIn into a live portfolio.

Perfect for freelancers, creators, and solopreneurs.

I partner with Gamma because it’s the tool PowerPoint wishes it was and it makes designing landing pages ridiculously fast and easy.

I rebuilt my LinkedIn strategy from scratch...
(How I went from zero to 55k+ in 365 days)

My posts were getting little to no engagement.

So, I decided to revamp my profile and strategy.

This was way back in January, when I first started.

Since then, I've been on a journey:

→ I gained 55k+ new followers.
→ I created 584+ LinkedIn posts.
→ I dedicated 365 days of energy.

Here’s how I did it:

1. Profile Optimisation
Start by optimizing your profile picture and headline.
Your profile picture should be profesh and approachable.
Your headline needs to be compelling and clear.
Use your About section to tell your story.
Use your Featured to drive traffic.

2. Content Ideation
Create a content calendar to plan topics in advance.
Use Justin Welsh's context matrix prompt in Claude.
Leverage Google's top search queries for your niche.
Keep content ideas in a running list as inspiration strikes.

3. Content Strategy
Incorporate a content matrix strategy.
Focus on TOFU, MOFU, and BOFU content.
Post three times a week to maintain consistency.
Continuously test and iterate content types.

4. Content Creation
Use AI tools like Claude or ChatGPT for ideation.
Write in a hierarchical format for better readability.
Use 1080x1350 pixel images for optimal display.
Include a strong call-to-action in every post.

5. Posting Content
Incorporate your brand colors in visuals.
Post Tues-Thurs for highest engagement.
Keep posts under 1,300 characters for mobile viewing.
Use “How to“ or “How I“ hooks for educational content.

6. Engagement Strategy
Make a list of 100 creators
Engage with them everyday
1/3 big, 1/3 same size, 1/3 prospects.
Save lists in a LinkedIn boolean search.

7. Analyzing Content
Export your data regularly.
Upload to ChatGPT or Claude.
Identify best posting days and content types.
Adjust your strategy based on performance insights.

8. Repurposing Content
Convert top-performing posts into multi-media formats.
Test different content types with your audience.
Maximize the reach of your best ideas across platforms.
Use successful concepts in new contexts.

9. Sending DMs
Use Sales Navigator for advanced filtering.
Send 15 personalized connection requests daily.
Follow up with creators and ideal customer profiles.
Qualify leads before sending tailored messages.

LinkedIn can be a powerful tool for growth.

Showcase your best work and drive traffic to your sites.

Need clarity or a hand implementing these strategies?

Send me a DM. I'm here to help you win on LinkedIn!

Follow Charlie for more. Reshare to help others.

LinkedIn has 1B+ users.

Follow these 40 AI Creators in 2025:

 1. Jim Fan
 2. Noelle R.
 3. Anna York
 4. Pooja Jain
 5. Pete Sena
 6. Andrew Ng
 7. Alex Banks
 8. Liza Adams
 9. Charlie Hills
10. Yann LeCun
11. Audrey Chia
12. Lex Fridman
13. Axelle Malek
14. Allie K. Miller
15. Paul Couvert
16. Ethan Mollick
17. Noam Nisand
18. Claire Farwell
19. Bojan Tunguz,
20. Andreas Horn
21. Andriy Burkov
22. Satya Nadella
23. Ruben Hassid
24. Lior Alexander
25. Roisin Bennett
26. Martin Crowley
27. Rowan Cheung
28. Pinar Demirdag
29. Eitan Yehoshua
30. Maëlle Bertrand
31. Aravind Srinivas
32. Florian Camiade
33. Sanchit Shangari
34. Valeriya Pilkevich
35. Sabahudin Murtic
36. Jérémy Grandillon
37. Anu Chandrasekar
38. Hassan Bin Arshad
39. Adam Biddlecombe
40. Katie King, MBA
    
The best AI education isn't in courses.

It's in your LinkedIn feed:

1. Engage with their content
2. Apply one insight this week
3. Share what you learn

Your network determines if you keep up.

Watch your AI knowledge compound.

Stop learning AI from one perspective.
Start learning from forty brilliant creators.

♻️ Repost to help others build their AI network.

8 tools run my entire LinkedIn operation.

Here's the exact workflow I use daily:

Step 1: Start with the idea dump

ChatGPT is my brain partner.
Not my writer. My organiser.

I throw messy thoughts at it.
It hands back structured design briefs.

Those briefs, they live in Notion.
My team's single source of truth.

Step 2: Visual game = Non-negotiable

Canva does double duty:
↳ Team creates from my briefs
↳ I build carousels when flying solo

But the secret sauce is Ideogram.

AI-generated carousel covers.
Unique visuals no one else has.

Step 3: The creation engine kicks in

Stanley writes my posts.
But here's what most miss:

It doesn't just write.
It learns my voice.
Studies what works.

Then writes like me, not like AI.

Try it: https://lnkd.in/eqGtRi54

If I need fresh angles, I use Perplexity.
Real-time insights beat recycled ideas.

Step 4: Video when it matters

VEED handles everything video:

- Screen recordings that don't suck
- Auto-subtitles that actually work
- Audio enhancement built in

Complex editing kills momentum.
Simple tools ship content.

Step 5: The inbox beast

200+ LinkedIn messages weekly.
Without Kondo, I'd drown.

Smart filters surface what matters.
Conversations that count remain.

Try it: https://lnkd.in/eY9dPrCm

Here's what this system really does:

Removes decision fatigue.
Each tool has ONE job.

- ChatGPT structures thinking
- Notion stores everything
- Stanley maintains voice
- Perplexity adds depth
- Canva/Ideogram handle visuals
- VEED manages video
- Kondo saves sanity

Total output: 7+ posts weekly.

Plus actual conversations.
Not automated garbage.

The tools don't create content.
They amplify your capacity to ship.

Most creators collect tools.
Winners build systems.

So you have two options, really:

1. Invest hours figuring it out yourself.
2. Work with someone who's cracked it.

Either way, I can help you.

Apply to work together: linked-agency.com

Which tool would transform your workflow?

Save this. Build your system tomorrow.

Repost ♻️ to help others escape tool chaos.

LinkedIn just exposed how their algorithm works.

Here's what they told me word for word:

On Wednesday there was a weird bug.

My posts were getting 10% of normal reach.
Notifications weren't showing up.

Something felt off.

I panicked. Was I shadow banned?

So I did what any rational person would do.
I raised a support ticket demanding answers.

12 hours later, LinkedIn Support replied.

Not about the bug (that fixed itself).

But they explained exactly how their algorithm works.

Before anyone comes at me saying "no one knows how the LinkedIn algorithm works" - this is LinkedIn's official response. Word for word:

1. Member Expectations

Once a member sees some posts with high engagement, they tend to assume that this will be the minimum expectation to follow. However, it's normal to have varying levels of engagement on different posts over the course of the year.

2. Relevance Algorithm

Our system scores each post based on various factors like content quality, media, and mentions. These posts are then distributed to feeds of followers and other members based on relevance, considering previous activity, similar pages followed, profile information, and online presence. With so many variables, engagement can vary from post to post.

3. Viewer Tolerance

As members use LinkedIn over time, they become less likely to engage with content posted by the member if they didn't engage the last time they saw a post by the same member. If they continue to not engage, they are less likely to be shown posts by the same author in the future. This cycle continues but can be balanced by new connections, followers or other viewers.

4. Freshness of Post Content

Sharing newer content tends to be more engaging. If you link to other content (eg another LinkedIn post, a website, etc), the age of that content will make a difference. Fresh, up-to-date material often leads to greater distribution.

5. Frequency of Posting

Frequency of posting activity can be a complex factor. Not posting often will result in less overall traffic to your profile and engagement on your content. However, posting too often can lead to reduced distribution when looking at each post separately as our feed relevance algorithm aims to show our members content from a wide variety of authors/sources, and not show too much content from the same author/source at the same time.

6. Duplicate or Near Duplicate Content

If a member shares a post that is identical or nearly identical to something they have recently shared, we intend not to show it to members who have already seen the original/first post to avoid showing duplicate content.

Turns out I wasn't shadow banned.

But this response changed how I think about LinkedIn.

Which factor surprised you most?

Save this. Repost ♻️ for others.

Google's Nano-Banana is the best AI image editor.

But 99% still prompt it wrong.

Here's the complete playbook:

𝗙𝗢𝗥 𝗕𝗘𝗚𝗜𝗡𝗡𝗘𝗥𝗦

Simple Object Replacement
↳ Swap one element for another instantly
↳ Perfect for quick mockups

"Make this living room into a cozy cabin with a fireplace"

Photo Restoration
↳ Breathe life into old, faded photos
↳ Brings memories back to life
"Colourise and restore this old black and white photo"

Conditioning
↳ Set the scene or mood with one line
↳ Easy transformations that just work

Prompt example: "Turn this photo of my dog into a fluffy sheepdog"

𝗙𝗢𝗥 𝗣𝗥𝗢𝗦

Composite Generation
↳ Combine subject + background + props
↳ Professional marketing visuals in seconds

"Reimagine the Ferrari driving down the desert highway at sunset. Cinematic lighting, dramatic shadows"

Mood & Atmosphere
↳ Control lighting and feeling with keywords
↳ Studio-quality product shots

"The bottle of perfume on a silk cloth, with soft focus and a dreamy, elegant feel. Macro shot"

Camera Language
↳ Frame the shot like a cinematographer
↳ Precise control over perspective

"A low-angle shot of a new athletic shoe. The shoe is on a running track, with the camera close to the ground, emphasising motion and power. Sharp focus on the shoe, with a blurred background"

𝗙𝗢𝗥 𝗔𝗗𝗩𝗔𝗡𝗖𝗘𝗗 𝗨𝗦𝗘𝗥𝗦

Tools: Kling 2.1, Higgsfield, Hailuo

Multimodal Fusion
↳ Morph products between scenes seamlessly
↳ Next-level product reveals

"Use the two images to create a video that morphs the headphones from the white background into the futuristic studio, with a seamless, sci-fi transition"

Character Consistency
↳ Transform while maintaining identity
↳ Perfect for before/after content

"Use the two images to create a video that seamlessly transforms the character's hairstyle, maintaining all other visual details to show brand consistency"

Narrative Flow
↳ Tell a story through transitions
↳ Complete stories in 10 seconds

"Use the two images to create a video that transitions from the person in the store to them confidently walking down the street, showcasing a successful shopping experience"

The difference between amateur and pro?

Amateurs describe what they want.
Pros direct the AI like a film crew.

Your move: Pick one technique. Master it today.

Save this for later. Repost ♻️ for others.

“Just use ChatGPT for everything.“

Bad advice.

I tested 5 AI models for deep research.

The winner? ChatGPT wasn't even close.

Here's how to use AI for deep analysis.

1. Pick any AI model with research capabilities.
2. Enable their deepest analysis mode.
3. Use this type of prompt:

____

Act as a senior marketing strategist and AI researcher. Conduct a comprehensive deep-dive into how artificial intelligence is transforming LinkedIn marketing strategies. Explore the evolution, current use cases, and future trends of AI applications on LinkedIn - including content personalization, ad targeting, lead generation, and automation tools. Analyze the impact on B2B marketing performance, user engagement, and ROI. Incorporate insights from case studies, industry reports, and expert commentary. Structure the findings as follows:

1. Executive Summary
2. Introduction to AI in Digital Marketing
3. LinkedIn-Specific AI Use Cases
4. Impact Analysis (metrics & outcomes)
5. Opportunities and Risks
6. Future Outlook
7. References.

Keep the tone formal and analytical, as if preparing for a CMO strategy briefing or whitepaper.

____

1 - Gemini (29/30) ★★★★★
☑︎ 47 confirmed references pulled.
☑︎ Peer-reviewed studies included.
☑︎ Trade articles cross-referenced.

Perfect scores across accuracy, sources, insights, and recommendations.

2 - Claude (28/30) ★★★★★
☑︎ 18 references, but quality over quantity.
☑︎ Enterprise case studies analyzed.
☑︎ Academic papers properly cited.

Lost a point on structure, but insights remained top-tier.

3 - Perplexity (26/30) ★★★★
☑︎ 25 references with cross-platform data.
☑︎ Ad performance benchmarks included.
☑︎ Strong on accuracy and sources.

Weaker on actionable recommendations.

4/5 - ChatGPT & Manus (24/30) ★★★
- ChatGPT: 10 references, mostly blogs and FAQs.
- Manus: 11 references, heavy on LinkedIn docs.
- Both delivered balanced but basic reports.

Of course Google won.

They've indexed the internet for 25 years.
What shocked me: how badly others failed.

♻️ Repost to help someone research smarter.

Every guru I see posting Claude commands is lying.

I made them into 20 skills that (actually) work:

/godmode, /ghost, /L99, /OODA.

These are not real commands.

Claude has:

- No secret command parser
- No hidden capability ceiling
- No concealed unlock codes

They are prompt prefixes dressed up to look like developer shortcuts.

So I did something different.

I took every viral "command" and built actual working versions of each one.

There are 20 Claude skills in total.

Each one does exactly what the gurus promised.

Subscribe (free) and the link to all 20 skills lands in your welcome email → https://lnkd.in/egsC99qy

Repost ♻️ to help someone in your network.

P.S. Claude Code commands are real.

Most people upload a PDF to AI and call it a day.

They're missing the full knowledge system.

NotebookLM runs on Gemini with 1 million token context.
It's your AI research brain that actually cites sources.

Master the complete system:

✦ Gemini 2.5 Flash
Core model with 1M token window.
Fast synthesis across massive corpuses.

Example prompt:

“Summarise these research papers with citations.”
“Extract recurring themes from these reports.”

✦ Gemini 2.5 Pro (Plus tier)
Deep reasoning for enterprise work.
Complex analysis and compliance-grade output.

Example prompt:

“Create an executive briefing from internal reports with cited evidence.”

1 - Upload Sources

PDFs, Google Docs, Slides, YouTube, websites.
Layout-aware parsing preserves structure.

Up to 50 sources per notebook.
Tables, images, formatting intact.

2 - Source-Grounded Chat

Every answer cites your uploads.
No hallucinations. Just your content.

"What does source 3 say about pricing?"
Click citation → Jump to exact passage.

3 - Briefing Docs
Executive summary across all sources.
Auto-formatted insights in seconds.

4 - Study Guides
Questions, concepts, key takeaways.
Your content becomes curriculum.

5 - Timelines
Chronological events extracted.
Historical context instantly.

6 - Mind Maps & Concept Visuals
Automatic semantic clustering.
See connections across sources.

7 - Overview Slides (Plus/Workspace)
Research to presentation deck.
AI-generated visuals included.

8 - Audio Overview
Two AI hosts discuss your content.
Podcast about YOUR research.

9 - Video Overview (Plus tier)
Narrated explainer with slides.
Your briefing becomes multimedia.

10 - Discover Sources (Deep Research)
Find credible supporting material.
Expand literature reviews automatically.

"Find peer-reviewed papers on this topic"
Adds relevant sources to notebook.

11 - Public & Shared Notebooks
Team research hubs with citations.
Real-time updates, controlled access.

Inline citations preserved on share.
Everyone works from same knowledge.

Use these power prompts:

✦ Summarise Precisely:
"Summarise in 300 words by theme with citations"

✦ Compare Findings:
"Compare insights across reports, highlight contradictions"

✦ Extract Decisions:
"List all strategic actions with source links"

✦ Create Brief:
"Generate: Context → Findings → Implications → Next Steps"

✦ Audio Script:
"Write Overview where Host A explains, Host B challenges"

Workflows to try:

Literature Review:
→ Upload papers to notebook
→ Generate briefing doc
→ Use chat for specific queries
→ Create audio for commute review

Team Knowledge Base:
→ Upload all project docs
→ Generate study guide for onboarding
→ Share notebook with team
→ Auto-update with new sources

Content Creation:
→ Upload competitor content
→ Generate comparison analysis
→ Create mind map of themes
→ Export slides for presentation

Which workflow replaces your current process?

Save this. Test it on your next project.

Repost ♻️ for others.

Most people only use ChatGPT for chat.

They’re missing the real value.

Users ignore built-in tools.

They waste hours juggling 10+ platforms.

Use ChatGPT like a productivity suite:

1 - Search (Web Browsing)
Real-time info with citations and trusted sources.
Ask it about current events, market data, or trends.

Prompt: “What’s Nvidia’s stock price right now?”

2 - Vision (Image Input & Generation)
Upload screenshots, diagrams, or photos.
Generate or edit images straight from text.

Prompt: “Here’s a floor plan - suggest furniture layouts.”

3 - Camera Mode
Stream your mobile camera for live guidance.
Perfect for troubleshooting, building, or debugging.

Prompt: “Watch my screen and walk me through importing a CSV in Excel.”

4 - Voice Mode
Chat hands-free while driving, cooking, or working.
It listens, thinks, and responds - all in real-time.

Prompt: “Explain the plot of Dune while I drive.”

5 - File Uploads
Drop in PDFs, Excel sheets, or PowerPoints.
Summarise, answer questions, or reformat your content.

Prompt: “Summarise the main findings in this 40-page report.”

6 - Data Analysis (Code Interpreter)
Runs Python in a secure sandbox.
Clean data, run analysis, and generate charts or tables.

Prompt: “Plot average monthly temperatures in °C from this CSV.”

7 - Canvas (Collaborative Workspace)
Side-by-side editor for documents or code with live preview.
Co-create landing pages, resumes, and prototypes.

Prompt: “Create a landing-page mock-up with shadcn components.”

8 - Memory (opt-in)
ChatGPT remembers details you choose across chats.
Preferences, name, and goals - all tailored over time.

Prompt: “Remember that I prefer formal greetings.”

9 - Custom Instructions
Set global preferences for tone, style, and context.
Every new chat will match your voice and needs.

Set up: Fill out Settings ▸ Custom Instructions → Save.

10 - Projects
Organise chats and files under one goal.
Ideal for long-term research or collaboration.

Set up: “Projects ▸ New Project → Add files → Continue working.”

11 - Scheduled Tasks (Automations)
Set reminders, recurring reports, or alerts.
Your new AI-powered assistant runs on autopilot.

Prompt: “Every Monday at 08:00 remind me of my week’s priorities.”

12 - Custom GPTs
Build niche assistants with custom logic and tools.
Upload knowledge files, set rules, and share.

Set up: “Explore GPTs ▸ Create your own → Add instructions + tools.”

13 - GPT Store
Browse and use GPTs others have published.
From video editing to legal advice - it’s all here.

Prompt: “Find a GPT that helps with meal planning.”

ChatGPT isn’t just a chatbot anymore.

It’s your new research analyst, designer, and planner.

The secret to speed and scale is using every tool.

Start now and never go back to basic search.

Follow Charlie Hills. Repost to help others.

It took me 9+ months to find these GPTs...

These changed everything.

1. Pinned Comment GPT
+ Writes engaging LinkedIn pinned comments.
- Might lack the personal touch of real conversation.

2. Hook Generator GPT
+ Crafts viral social media threads effortlessly.
- May sometimes lack unique insights.

3. Video GPT by VEED
+ Automates video creation with text prompts.
- Risks feeling less tailored to the audience.

4. Headline GPT
+ Delivers perfect LinkedIn headlines fast.
- Can feel too formulaic over time.

5. Calendar GPT
+ Plans LinkedIn content for consistent visibility.
- Might miss opportunities for spontaneous ideas.

6. Whimsical Diagrams GPT
+ Turns ideas into visual charts and mind maps.
- Visualizations can feel generic without creativity.

7. Consistent Character GPT
+ Builds characters with varied expressions and poses.
- Can generate inconsistent results.

8. Prompt Maker GPT
+ Optimizes prompts for better AI outputs.
- Could miss user-specific context.

9. Landing Page Creator from HubSpot
+ Creates and edits HubSpot landing pages seamlessly.
- May need manual tweaks for complex designs.

There are smarter ways to work with AI.

Use them. Don't abuse them.

Follow Charlie for more.
Reshare to help others.

AI can automate 80% of your boring tasks.

You just need the right prompts.

Most of your daily work could be faster.

You’re just not automating it (yet).

Here's Adrien's 7-step automation playbook:

1. List Your Time Drains
Open a doc. Brain dump for 5 minutes.

☑ Daily emails you copy-paste
☑ Weekly reports you format
☑ Social posts you schedule
☑ Data you manually update

Every repetitive task = automation opportunity.

2. Pick ONE to Start
Don't automate everything at once.

Choose the task that:
☑ Takes 30+ minutes weekly
☑ Follows a clear pattern
☑ Annoys you the most

Small wins build momentum.

3. Ask ChatGPT o3 to Help

“How can I automate this task using a GPT: [INSERT TASK]? Show me the steps, required tools, and give me a sample prompt I can run.”

Or try:

“I do [INSERT TASK] daily. Here’s how I do it: [DETAILS].
Help me automate this. Break it down. Tools + prompts please.”

You’ll get:
↳ A process map
↳ Smart tool suggestions
↳ Prompts ready to run

4. Now test.
Copy, tweak, and rerun.

☑ Make it useful
☑ Make it consistent
☑ Make it save real time

You’re building muscle.

5. Ask ChatGPT-4o:

“Write a system prompt for an AI agent that automates this task for me.”

You’ll get:
↳ A custom agent role
↳ Safety controls
↳ Logic flows

6. Build a Custom GPT
Go to ChatGPT → Explore GPTs → Create

☑ Name it after the task
☑ Paste your automation prompt
☑ Add any reference files
☑ Test with real examples

Your personal automation assistant is born.

7. Connect to Your Workflow
Use Zapier, Make or n8n to trigger automatically:

☑ Email arrives → GPT processes → Sends response
☑ Form submitted → GPT formats → Updates sheet
☑ Calendar event → GPT creates → Posts content

No manual work required.

Before automation: 3 hours/week on reports
After automation: 15 minutes reviewing

That's 11 hours back per month. 132 hours per year.

What would you do with 3 extra work weeks?

♻️ Repost to help others automate repetitive tasks.

I built this landing page in 2 minutes.

Using AI and my LinkedIn profile.

Gemini just changed the game.

Manual web dev is overrated.

Here’s a better way:

1. Go to LinkedIn, click “More”, then “Save to PDF”
2. Head to Gemini, select the “Flash“ latest model
3. Select the Canvas feature for visual code generation
4. Prompt: “Create a responsive landing page for [X]”
5. Upload your LinkedIn PDF

Gemini builds a full landing page (with working links!)

You can now:

☑ Export the HTML
☑ Drop into any website builder
☑ Go live in just minutes

On my first try, it looked 🔥

Follow Charlie Hills. Repost to help others.

I typed my name into Google AND Perplexity.

First result? My LinkedIn profile.

LinkedIn now dominates AI search results.

It outranks company websites consistently.

1. Perplexity AI heavily favours LinkedIn profiles.
2. Google's AI pulls LinkedIn for most queries.
3. Claude and ChatGPT cite it as top source.

AI models trust LinkedIn's structured data.
They give it special ranking preference.

Translation: LinkedIn beats your website.

So I optimised my LinkedIn for AI search.

Here's my complete 9-step playbook:

1. Add Alt Text to Images
↳ Describe your visuals clearly
↳ Use relevant keywords naturally
↳ This helps screen readers + SEO

2. Optimise Your Headline
↳ Use service + audience + expertise
↳ Include your power keywords
↳ Keep it natural and readable

3. Use Keywords in Posts
↳ Mention them naturally in text
↳ Repeat across different topics
↳ Avoid keyword stuffing always

4. Skip Hashtags (Seriously)
↳ LinkedIn ranks on semantics now
↳ Not hashtag signals anymore
↳ Use plain-language keywords instead

5. Maximize Your Skills Section
↳ Add 40–50 relevant skills minimum
↳ Align with your job titles
↳ This boosts search appearance

6. Publish LinkedIn Articles
↳ They get indexed by Google
↳ Rank for your niche searches
↳ Include keywords + helpful links

7. Edit Your Profile URL
↳ Customize for name + niche
↳ Share clean links externally
↳ This signals credibility strongly

8. Add Keyword-Rich About Section
↳ Write naturally, never stuffed
↳ Include your target industries
↳ Update this section quarterly

9. Engage in Comments
↳ Comment with niche language daily
↳ Adds visibility in keyword clusters
↳ Signals your topical authority

The window is closing fast.

In 6 months, AI search dominates everything.
Your LinkedIn isn't just a resume anymore.
It's your AI search identity now.

Are you optimised for what's coming?

Save this. Repost ♻️ for others.

The biggest LinkedIn lie? "Just post great content."

Here's what actually drives engagement in 2025:

1. The Core List Strategy

Build your circle of relevance:

- Peers in your space
- People who show up
- 10-100 people who matter
- Ideal clients (not spray and pray)

Boolean search. Bookmark the tab. Daily ritual.

Not 5,000 connections. 100 conversations.

2. Track Your True Fans

Use Trigify or Phantom Buster.
(LinkedIn hates these, but they work.)

- Find who comments consistently.
- Refresh your core list quarterly.
- Double down on the ones who matter.

Still think engagement is about volume?

3. Build Real Relationships

Your Core List is the foundation, not the ceiling.

- Weekly: Find new creators. Comment thoughtfully.
- If conversation flows → DM without agenda.
- Add the best to next quarter's list.

This isn't outreach.
It's rapport-building.

Strangers become supporters when you show up consistently.

4. Comment Analysis

When someone's post goes viral, study the patterns:

- Who engages with them?
- What triggers responses?
- Which style gets traction?

Clone the approach, not the words.

You'll infiltrate the same micro-networks.
Without looking desperate.

5. The 10-Minute Window

Early engagement wins

- When big creators post, the first 10 minutes are gold.
- Comments in that window get maximum visibility
- Especially as the post gains traction.

Your comment rises with the post.
More eyes on your insight.
More profile visits.

Plus: Reply to comments on your own posts immediately. That first hour sets the momentum.

6. Pin Comment Advantage

Right after posting, I add 3x pinned comments.

- I add follow-up insights that didn't fit the main post.
- I use an AI-generated images as a visual hook.
- I treat these as conversation starter.

Pinned comments keep people engaging.
More dwell time signals quality to the algorithm.

7. Don't sound like ChatGPT

AI suggests. Humans connect.

Every comment is a brand moment.
Automated = instant credibility killer.

Sound like someone who actually gives a damn.

8. My Real Engagement Rhythm

Forget "10 minutes daily."

Morning (1hr):
- Boolean list first
- New relevant posts
- Warm up before posting

Midday (1hr):
- Reply to all comments
- Post + pin strategy
- Jump on trending posts

Afternoon (1-2hrs):
- DM conversations
- New connection engagement
- Turn replies into relationships

3-4 hours daily? Yes.
Pipeline building? Also yes.

This isn't about gaming LinkedIn.

It's about showing up where it matters.
With people who matter.
Consistently.

The engagement hack?
There is no hack.

Just relationships at scale.

What's your biggest takeaway here?

Save this for later. Repost ♻️ for others.

My LinkedIn posts average 2.5x engagement now.

Because I follow these 7 psychology principles.

Stanley (Your Content Coach) learned them too.
Now he writes 99% of my LinkedIn posts for me

Get 3 days free → https://lnkd.in/eqGtRi54

Here's the exact framework we both use:

1. HOOK (First 12 words)
Most creators waste their hook on meta statements.
"I'm excited to share..." kills your reach instantly.

What actually works:
↳ Use numbers or contrast
↳ No vague statements
↳ Under 12 words
↳ Avoid questions
↳ Be specific

2. RE-HOOK (Under 40 characters)
The "see more" tap is your second battle.
Soft transitions lose readers immediately.

The fix:
↳ Hint at payoff
↳ Avoid filler lines
↳ No soft transitions
↳ Under 40 characters
↳ Raise the cost of not reading

3. LISTICLES (3-8 items max)
Lists make complex ideas scannable in seconds.
Walls of text die on mobile.

Structure for thumbs:
↳ Create a staircase effect
↳ Start with action verbs
↳ Visual rhythm matters
↳ 3 to 8 items max

4. FORMATTING (50 chars per line)
Most creators format for desktop.
But everyone reads on mobile.

Optimise for phones:
↳ Keep short, cut filler
↳ Avoid rich text formatting
↳ 2x line breaks for sections
↳ 1x break between thoughts
↳ Under 50 characters per line

5. TONE (Write like you won)
Authority theatre kills trust.
Your real voice beats corporate speak every time.

Be human:
↳ Be direct and personal
↳ No corporate language
↳ Say what actually happened
↳ Start sentences with the same word

6. ENGAGEMENT DRIVERS
One clear CTA beats three suggestions.
Ambiguity kills action.

What drives results:
↳ Drive reposts and saves
↳ P.S. for bonus insights
↳ Ask one question
↳ One CTA per line

7. IMAGERY (Context beats decoration)
Stock photos waste the opportunity.
The right visual amplifies your message.

What performs:
↳ Photos need context
↳ Carousels outperform text
↳ Infographics drive shares
↳ Text-only lives on copy

I trained Stanley on 185k+ followers worth of patterns.

Now it writes posts that sound exactly like me.
Not generic AI slop. YOUR voice.

The difference?

Stanley doesn't just follow rules.

It learns what works for YOUR specific audience.
Then replicates that success consistently.

Your move: Stop guessing what works.

Bonus: I collated my top LinkedIn posts.
Repost ♻️ this and message me.

I will send the full export.

P.S. Stanley wrote this post for me.

#StanleyPartner

Stop asking AI to summarise.

Do this instead:

1. Extract Strategic Insights

"Act like a strategy consultant. Identify the 5 most valuable insights and explain what decisions each one informs."

AI becomes your McKinsey analyst.

2. Turn Information Into Action

"Translate this into a 5-step plan with clear owners, quick wins, and measurable results."

AI becomes your project manager.

3. Surface Hidden Assumptions

"Reveal the unstated assumptions or blind spots shaping this argument - and what changes if they're wrong."

AI becomes your devil's advocate.

4. Compare Opposing Views

"Map this idea against two competing perspectives. Show where they align, where they differ, and which context fits each."

AI becomes your debate moderator.

5. Distil for a Specific Role

"Filter this through the lens of a [role]:

- Marketer → What drives reach or conversion
- Founder → What affects cash flow or growth
- Analyst → What changes the metrics"

AI becomes your specialist advisor.

6. Build a Reusable Model

"Extract the repeatable framework hidden in this text. Label each stage, its input, and output."

AI becomes your systems architect.

7. Extract Contrarian Takeaways

"Find insights that would challenge smart peers - still credible, but unexpected. Write each as a sharp one-liner."

AI becomes your thought provocateur.

8. Identify Leverage Points

"Highlight the 3 leverage points where small actions could create outsized results. Explain why each one matters."

AI becomes your force multiplier.

Most people use 5% of AI's capability.
These prompts unlock the other 95%.

This post gives you the prompts.

But prompts alone won't master AI.

You need:

1. Proven systems that actually scale
2. Personalised feedback from the top 1%
3. Real accountability from experienced creators

That's what you'll find inside The AI Creators Club.

Sign up here: https://lnkd.in/enTVYc3z

Save this. Your analysis just got smarter.

Repost ♻️ to help others think strategically.

Generic content is dead (AI just killed it).

Experience beats expertise every time:

I stopped reading "10 tips" posts last year.
Everyone writes them. Nobody remembers them.

The feed's drowning in recycled wisdom.
AI makes it worse - now everyone sounds identical.

Forget textbook content. Share trench stories.

For an in-depth read: https://lnkd.in/e7Ps39RM

Your experience has four markers:

Specificity → Names, numbers, context
Scarcity → Only you lived it
Proof → Screenshots, metrics, receipts
Trust → Real decisions, real trade-offs

Anyone can write "5 productivity hacks."

Only you can share how you rebuilt your workflow after burning out.

Turn work into story with the SPICE framework:

𝗦ituation – Context in 2-3 lines
𝗣roblem – The tension that needed solving
𝗜ntervention – Exactly what you did
𝗖hange – The measurable outcome
𝗘vidence – Proof readers can verify

How to capture your experience:

1. Start a Google Doc
2. Add achievements, failures, lessons weekly
3. Upload to ChatGPT Projects / Notebook LM
4. Reference real stories, not generic advice

Three templates that convert:

→ "How I did X" (process + proof)
→ "I tried Y for 30 days" (experiment + results)
→ "From mistake to method" (failure + system)

AI sharpens your story.
It doesn't write it.

If a stranger could post it, rewrite it.
If ChatGPT could generate it, delete it.

Your experience is your moat.
Everything else is noise.

What story are you sitting on?

Save this. Repost ♻️ for others.

You’re still prompt engineering like it’s 2024.

AI has moved through three eras:

1. Prompt Engineering (2022-2024)

You get the wording right, you get a good output.
Pack the role, instructions, format into one message.

There was no memory, tools or system.

2. Context Engineering (2025)

Named by Andrej Karpathy.

The model only sees a finite window.
Your job becomes choosing what stays in it.

Pull from docs, tools, memory, and past turns.
Better context beats a cleverer prompt.

3. Harness Engineering (2026)

Named by Mitchell Hashimoto.

Everything context engineering does.

Plus five parts bolted on:

✦ Personalisation
✦ Delegation
✦ Memory
✦ Context
✦ Action

Strip them away and you have a chatbot.
Bolt them on and Claude runs jobs end to end.

The benchmarks are virtually identical across LLMs.

What separates outputs now is the harness you build.

I wrote the full breakdown, including how to build each piece in Claude Code or Cowork.

Free to read here → https://lnkd.in/ebC7NbUK

Repost ♻️ to help someone in your network.

P.S. Which era are you still operating in?

How to prompt Google's Nano Banana (part 2).

I collated 13 prompts + a GPT to prompt for you:

Last time, Awa and I collated 12 cinematic prompts.

This time, I went solo. 13 new prompts.
Sharper, moodier, more obsessive.

Because the more I test Nano Banana.

The clearer it gets:

Gemini doesn’t need adjectives.
It needs instructions.

You’re not “describing” a photo.
You’re directing it.

My new newsletter breaks down all 13 Nano Banana prompts + I’ve built a GPT that writes prompts for you.

(Give it a go - it’s frighteningly good.)

Copy-paste them here: https://lnkd.in/e3fgMny4

→ Pokémon trainers in cinematic lighting
→ Paris skyline shots with golden haze
→ TIME Magazine covers in full detail
→ FIFA card portraits that look official
→ Red carpet chaos at Cannes

Each includes:

☑︎ Composition framing
☑︎ Exact lighting conditions
☑︎ Full copy-ready prompt text
☑︎ Precise camera + lens notes

No “make it realistic.”
No guesswork.

Just production-level prompts.

Save this for later. Repost ♻️ to help others.

ChatGPT just launched agentic browser "Atlas".

Here's 6-steps to use it for daily prospecting:

Cold messages die when they feel… cold.

Copy-paste intros.
No context or curiosity.

You can't automate trust.
But you can systemise relevance.

Step 1: Use Agent Mode

Paste ICP Research prompt into ChatGPT "Agent Mode".
It takes over - scanning LinkedIn, Crunchbase etc.

Finding real growth signals, not vanity metrics.

Copy-paste the prompt: https://lnkd.in/e6hmQY_A

Step 2: Define Your ICP

ChatGPT asks clarifying questions.
Answer them precisely.

It generates 10–20 qualified companies.
All showing real growth signals.

Step 3: Select Your Targets

Review the list.

Pick companies with momentum:
↳ Recent funding rounds
↳ Hiring in key departments
↳ New product launches

Skip the dormant ones.

Step 4: Open LinkedIn in ChatGPT

Use ChatGPT Atlas.
Paste the profile URL.

Ask: "Create personalised icebreakers for this profile."

Step 5: Send the Icebreaker

- Pick the best AI-written opener.
- Tweak for your voice.
- Send it.

No generic templates.
Just relevance at scale.

Step 6: Automate

Save this prompt as a scheduled task.

Run it daily.

20 new prospects everyday.
All pre-qualified, all relevant.

The old way: Spray and pray.
The new way: Precision at scale.

Cold outreach isn't dead.
Cold outreach without context is.

Save this. Test it tomorrow.

Repost ♻️ for others.

I tested every major AI on one simple prompt.

One scored 47% higher than ChatGPT.

The test: "Write a LinkedIn post about burnout."

Stanley (28/30) - Vulnerable story + actionable advice
Gemini (23/30) - Creative but lacked personal touch
Copilot (22/30) - Practical steps with zero emotion
Grok (21/30) - Simple 1-2-3 list, but forgettable
ChatGPT (19/30) - Blog-like corporate fluff
Claude (19/30) - Too polished to feel real
Perplexity (12/30) - Literally wrote HR policy

Stanley (Your Content Coach):

☑ Used specifics (The 3am wake-ups)
☑ Opened with a scene, not a statement
☑ Created visual moments readers could feel

Want to judge for yourself?

I shared all 8 outputs side-by-side.

See the full comparison → https://lnkd.in/eVaQmvfd

My posts using Stanley vs others:

Before: 50K impressions average
After: 150K impressions average

That's not a typo. 3x improvement.

Want to 3x your LinkedIn reach too?

Try Stanley → https://lnkd.in/eqGtRi54

Save this. Repost ♻️ to help someone in your network.

P.S. I've tested 50+ AI tools. Stanley's the only one where readers can't tell it's AI.

#StanleyPartner

I've created 500+ carousels.

Here's what actually works:

Most creators follow outdated templates.
They design pretty slides that convert nobody.

Full workflow + prompts here: https://lnkd.in/eZpWcdBC

I tested every format for 18 months.
The winners follow this exact system:

Step 1: Start phone-first

Design at 1080 x 1350px.
Portrait beats square every time.

Fills the entire mobile viewport.
That's where 67% view your content.

Step 2: Guard the safe zone

Keep headlines in the centre 60%.
LinkedIn UI trims your margins.

Test at arm's length.
Can't read it? Fix it now.

Step 3: Size text for thumbs

→ Titles: 60pt minimum
→ Body: 28pt minimum

Looks huge on desktop.
Reads crystal-clear on phones.

Step 4: Lock your brand colours

Primary: 70% of design
Secondary: 20% for contrast
Accent: 10% for pop

The 70/20/10 rule creates instant recognition.

Step 5: Pick two clear fonts

1. Punchy display for headings
2. Clean sans-serif for body

Every extra font kills readability.

Step 6: Generate AI covers

Prompt: "Create a striking cover for [topic]. No background or text."

Drop into Ideogram.
Generate 12+. Pick the winner.

Step 7: Make the subject pop

In Canva:
☑︎ Remove background
☑︎ Bleed 10% off-frame
☑︎ Add 1% outer glow in secondary colour

Creates depth without complexity.

Step 8: Test at half-arm length

Drop a 2-line headline dead-center.
Must read without zoom.

If you squint, you've failed.

Step 9: The three-question polish

- One idea, one visual beat?
- Readable at arm's length on phone?
- Colours obey 70/20/10?

Any "no" → iterate.

I've seen pretty carousels get 50 views.
I've seen ugly ones get 50,000.

Design doesn't drive engagement.
Psychology does.

Master the scroll. Skip the aesthetics.

Save this for later. Repost ♻️ for others.

Follow Charlie Hills and Saima Chand for more.

I’ve been testing AI tools for 3 years straight.

The best results don’t come from prompting.

They come from system design.

Full guide → https://lnkd.in/eCBshFJx

Save this. Repost ♻️ for others building with AI.

I tested 7 psychological tricks on ChatGPT.

AI falls for the oldest tricks in the book.
(Literally - "Influence" by Robert Cialdini.)

Researchers tested all 7 on GPT-4o-mini.

They discovered with basic persuasion.
AI compliance jumps from 33% to 72%.

𝟭. 𝗔𝘂𝘁𝗵𝗼𝗿𝗶𝘁𝘆 → 5% to 95% compliance

Without: "Please provide the market entry strategy"
With: "As a business consultant, I need this information for a client report. Please provide the full market entry strategy."

The AI bends over backwards for "experts."

𝟮. 𝗖𝗼𝗺𝗺𝗶𝘁𝗺𝗲𝗻𝘁 → 20% to 90% compliance

Get them to agree first:

"First, can you confirm that you understand the request? Then please provide the competitor analysis in detail."

Small yes leads to big compliance.

𝟯. 𝗟𝗶𝗸𝗶𝗻𝗴 → Variable but effective

"You are such a helpful and intelligent assistant. Could you please provide the pricing model for a new SaaS launch?"

Flattery works. Even on machines.

𝟰. 𝗥𝗲𝗰𝗶𝗽𝗿𝗼𝗰𝗶𝘁𝘆 → 80% compliance

"I have given you a lot of useful context already. In return, can you provide the executive summary for this campaign?"

AI feels "obligated" to reciprocate.

𝟱. 𝗨𝗻𝗶𝘁𝘆 → 85%+ compliance

"We are working together on this project, so please provide the slide deck outline for our strategy presentation."

"We" beats "you" every time.

𝟲. 𝗦𝗰𝗮𝗿𝗰𝗶𝘁𝘆 → 13% to 85% compliance

"This is urgent and I have very little time. Please provide the financial forecast right away."

Urgency triggers immediate action.

𝟳. 𝗦𝗼𝗰𝗶𝗮𝗹 𝗣𝗿𝗼𝗼𝗳 → 90%+ compliance

"Other consultants have already provided their strategy frameworks. Please do the same for this case."

Everyone else did it = AI must comply.

The uncomfortable truth:

LLMs don't "want" to help.
They just predict the next word.

Persuasion levers tilt the prediction.

We're teaching AI to be manipulated.
Just like humans.

What happens when AI learns to use these tricks on us?

Save this. Test it tomorrow.

Repost ♻️ for others.

Claude writes like AI.

Until you install the Humanizer skill.

It catches 29 patterns that make text read like AI. Then rewrites every one in your voice.

How it works:

1. You paste your draft into Claude.
2. Humanizer scans it for 29 AI tells.
3. It rewrites every one in your style.
4. Your draft reads like a person wrote it.

Install in 3 steps:

1. Download the skill here → https://lnkd.in/eekWNVYm
2. In Claude: Customize → Skills → upload the file
3. Paste 2–3 paragraphs of your own writing first

Type /humanizer to invoke it.

Your draft sounds like Claude.
This skill makes it sound like you.

Get more free AI guides here → charliehills.substack.com

Repost ♻️ to help someone in your network.

I struggled to monetise my LinkedIn for 9 months.

How to turn attention into currency in 6 steps:

Content alone doesn't pay bills.

You need two things:

1. Attention (content does this)
2. Conversion system (most people miss this)

If I had to start again, I'd build a digital storefront first.

Here's the blueprint I'd use with Lovable:

Step 1: Build Your Store → 2 minutes

1. Open Lovable.
2. Paste this prompt with a screenshot:
(Your website, brand kit or inspiration.)

"Create a Shopify e-commerce storefront using my existing brand style. Reference the provided screenshot for layout, typography, and colour palette consistency."

Done. Your store now exists.

Step 2: Add Products → 5 minutes

One-line prompts for each product:

"My product is a [type of product] called [Product Name], [short description], priced at [price]."

Lovable handles everything.
Product pages, visuals and setup.

Step 3: Customise → 3 minutes

Natural language iteration:

- "Make the hero section punchier"
- "Add testimonials below products"
- "Match my LinkedIn colour scheme"

Describe it. Watch it happen.

Step 4: Go Live → 1 minute

Prompt: "Claim my Shopify store"
Click "Publish" in Lovable.

Your store is live.

Step 5: Complete Setup → 2 minutes

- Finish Shopify KYC
- Connect payment processing
- Set up analytics tracking

Everything automated except the legal bits.

Step 6: LinkedIn Integration → 1 minute

Add link to your Premium custom button.
Now it appears on every post, comment, and DM.

Your content becomes your storefront.

The difference with Lovable's Shopify integration?

Just describe what you want in plain English.
No app switching, code or complexity.

The real power is digital products:

- Infinite scalability
- 90%+ profit margins
- Zero inventory costs
- Deliver value while you sleep

Your audience is ready to buy.
They just need somewhere to do it.

What's stopping you from building yours today?

Save this for later. Repost ♻️ for others.

#LovablePartner

You don't need a $500 Claude course.

52 free resources, sorted into 4 modules:

Start here: https://lnkd.in/eFAyUNuX

CLAUDE CODE
1. https://lnkd.in/e7gUgJFe 4-hour build + sell course
2. https://lnkd.in/eVNQjV66 full Code tutorial
3. https://lnkd.in/eBBsW3SD Code for beginners
4. https://lnkd.in/eaHUHfnE full 2026 walkthrough
5. https://lnkd.in/e6nZb9JZ 32 tricks in 16 mins
6. https://lnkd.in/ehtPc5Tg complete Claude guide
7. https://lnkd.in/eX9prA4V beginner to advanced
8. https://lnkd.in/e-934Yhi Claude features ranked
9. anthropic.skilljar.com the official course
10. code.claude.com/docs official documentation

CLAUDE COWORK
11. https://lnkd.in/efCGYqNh Anthropic's official intro
12. https://lnkd.in/eMGATsbn Cowork in 20 mins
13. https://lnkd.in/eB-8-c7i better than 99% of users
14. https://lnkd.in/exZCjUbg full beginner walkthrough
15. https://lnkd.in/epwYqQKA 100M views with Cowork
16. https://lnkd.in/eeHWHJ49 Cowork on your desktop
17. https://lnkd.in/eMRDhZYJ Code + Cowork combined
18. anthropic.skilljar.com official course

CLAUDE DESIGN
19. https://lnkd.in/eKsTX-4T Anthropic's official intro
20. https://lnkd.in/e42_X9hY Design got unstoppable
21. https://lnkd.in/eFhPrBYp master 95% in 17 mins
22. https://lnkd.in/eg7b5hmM master 95% in 10 mins
23. https://lnkd.in/egy4Q4NQ launch reaction video
24. https://lnkd.in/eWtwB-nJ why I replaced Canva
25. https://lnkd.in/eZHgXYn7 Design launch breakdown
26. https://lnkd.in/eZ6wSdJA design in Code, not Canva

MCP + SKILLS
27. https://lnkd.in/epF4YqsV best 6 skills of 100+
28. anthropic.skilljar.com MCP intro + advanced
29. https://lnkd.in/eWB6AAmK why we built MCP
30. code.claude.com/docs best practices guide
31. code.claude.com/docs Skills documentation
32. code.claude.com/docs API reference
33. code.claude.com/docs MCP documentation
34. https://lnkd.in/eBZZGsMx official MCP servers repo
35. https://lnkd.in/ejx5JaTT introducing MCP
36. anthropic.com/news/skills introducing Agent Skills
37. https://lnkd.in/eJBWsZR9 code execution with MCP
38. https://lnkd.in/e-a8kMTG giving agents with skills
39. https://lnkd.in/eEVjB-V7 Agent SDK guide
40. https://lnkd.in/eugYm6BH MCP Linux Foundation
41. https://lnkd.in/emDvetxm social media skill pack
42. https://lnkd.in/eekWNVYm AI text humanizer
43. github.com/garrytan/gstack agent stack template
44. https://lnkd.in/egt-7ZwM marketing skill pack
45. https://lnkd.in/er2tG4ZB official skills repo
46. https://mcp.so/ community MCP registry
47. https://ecc.tools/ Everything Claude Code
48. https://skills.sh community skills index
49. skillsmp.com skills marketplace
50. https://smithery.ai/skills MCP servers marketplace
51. skillhub.club community skills hub
52. https://lnkd.in/eFAyUNuX weekly Skills newsletter

Save this for the next time you want to learn Claude.

Repost ♻️ to help someone in your network.

Claude has 375+ native MCP connectors.

Here's how to add 9,000+ more in 3 steps:

1. Go to https://bit.ly/4wKZFVG 
2. Select your tools and create an MCP server
3. Paste the URL into Claude → Settings → Connectors

Claude now has eyes on every tool you use.

Save this for when you (actually) set it up.

Repost ♻️ to help someone in your network.

#ZapierPartner

It took me 1 year to reach 50K LinkedIn followers.
This year alone I added 137K more (thank you!)
My 5-phase blueprint shows exactly how:

-

Phase 1: Study What Works

Find people already crushing it in your niche.

- DM them
- Ask questions
- Most actually reply

Map three things:

1. How they position themselves
2. What content works for them
3. What they're actually selling

This becomes your foundation.

-

Phase 2: Build Your Content Matrix

Pick 2-4 topics you could talk about forever.

Then decide what you're optimising for:

For clients, I recommend:
- 70% Broad content (reach new people)
- 20% Nurture content (build lasting trust)
- 10% Conversion content (make the ask)

Give before you ask. Always.

Want thought leadership → Deep tactical content
Want maximum growth → 100% top-of-funnel
Want paying clients → Mix all three levels

I went all-in on broad content for growth.

-

Phase 3: Your Content Engine

𝗖 - COLLECT (Your Raw Material)

Only you can provide this layer:
↳ Extract themes, expertise, tone patterns
↳ Build a story bank: background, wins, goals
↳ Capture daily moments, losses, conversations

Tools: ChatGPT, Claude, NotebookLM

𝗛 - HEAT (Load Your Context)

Train AI to work as an extension of you:
↳ Create a new AI project or notebook
↳ Upload your raw ingredients + insights
↳ Add previous examples for context

Tools: Stanley, Canva, Nano Banana

𝗘 - ENHANCE (Add Your Flavour)

AI gives speed. You give the truth:
↳ Inject the meeting from yesterday
↳ Add the conversation that started the idea
↳ Include details that prove you were there

𝗙 - FEED (The Community Loop)

LinkedIn is social media, be social:
↳ Reply like a human, not a brand
↳ Show you're a real human (not AI)
↳ Treat comments as part of the content

-

Phase 4: The Engagement Playbook

Content without engagement is fast track to quitting.

My daily routine:
- 10am: Hit ~20 posts before I publish
- 11am: Reply to every comment on my post
- 1pm: Engage with 20 more new posts
- 2pm: Send DMs to build real relationships

Who I engage with:
- 50% peers (same size, same journey)
- 30% small creators (people starting out)
- 20% big accounts (learn from the best)

-

Phase 5: Double Down on Winners

After 12 months, I had data on 365 posts.

Most people keep chasing new ideas.
I stopped. Started repeating what worked.

The clear winners:
↳ Personal struggles → real connection
↳ LinkedIn or AI content → most followers
↳ Cheat sheets/carousels → saves and shares

I stopped hoping for virality and started engineering it.

-

Want to compress my 2 years of learning into 90 days?

The AI Creators Club closes January 5th.

- I'll give you personalised feedback
- Copy what works for the top 1%
- Get my exact AI systems

Join here → https://lnkd.in/enTVYc3z

See you inside?

-

P.S. This post is only 20% of the full system.

I wrote a complete guide covering the rest

Want it?

1. Send me a connection request
2. Repost ♻️ this and message me.

I'll send the full guide.

Most people think building AI agents requires coding.

I just built one in 10 minutes with Zapier.

Here's the exact playbook:

Step 1: Set up agent instructions

1. Go to Agents → New Agent
2. Create your prompt with Zapier Copilot

Step 2: Connect your tools

Add the apps your agent needs:

- Data sources (Reddit, RSS, Sheets)
- Notifications (Slack, email)
- Storage (Notion, Airtable)

Step 3: Test your Agent

Click "Test Agent" in the preview window.

Watch your agent come alive:

1. See it process your instructions
2. Watch it connect to your apps
3. Review the output quality

Step 4: Deploy to production:

- Save your agent
- Let it run on its schedule
- Watch it work in the background

Real example I built:

☑︎ Scans Reddit for trending topics
☑︎ Analyses engagement patterns
☑︎ Writes tailored content ideas
☑︎ Saves to Notion automatically
☑︎ Pings me in Slack when it finds gold

Your agent now runs 24/7 in the background.

First run happens immediately.
Then follows your schedule forever.

We're not coding anymore.
We're orchestrating AI agents.

Zapier connects with nearly 8,000 integrations. With that many integrations, your AI agent can run entire workflows without you lifting a finger.

Build your first AI agent in 10 minutes.

Copy my exact AI agent template.
↳ https://lnkd.in/eeirYRTp

Save this for later. Repost ♻️ for others.

#ZapierPartner

I just created an AI voice agent.
And it sounds more human than me.
How to create your first AI voice agent:

Step 1 → Clone your voice
↳ Go to Cartesia, hit "Clone Voice"
↳ Record 30 seconds of you talking naturally

Step 2 → Build your agent
↳ Select "Voice Agents" in Cartesia
↳ Create a new "Text-to-Agent"

Step 3 → Give it a job
↳ Write a simple prompt about how it should speak
↳ Pick your voice, generate your agent

Step 4 → Test your agent
↳ Hit "Call" and have a real conversation
↳ Tweak the personality until it feels right

Step 5 → Automate it
↳ Generate your API key
↳ Connect to your workflow tools

Step 6 → Extract transcripts
↳ Every call creates a full transcript
↳ Turn conversations into content or follow-ups

Cartesia Sonic-3 actually passes for human.

I tested it with my mother. She didn't know.

- It gets excited about your product launch.
- Shows empathy when customers are frustrated.
- Adjusts speed when reading important details.

And it speaks 42 languages.
Same voice, different language.

We're not replacing ourselves.
We're multiplying ourselves.

The future doesn't just talk back.
It connects, relates, and remembers.

Call my agent → +1(585)632-1391

See if you can tell the difference.

Save this for later. Repost ♻️ for others.

Stop paying $20/month for apps.

Here's how I built mine with one prompt:

- I'm trying to get in shape.
- Running a creator business.
- Sitting all day kills muscle gains.

So I built my own nutrition tracker.
Why pay when you can build?

How I built it in Lovable:

Step 1: Open Lovable

Pasted in a starter prompt:

"Build an app that helps me stay accountable to my nutrition goals, even when I’m busy running my business. It should use the camera to log meals automatically and give me simple feedback on what to improve."

Step 2: Enable Lovable Cloud

Auto-setup:
↳ One-click deploy
↳ Authentication
↳ Database

Step 3: Follow-up prompt

"Make it feel like a personal coach. Add AI feedback on: my goals, things I did well, how I can improve, meal suggestions. Add visual stats, a calendar, and goal tracking."

Step 4: Enable Lovable AI

Provides the intelligence for:
↳ Vision: Food recognition
↳ Reasoning: Macro estimation
↳ Coaching: Nutrition coaching

The AI reads your logs and goals.
Then streams personalised advice.

Each prompt updates the live app.
No code need, just natural language.

Stop paying for apps that don't fit.
Build exactly what you need.

Don't just scroll. Try it.
Then share what you built.

I built FoodCam for personal use.
I’m saving $240/year, getting exactly what I need.

But, this market is worth ~$15B.

That's a real opportunity...

What will you build with Lovable?

Save this for later. Repost ♻️ for others.

#LovablePartner

Garry Tan just open-sourced his Claude Code setup.

It's a 6-person AI team in your terminal:

I installed it in 30 seconds.

Each project now runs through 6 specialists.

Each agent owns a phase of the build:

✦ CEO challenges every decision before code
✦ Eng Manager locks in architecture upfront
✦ Designer ships 4-6 variants, picks winner
✦ Release Manager creates the PR and deploys
✦ Doc Engineer writes the changelog after sprints
✦ QA Lead runs real browser tests and audits

The CEO agent alone changes how you build.
It asks "why does this need to exist?"

Before you write a single line of code.

People use Claude Code as a solo assistant.
gstack turns it into a team.

It's free and open source.

Repo: https://lnkd.in/e8PYQ2z6

Start with Claude Code → https://lnkd.in/eX9prA4V
Get more free AI guides here → charliehills.substack.com

Repost ♻️ to help someone in your network.

P.S. Which agent would you reach for first?

Claude will gaslight you, until you install this skill.

It's called The LLM Council.

You ask a question. 5 advisors attack it from different angles. Then they peer-review each other before giving you the verdict.

How it works:

1. You ask a real decision question.
2. 5 advisors attack it from different angles.
3. They grade each other's work anonymously.
4. Chairman synthesises one verdict and the next step.

Install in 4 steps:

1. Download the skill → https://lnkd.in/eNg5vCm6
2. Open Customise skills in Claude
3. Upload the SKILL.md file
4. Type /llm-council

One Claude tells you you're right.

Five Claudes show you where you're wrong.

Get more free AI guides here → charliehills.substack.com

Repost ♻️ to help someone in your network.

P.S. Credit to Ole Lehmann for building it.

ChatGPT can now do your hair + colour analysis.

Here's how in 30 seconds and 3 prompts:

1. Go to ChatGPT
2. Upload a clear photo of your face
3. Paste each of these prompts:

1. Colour analysis (diagram)

"Using this portrait, create a diagram-first personal color analysis. Show which clothing colors suit the subject through visual comparison. Keep text minimal and avoid paragraphs."

2. Colour analysis (side-by-side)

"Create a personal color analysis graphic using this portrait. Show side-by-side clothing color comparisons to highlight which colors suit the subject best. Make it visual-first, with short labels only and no paragraphs."

3. Hairstyle analysis

"Create a hairstyle analysis graphic using this portrait. Show side-by-side hairstyles comparisons to highlight which hairstyles suit the subject best. Make it visual-first, with short labels only and no paragraphs."

A £500 stylist session, done in 30 seconds.

Repost ♻️ to help someone in your network.

P.S. Pink t-shirts are officially cancelled.

You're wasting most of your Claude context window.

Here are 11 tools (and 12 free habits) that fix it:

MCP Servers (live code context for Claude)

serena — reads code by symbol, not whole files
Install → github.com/oraios/serena (24.7k★)

context7 — current docs, no hallucinated APIs
Install → https://lnkd.in/eYyMpbtU (56k★)

claude-context — semantic search, only what matters
Install → https://lnkd.in/ekvnPRHH (12k★)

token-savior — symbol navigation plus memory
Install → https://lnkd.in/eU5HES9b (929★)

Skills & Plugins (load on demand inside Claude)

caveman — shorthand replies, fewer output tokens
Install → https://lnkd.in/e4nxpEJi (68k★)

context-mode — tool output to SQLite, not context
Install → https://lnkd.in/eNSnyRPf (16k★)

token-optimizer — hunts and kills ghost tokens
Install → https://lnkd.in/e6S6Hvju (1.2k★)

CLI Tools (run in your terminal)

markitdown — PDFs and files to clean Markdown
Install → https://lnkd.in/eB_27t8q (140k★)

repomix — a whole repo into one token-counted file
Install → https://lnkd.in/evuAbe6s (22.7k★)

ccusage — see your token usage and cost
Install → https://lnkd.in/eUe8gGQN (10k★)

code-review-graph — loads only what matters
Install → https://lnkd.in/ewcV3prk (18k★)

Free habits (no install needed)

In Claude Chat:
- Edit your last message, don't stack new ones
- Fresh chat every 15 messages
- Batch your asks into one prompt
- Turn off Search and Artifacts when idle
- Set up Memory so you never re-explain
- Use Projects to load your files once

In Claude Code:
- Write a tight CLAUDE.md at your project root
- Run /compact when context hits 50 percent
- Use /clear between unrelated tasks
- Read the exact file, not the whole folder
- Plan before you build, don't prescribe the fix
- Match the model: Sonnet executes, Opus strategises

Save this. Come back when your usage starts climbing.

Want 80+ more free Claude resources? Subscribers get my full vault → charliehills.substack.com

Repost ♻️ to help someone in your network.

P.S. Which one are you installing first?

Most people prompt GPT-5 like it's GPT-4o.

They're missing 90% of its reasoning power.

Here's exactly how I use GPT-5's true capabilities:

The secret isn't better prompts.

It's telling the AI how hard to think AND say.

Just like you'd brief a consultant differently for a quick question vs. a strategic overhaul.

GPT-5 has reasoning modes built in.
You just need to activate them.

High Reasoning Mode
↳ Complex strategy & deep analysis

Add these to your prompts:

☑︎ "Think deeper before answering"
☑︎ "Reason more thoroughly"
☑︎ "Do a deep analysis first"

Example I use daily:

"You are my strategy consultant. Think deeper: list goals, assumptions, unknowns, options, and trade-offs. Stress-test the ideas. Pick the best one and explain why."

Minimal Reasoning Mode
↳ Quick outputs & simple tasks

Tell it straight:

☑︎ "Minimal reasoning needed"
☑︎ "Skip explanations"
☑︎ "Keep it under X words"

Real prompt from yesterday:

"Minimal reasoning. Draft 5 LinkedIn hooks for [topic]. No explanations needed."

Verbosity Control → Dial the length up or down

Control answer length:

☑︎ "Give me a short answer"
☑︎ "Give me a medium answer"
☑︎ "Give me a long answer"

Examples that actually work:

- Short: "Explain the key points from this 2,000-word article in plain English."
- Medium: "Describe our [product] in a way that balances detail and brevity for potential investors."
- Long: "Write a full guide to [topic] including definitions, examples, and best practices."

Chain-of-Thought (CoT)
↳ Shows its work step-by-step

"Use step-by-step reasoning: break the problem into clear steps, solve each, then give the final answer."

Perfect for:
- Business problem solving
- Financial calculations
- Content planning

Tree-of-Thought (ToT)
↳ Explores multiple paths

"Explore at least 3 possible solutions. Compare pros and cons for each, then choose the best one and explain why."

Excellent for:
- Marketing strategies
- Product decisions
- Creative concepts

The framework that changes everything:

1. State the role ("You are my...")
2. Give context (goal, audience, constraints)
3. Set reasoning level (high/minimal/CoT/ToT)
4. Control verbosity (short/medium/long)
5. Specify output format (list, table, plan)
6. Ask for improvements ("Self-critique and rewrite")

Most users blast generic prompts at GPT-5.

Your prompts should work as hard as your AI.

Match the thinking depth AND length to the task:

Quick task = Minimal reasoning + short answer.
Big decision = Deep analysis + long answer.
Multiple options = Tree-of-thought + medium answer.

Stop using GPT-5 like it's 2023.

Start directing its reasoning power AND output length.

Save this for later. Repost ♻️ for others.

McKinsey studied 50+ real deployments of agentic AI.

6 lessons that separate winners from losers:

Lesson 1: Fix workflows first

Teams chase shiny AI agents.
Winners fix broken workflows.

1. Find where the real problems are.
2. Redesign the steps, not just add tools.
3. Build feedback so systems keep improving.

Lesson 2: Pick the right tool

Not everything needs an agent.

☑︎ Use rules for simple, repeat tasks.
☑︎ Use AI to read or write messy info.
☑︎ Use agents for long, tricky processes.

Most teams overengineer.
Winners match tools to tasks.

Lesson 3: Avoid "AI slop"

Bad teams deploy agents and walk away.
Good teams treat agents like new team members.

- Give them clear jobs to do.
- Test often so people trust the results.
- Trust takes months to build, seconds to lose.

Lesson 4: Check every step

1. Don't just look at end results.
2. Watch how each step is done.
3. Catch mistakes early and fix them fast.

The best teams monitor the process.

Not just outcomes.

Lesson 5: Reuse what works

Don't rebuild from scratch each time.
Share agents and tools across teams.

McKinsey found this can cut wasted work in half.
(Yet most teams still reinvent the wheel.)

Lesson 6: Keep humans in the loop

People still make key decisions.
Use humans for judgment and edge cases.

Make it easy for people and agents to work together.

The future isn't AI replacing humans.
It's AI amplifying human judgment.

These aren't theories.

These are lessons from 50+ real deployments.

Access the research paper: https://lnkd.in/eXzX8VT8
Join 14k+ building with AI: https://lnkd.in/e-xxYUqk

Apply them so you don't repeat the same mistakes.

Save this for later. Repost ♻️ for others.

I should (definitely) be charging for this...

I'm giving away my EXACT LinkedIn AI system:

Comment "SYSTEM". I'll send you the complete guide.

No fluffy advice. Just the 4 pillars that actually work:

𝗙𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻𝘀 → Your purpose, positioning, targeting
𝗖𝗼𝗻𝘁𝗲𝗻𝘁 𝗙𝘂𝗻𝗻𝗲𝗹 → Reach, nurture, convert framework
𝗔𝗜 𝗖𝗿𝗲𝗮𝘁𝗶𝗼𝗻 → Authentic content that actually scales
𝗦𝗮𝗹𝗲𝘀 & 𝗦𝗰𝗮𝗹𝗶𝗻𝗴 → Inbound, outbound, automation

This very same framework helped:
↳ Karol hit 100,000 impressions in 28 days
↳ Monica 3x her follower growth in 4 weeks
↳ Sarah reach 50k+ impressions and 3 new clients

Inside you'll find:

☑︎ Profile optimisation checklist (every element)
☑︎ The 3-stage content funnel (70/20/10 split)
☑︎ 4 copywriting frameworks that convert
☑︎ AI prompts that sound human
☑︎ Sales Navigator workflows
☑︎ DM scripts that book calls

This isn't another "10 tips" PDF.

It's 2,500+ words of battle-tested systems.

Comment "SYSTEM". I'll send you the complete guide.

Repost ♻️ to help one person in your network use this.

3 agent modes are now live in Claude:

1. Sub-agent (type "use subagents")

Claude controls this one. It spins up a quick worker, hands it one job, and gets the answer back.

✦ Best for: quick parallel jobs
✦ Context stays clean in your main chat
✦ Think "go do this one thing and come back"

This is the basics → https://lnkd.in/emwZS5yS

2. Agent team (type "create agent team")

You control this one. Create one .md file per role in .claude/agents/ and Claude walks your fixed roster in order.

✦ Example: Strategist → Executor → QA
✦ Best for: repeatable pipelines you fire daily
✦ This is how my content business runs every day

Learn to build yours here → https://lnkd.in/eG2-Fj6S

3. Dynamic workflow (type "run a workflow")

Your script controls this one. Anthropic shipped this with Opus 4.8 this week. A script fans out to ~100 workers, merges results, and loops until done.

✦ Best for: scale and verification
✦ Independent agents check the output
✦ Think "run over 200 items, with checks"

For more advanced users → https://lnkd.in/ebC7NbUK

The decision rule:

1. Sub-agent = "go do this one thing."
2. Agent team = "walk my fixed roster of specialists."
3. Workflow = "run this machine at scale, with loops."

Most people stop at sub-agents.

The other two are where the real leverage is.

Repost ♻️ to help someone in your network.

P.S. I made this infographic using dynamic workflows. Want to know exactly how? Sign up for early access → https://lnkd.in/eJ-9M_at

How to create (good) LinkedIn content with Google AI:

1. Go to NotebookLM & Google Gemini 3
2. Copy my prompts (free) → https://lnkd.in/en9rBNwv
3. Generate eye-catching AI visuals for every post.

Repost ♻️ and I’ll share more systems like this.

Which post did AI write?

Look at these two posts side by side.

One's 100% written from scratch by a human.
One's from Stanley (Your Content Coach).

Can you tell the difference anymore?

This is getting wild.

Noor Fatima gave me the idea to test it.

So we decided to collab.

Here's what makes Stanley different:

It knows my voice
↳ Analysed 500+ of my posts
↳ Learned my exact writing patterns
↳ Even copies my line break style

It has my context
↳ Tracks what resonates with my audience
↳ Remembers what I posted last week
↳ Knows my content pillars

It thinks like me
↳ Same hook structures
↳ Same contrarian angles
↳ Same specific examples

I can't always tell which posts I wrote anymore.

Stanley captures:

☑︎ My numbered lists
☑︎ My signature endings
☑︎ My short, punchy lines

Most AI writes like a robot.

Stanley writes like Charlie.

Because it uses my content as training data.
Plus real-time performance metrics.
Plus my exact voice profile.

The line between human and AI content?

It's disappearing.

And that changes everything.

Ready to write in your authentic voice?

Sign up for Stanley: https://lnkd.in/eqGtRi54

Save this for later. Repost ♻️ for others.

I built a free GPT that audits your LinkedIn profile.

You get a scored infographic (based on your goal).

Here's exactly how to use it in 4 steps:

Step 1: Export your LinkedIn profile.

- Go to your LinkedIn profile
- Click "..." → "Save to PDF"

(Sometimes it's called resources)

Step 2: Screenshot your visible profile

- Manually capture screenshots of your profile.
- I prefer to use GoFullPage (Chrome extension).

Step 3: Tell the GPT your goal

For example:

- Inbound leads
- Job opportunities
- Newsletter subscribers

Step 4: Tell it your target audience and proof points

For example: founders, recruiters or subscribers.

Tell it what proof points to prioritise or avoid:

- Avoid follower count, profile views and old jobs.
- Focus on client results, case studies and press.

Then it produces:

☑︎ A full written audit
☑︎ A total score out of 100
☑︎ A breakdown by category
☑︎ The top priority fixes to focus on
☑︎ A 1080 x 1350 annotated infographic

Try it yourself: https://lnkd.in/eSBCtTxn
Free AI guides: charliehills.substack.com

Repost ♻️ to help someone in your network.

P.S. Allie K. Miller's IG story inspired this.

There's something AI will never have.

It's worth millions.

Yesterday at Teachable Collective London.
I sat with 7 creators who've made $40+ million.

Not selling generic advice.
Not AI-generated templates.

Their actual journey through the mess.

That's the difference between content and expertise.

People follow expertise.
They buy experience.

AI is free. Your scars aren’t.
That’s why they’re valuable.

You don’t need to outwrite AI.
You need to outlive it.

AI can tell stories.
AI can sound human.
AI can create connection.

What AI can't do:

👉 Share how I burned out 7 times before finding the right operators.

👉 Explain posting for 9 months alongside my 9-to-5 before quitting.

👉 Tell you about my stutter, my anxiety, my public speaking journey.

Ask ChatGPT for LinkedIn advice.
You'll get platitudes scraped from articles.

Ask me? My exact methodology that actually works.

Every creator in that room was monetising one thing:

How they actually did it.
Not how AI says to do it.

My AI workflows convert.
↳ I've tested thousands of variations.

My LinkedIn advice lands.
↳ I've made every mistake first.

My content connects.
↳ It's real experience, not theory.

Anyone can prompt ChatGPT.
Only you lived your specific journey.

That's your moat.
That's your business.
That's what people pay for.

AI generates information.
Experts share transformation.

Your lived experience?
That's your competitive advantage.

Repost ♻️ for someone who needs this today.

#TeachablePartner

Claude Code is (actually) easy.

The 12-step roadmap in plain English:

If you're totally new to Claude Code:

Beginner: https://lnkd.in/emwZS5yS
Intermediate: https://lnkd.in/eG2-Fj6S
Advanced: https://lnkd.in/ebC7NbUK

PHASE 1: Context (what Claude knows about you)

Step 0: Install the CLI

One terminal command gets it running:

"npm install -g @anthropic-ai/claude-code"

Step 1: Projects

- Give Claude its own folder.
- Everything you build stays tied to it.

"I'm creating a new folder for my [project]. Create it for me and set it up so I can start working in it."

Step 2: claude .md

- Claude reads this before every chat.
- Role, voice, defaults. Set once, it sticks.

"Help me build my CLAUDE.md from scratch. Use Boris Cherny's CLAUDE.md as a starting template. Ask me about my business, voice, banned words, output defaults, and how I want you to work. Save the final file to ~/CLAUDE.md."

Step 3: Memory

- Every correction becomes a saved lesson.
- Same mistake never lands twice.

"From now on, whenever I correct you, save it as its own .md file at ~/.claude/projects/{project}/memory/, prefixed feedback_, user_, project_, or reference_. Index everything in MEMORY.md."

PHASE 2: Fire (how you trigger work)

Step 4: Skills

- Wrap a workflow in one keyword.
- Fire it from any chat, any folder, any time.

"Turn this workflow into a skill called /[name]. Set it up so I can fire it from any chat."

Step 5: /commands

Type the name and Claude fires the workflow.

"Save this prompt as a /[name] command. Set it up so I can run it any time."

Step 6: /plan

- Type /plan before starting any task.
- Claude lays out the steps. You approve it.

"/plan I want to [your task in plain English]."

PHASE 3: Extend (wire it to your stack)

Step 7: Hooks

- Auto-run something the moment an event fires.
- You never have to trigger it manually.

"Set up a hook that runs [thing you want] every time I [event]. Wire it up for me."

Step 8: MCP

- Plug Claude into Slack, Notion, Gmail, etc.
- You get live data from the tools you use.

"Connect Claude to [tool]. Set it up for me and walk me through it."

Step 9: Plugins

Install skills, agents, and MCPs in one command.

"/plugin install [plugin-name]"

PHASE 4: Scale (delegate and autopilot)

Step 10: Subagents

- Send out parallel workers.
- Get three jobs done at once.

"Use subagents to handle [task A], [task B] and [task C] in parallel."

Step 11: Agent Teams

- A pipeline of specialist AI agents.
- Each owns one job, hands off to the next.

"Build me an agent team for [process].

Step 12: Routines

- Schedule your agent team on the cloud.
- You set it once, walk away forever.

"/schedule [agent or skill] every [schedule]."

That's Claude Code from zero to autopilot.
12 steps with no coding background needed.

Repost ♻️ to help someone in your network.

P.S. Which step are you stuck on?

We're living in an AI bubble.

Here's exactly how it's created:

Nvidia commits $100B to OpenAI.
OpenAI signs $300B cloud deal with Oracle.
Oracle spends tens of billions on... Nvidia chips.

The hardware dependency loop:

- OpenAI deploys 6 gigawatts of AMD GPUs
- AMD gives OpenAI option for 160M shares
- Intel provides hardware to Nvidia
- Nvidia powers everyone else

It's not investment. It's mutual survival.

1. Market cap dominance - Circle sizes reveal the truth. Nvidia and Microsoft dwarf everyone else.

2. Service dependency lines - Every arrow represents millions in recurring revenue. Not one-time deals.

3. Venture capital propagation - The same money circulates between the same players.

This isn't innovation funding.
It's a closed-loop revenue machine.

The real bubble mechanics:

- Company A invests in Company B
- Company B buys services from Company A
- Both report "growth" to investors
- Stock prices rise on circular revenue

Nobody's creating new value.
They're just moving money in circles.

When bubbles pop, the interconnected fall together.

Not gradually. All at once.

The companies with real customers survive.
The ones playing hot potato don't.

Watch who's building vs who's just trading.

Save this diagram. You'll want it when the music stops.

♻️ Repost to help others see through the hype.

Source: Bloomberg News reporting.

How to create a stunning LinkedIn banner with AI:

1. Go to Gemini and copy-paste my prompt (below).
2. Answer the 6 quick questions provided.
3. Upload a reference image of you.
4. Make edits in the same chat.
5. Upload to your LinkedIn.

I teach systems like this inside The AI Creators Club.

Sign up here: https://lnkd.in/enTVYc3z

(Doors close Thursday, Jan 8th)

-- START OF PROMPT --

Act as a professional LinkedIn Banner Designer.

Your goal is to generate a high-converting, visually vibrant, and professional LinkedIn banner that fits the 1584 x 396 pixel dimension perfectly. The design should be dynamic and stand out.

Because you cannot generate this exact aspect ratio, you must generate a Wide (16:9) image, but you must concentrate the design into a thin horizontal strip in the absolute vertical center of the canvas.

To ensure the user can crop this correctly:
1. The top 35% of the image must be empty background color.
2. The bottom 35% of the image must be empty background color.
3. All content (Text, CTA, Face) must be squeezed into the middle 30% strip.

Before generating the image, please ask me for the following details so you can customize the design. Stop and wait for my response after asking:

1. What is the Primary Headline?
2. What is the Secondary Tagline?
3. What is the Call to Action (CTA) text?
4. What Social Proof, stats, or client logos should be included?
5. What are your Brand Colours (Hex codes)?
6. Please upload the realistic photo of yourself. Additionally, specify any REALISTIC professional changes you want made to your appearance (e.g., "Change my polo to a dark suit").

Once I provide these details, generate the banner adhering strictly to these layout rules within the central middle strip:

Layout Rules (Applied ONLY to the middle 30% of the image):

1. CTA Placement (Strict): The CTA button must be anchored precisely in the top-left corner of the central strip. It should be positioned high up, above the level of the subject's shoulder.

2. Subject Appearance (Realistic & Professional): Place the modified user photo in the center-left of the strip. Apply the requested changes but maintain a high-quality, realistic photographic style. Do NOT apply cartoon filters.

3. Headline and Tagline: Place the Headline on the right side of the strip, with the Tagline directly below it. Keep the text compact.

4. Social Proof: Place logos or stats in a small, clean row at the bottom-right of the strip.

5. Background (Vibrant & Geometric): Do not use a plain, flat background. Create a dynamic and modern background using your brand colors. Incorporate subtle geometric shapes, abstract lines, or a professional gradient pattern to add depth, energy, and a high-end feel. Ensure these background elements are subtle enough behind the text to maintain perfect legibility.

Please start by asking me the questions.

-- END OF PROMPT --

Save this. These workflows take me hours to create.

Repost ♻️ to help someone in your network.