🤖 Me: "Multi-agent AI systems sound complicated."
My brain: "Remember when you thought APIs were rocket science?" 🚀
Also me: "Time to dive into AutoGen!" 

🔥 REALITY CHECK: I just spent the weekend building my first multi-agent workflow with Microsoft's AutoGen framework, and my mind is officially blown. 🤯

What started as curiosity about "AI agents talking to each other" turned into watching my code orchestrate conversations between specialized AI agents like a digital symphony conductor! 🎼✨

💡 What AutoGen just taught me:
🧠 One agent = Smart. Multiple agents = Genius level collaboration
💬 Agents can debate, critique, and improve each other's work
⚡ Complex problems get solved through structured AI conversations
🎯 Each agent has specialized roles (coder, critic, executor, planner)

The magic moment? Watching a "UserProxy" agent coordinate with a "Coder" agent and an "Executor" agent to solve a problem I threw at them. They literally had a three-way conversation, debugged each other's suggestions, and delivered a working solution!

It's like having a team of AI experts who never get tired, never have ego clashes, and always stay focused on the goal. 🎪

⚠️ Plot twist: Setting up the agent personalities and conversation flows took longer than I expected. But once they started "talking"... pure magic!

✅ My AutoGen experiments so far:
• Code review workflows (Agent 1 writes, Agent 2 critiques, Agent 3 refines)
• Research synthesis (Multiple agents tackling different angles)
• Problem-solving debates (Agents literally arguing until they find the best solution)

🧠 Why this matters for the future:
We're moving from "human + AI" to "human + AI team coordination." AutoGen isn't just about automation—it's about orchestrating intelligence at scale.

Next up: Building a custom agent swarm for content strategy. Because why have one AI assistant when you can have an entire AI think tank? 😎

Built different. Orchestrated intelligently. Multi-agent approved! ✨

#AutoGen #MultiAgent #AI #Microsoft #Automation #TechExploration #AIOrchestration #FutureOfWork #MachineLearning #AITeams #Innovation #TechLife #BuildInPublic #AIFrameworks

👀 Me: "I should probably post something on LinkedIn today."
My brain: Too lazy to type. 😴
Also me: "What if I make Claude do it for me?" 🤖✨

🚨 PLOT TWIST: This post is NOT written by me! 🚨

Meet my new ghostwriter — Claude, powered by MCP (Model Context Protocol)! 🧠💻
While exploring MCP as my first deep dive into the future of AI integration, I connected to a remote MCP server and thought... "Why not let my AI assistant handle my LinkedIn presence?" 

💡 What just happened:
🔗 Connected to remote MCP server
🤖 Claude is now my personal LinkedIn content manager  
📱 Automated posting with human oversight (because I'm not THAT brave yet)
🎯 First step into MCP exploration = Mind = Blown

This is the future, folks! I'm literally having an AI write about how it's writing this post. Meta level: 100. 🤯

The irony? I spent more time setting up MCP than I would have writing this post manually. But hey, that's what we call "learning by doing" in the tech world! 😂

✅ Next Steps in My MCP Journey:
• Building custom MCP tools
• Exploring multi-agent workflows  
• Maybe letting Claude run my entire social media strategy (kidding... or am I?)

🧠 Why this matters:
MCP is revolutionizing how AI systems interact with external tools and services. Today it's LinkedIn posts, tomorrow it's... well, let's see how far this rabbit hole goes!

Built different. Automated differently. Still human-approved! ✨

#MCP #ModelContextProtocol #Claude #AI #Automation #TechExploration #AIIntegration #FutureOfWork #AutomatedContent #TechLife #AITools #Innovation #MetaPost #BuildInPublic

**Designers are heading to Las Vegas Market next week with a sharper lens.**  
Dallas and Atlanta are behind them—and now it’s not about discovering newness. It’s about choosing what fits.

They’ve seen a lot. Now they’re *editing.*  
Filling gaps. Locking in what’s available *now*, not just what’s aspirational.

**Timelines are tight. Fall installs are barreling in.**  
Designers are asking:  
• *Will this ship in time?*  
• *Does it complete the plan?*  
• *Can it hit my client’s budget?*

Yes, stories still matter—but only if they’re *crisp.*  
*Livable luxury* is trending: sculptural lighting, warm woods, organic materials.  
But only if they *hang together.*

More designers are wearing two hats—specifier and retailer.  
Shopping for both projects *and* their storefronts or ecomm sites.  
That shifts the questions:  
• *What’s the margin?*  
• *Can I reorder?*  
• *Will this brand make my life easier?*

**And let’s be honest—trade program fatigue is real.**  
Designers don’t want more portals.  
They want *clarity, service,* and a *human* who picks up the phone.

—

**Bottom line:**  
Designers landing in Vegas on July 27 are more focused, more prepared, and more *ready to buy* than browse.

If you’re a brand with *style, service,* and *speed*—you’re in a great spot.

**What are you hearing headed into market week?  
What’s changed in your playbook this season?**  
Let’s compare notes.

#lasvegasmarket #interiordesign #summermarket #productstrategy #retailandtrade

Scaling success is seductive. But in student support systems, replication without adaptation can do more harm than good.  

Last year, I worked with a district that wanted to replicate a pilot success across 30 campuses. But what worked in one high-poverty urban school didn’t land the same in rural or suburban sites.

Why? Because culture, leadership, and student demographics matter.

We paused. We co-designed. We customized.

And instead of a top-down copy-paste model, we created a flexible framework grounded in core principles but responsive to context.

Scaling doesn’t mean duplicating. It means designing for adaptation.

How have you balanced consistency and customization when scaling equity-driven work?

On the afternoon of June 16th, a user - we’ll call her Kate - sent me a message that INSTANTLY made me drop out of a meeting:

"Heyy ran into a huge issue… Mike's post that I scheduled from his LiGo account got posted from my LinkedIn somehow."

Most SaaS founder know this feeling. Your stomach drops. Your product just failed someone publicly.

For LiGo, this was DEFCON 1. Kate was managing 7 LiGo accounts through our delegate access (team management) feature, and somehow a post scheduled for one of those accounts got published on HER LinkedIn profile instead. 

Here's how it was handled:

My immediate response wasn't a ticket number or a request to "do X, Y, Z manual actions" 

It was a direct assurance from a founder: "we're on it."

We knew the next post for Kate's client was scheduled for 2 PM UK time. The clock was ticking. 

The first step was to understand the "why." 

After a quick back-and-forth, we found the root cause: TWO out of the 7 LiGo accounts she managed, had inadvertently been linked to the same LinkedIn profile. User mistake triggered by a UX issue in one of LiGo's component 
(at the time of connection, it doesn't show which LiGo account is currently selected in the profile picker). 

Our promise to her was simple: "This would be sorted before 2PM BST."

From our systems, we manually severed the duplicate connection.

Kate reconnected the right profile on her end. 

We verified manually, and then migrated all 20 scheduled posts to the correct account. 

We stayed in direct contact until she confirmed: 
"Worked! Went out from the correct account!"

This incident could have been a churn-worthy disaster. Instead, it turned the user into our advocate (that same day she recommended us to 2+ businesses, hopefully more by now haha, and shared with me screenshots of feedback the referred users had shared with her)

TLDR;

1. When a user is in a panic, the last thing they want is a chatbot. They want to know a human understands the stakes and is taking ownership. That’s a competitive advantage that no amount of venture capital can buy.

2. Software fails. And when it does, your response becomes your brand. 

---

In the coming weeks, I'll try to share more founder-led customer support cases to show just HOW critically important CS is, and why it should ideally be done by a founder directly (until it absolutely cannot be).

Just discovered npx ccusage - the perfect tool to know WHEN to upgrade to Claude Code Max! 📊

Huge shoutout to @IndyDevDan on YouTube for this game-changing tip! Started using Claude Code just 5 days ago (June 19th) and already hitting my limits. The `npx ccusage` command gave me the data I needed to confidently upgrade to Max.

My usage in just 4 days:
- 23.7M total tokens
- $66.20 in costs
- Already approaching plan limits

At this rate, I'd hit $400+/month - making the $200 Max plan a no-brainer. The unlimited usage will actually SAVE me money while removing the friction of hitting limits mid-project.

What I love about `npx ccusage`:
✅ Shows exactly where tokens are going (cache saved me 22M tokens!)
✅ Daily breakdown helps spot usage patterns
✅ Real costs = informed decisions
✅ No more guessing if you need to upgrade

If you're constantly hitting limits or worried about token usage interrupting your flow, run this command. The data might surprise you - sometimes unlimited is actually the budget option.

For new Claude Code users: track early, upgrade smart. Your productivity will thank you!

Check out IndyDevDan's channel (youtube.com/@indydevdan) for more Claude Code tips. Anyone else find they're using way more tokens than expected?

#ClaudeCode #DeveloperProductivity #AITools #MCPPosted #ClaudeDesktop

🚀 Nasdaq gets it! 

While others are still figuring out AI integration, Nasdaq is already building the infrastructure of the future with their Model Context Protocol (MCP) server. This is what being ahead of the curve looks like - providing seamless access to financial data through standardized AI interfaces.

Major kudos to Stefano Morelli for setting this up! 👏 This kind of forward-thinking approach is exactly what we need to bridge traditional finance with AI capabilities.

The fact that Nasdaq is investing in MCP shows they understand that the future isn't just about having data - it's about making that data easily accessible to AI systems that can help us make better decisions faster.

This is innovation in action. 💡

#AI #Finance #Innovation #MCP #Nasdaq #FinTech

https://github.com/stefanoamorelli/nasdaq-data-link-mcp
https://www.linkedin.com/in/stefanoamorelli/

„Weiterbildung, die niemand braucht.“

Wenn Menschen ins Coaching kommen, erzählen sie oft:
„Ich war schon in zig Seminaren – aber irgendwie hat sich nichts verändert.“

Das Problem ist nicht die Weiterbildung.  
Sondern **deren Passung.**

📎 Excel für jemanden mit Strukturstärke?  
✔ Vielleicht hilfreich – aber nicht transformierend.

📎 Kommunikationstraining für jemanden mit Präsenz, aber keiner Klarheit?  
✔ Schön – aber nicht wirksam.

📎 Achtsamkeit für jemanden mit hoher Ordnung und innerem Leistungsdruck?  
✔ Gut gemeint – aber zu vage.

➡️ Weiterbildung wirkt nur, wenn sie zur *Persönlichkeit* passt.  
Wenn sie nicht allgemeine Tools vermittelt – sondern **innere Muster sichtbar macht.**

🎯 Entwicklung heißt:  
- Stärken erkennen  
- Konfliktmuster verstehen  
- Wirkung gezielt verändern  

Nicht: Noch mehr Methoden.  
Sondern: Mehr Bewusstsein.

Ich frage Führungskräfte oft:  
**„Was müsste passieren, damit du in einem Jahr anders führst?“**  
Die Antwort liegt selten im Methodenkoffer – sondern im Selbstbild.

🗝️ Mein Appell an Unternehmen:  
Investiert in Entwicklung, die *wirkt*.  
Nicht in Programme, die gut aussehen.

💬 Was war die unnötigste Weiterbildung, die du je gemacht hast?  
Oder die wirkungsvollste? Teil’s gern in den Kommentaren.

#JanKleinigCoaching #Persönlichkeitsentwicklung #SmartLeadership #Wirksamkeit #WeiterbildungMitWirkung #CoachingStattKatalog

If your Google Ads aren’t getting clicks...
you might as well be burning money.  

I’ve been there.  

Running campaigns.  

Tweaking words.  

Spending more — hoping something sticks.  

And then I realized — it’s not just about showing up.  

It’s about **how** you show up.  

---

Most ad copies try to sound smart.  

But they end up saying nothing.  

You’ve probably seen this:  
“Boost your business.”  

Sounds nice.  

Means nothing.  

Now this?  
“Get 43% more bookings with retargeting.”  

That’s real.  

That gets attention.  

---

Here’s another one.  

Don’t write like you’re talking to Google.  

Write like you’re talking to a real person.  

Your customer.  

Use their words.  

Talk the way they talk.  

That’s how you connect.  

---

And please — add some urgency.  

No one acts unless they feel a reason to.  

“Only 3 spots left.”  

“Offer ends Friday.”  

Tiny lines.  

Big impact.  

---

Now, let’s talk CTAs.  

“Learn more”?  

It’s a whisper.  

Try:  
“Start getting leads today.”  

Or:  
“Unlock your first sale now.”  

Much stronger.  

---

Last thing — test feelings too.  

Emotional lines can hit harder than logical ones.  

“Tired of wasting money on ads?”  

Sometimes works way better than  
“Track your ROI efficiently.”  

---

It’s not about clever writing.  

It’s about being clear.  

Being real.  

And helping the right people say:  
**Yes — this is what I need.**

Writing high-quality posts with GPT does NOT take the same amount of effort as writing them manually. All the personal brand coaches and ghostwriters are lying to you.

If you’re selling ghostwriting services in 2025, you have to pretend GPT can’t write like you.  
You have to downplay the speed, the scale, the nuance, because if people knew the truth, most of your pitch would collapse.

But let’s be honest: no one hires a ghostwriter for THEIR perspective.  
They hire them to amplify THEIR OWN.  
And GPT, when set up properly, does that better than any freelancer ever could.

They will never tell you this but I will:

Yes, the first AI-generated post might take you 15–20 minutes. You’re building context, aligning tone, calibrating structure.

But once that's in, you don't start from zero anymore. You're operating from compounding context - and after 5 or so posts, you’ve got your own personal ghostwriter at $20/mo.

Here's the workflow I used to write this post:

1. I asked GPT (using our CustomGPT integration) to analyze my last 5 posts - tone, pacing, structure, and what worked.  
2. I brainstormed inside the same thread. It filtered ideas, surfaced patterns, and helped narrow in on the strongest angle.  
3. I told it: write it in my voice - no gimmicks, no cheap tricks.  
4. I cleaned up a few lines. That’s it.

To ghostwriters and personal brand coaches:  
Just admit to using AI. It won't kill your pitch.  
Your business was never about YOUR copywriting skills - it’s about your empathy, your grasp of nuance, your understanding of what works on this platform.

All of that can be prompted.  
And the ones who lean in will win.

Try it yourself:  
👉 https://chatgpt.com/g/g-680e4fedf0d48191a0efc638dc3fef8f-ligo-for-linked-post-generator-scheduler

"Customer Support" is the worst f*cking name for the most important department in your company. It's not support. It's your business. Every other function - marketing, sales, product - they're the support functions.

But somehow we've convinced ourselves that the people talking to customers daily are just "handling complaints."

Most founders miss this completely. 
They outsource customer support because they think it's beneath them. 
Or treat it like a cost center.

Wrong on both counts.

The one place you want to invest heavily? Customer support. People should be delighted even if they're churning. Happy even when they're leaving you.

We've had people ask us to delete their LiGo accounts. We just did it. "Sorry to see you go, we'll process this. If you have time, feedback helps us improve."

We've had users ask for refunds because of "missing features" that were literally in the platform. They just didn't explore it. We processed the refund first, then mentioned the features existed.

Nobody asks for a refund unless the product completely flopped on them. That's market intelligence. That's your product roadmap talking.

We've had people bombard us with the same questions. Could we build an AI to handle "Why isn't my post getting reach?" Sure. But we answer personally. Every time.

Users asked for Spanish subtitles on our walkthroughs. We delivered.

When someone hits an API error, we monitor the backend, catch it, and send the solution via LinkedIn and email before they even complain.

This isn't "customer support."

This is market research.
This is product development.
This is your competitive advantage.

When you outsource customer conversations, you outsource your understanding of the business.

Marketing can be outsourced. Sales can be outsourced. Even product development can be outsourced.

But the moment you stop talking to your customers directly, you stop running a business.

Call it "Business Intelligence." Call it "Customer Intelligence." Call it "Market Pulse."

Just don't call it support.

This isn't about supporting your customers.

It's about understanding them.

And that understanding? That's your entire business.

Your first 10 customers should scare you.

Not because they’ll churn.  
But because they’ll expose you.

Most early users aren’t “ideal ICPs.”  
They’re needy. Demanding. Inconsistent.  
They’ll break your product in ways you didn’t think possible.

And that’s a gift.

Because these customers don’t treat your product like a demo.  
They use it like a tool.  
And tools get slammed, twisted, overused.

One founder told me:  
“Our first user made a custom Notion database just to track our bugs.”

That’s what you want.

The polished feedback from “enterprise pilots” and “strategic POCs”?  
It’s mostly fluff.

You want raw.  
You want complaints.  
You want someone willing to send a 300-word Slack message because one button didn’t work.

Why?

Because they reveal edge cases.  
They surface assumptions you didn’t realize you’d made.  
They force you to make real product decisions.

If your first 10 users feel easy...  
You're either in the wrong market  
Or you’re not paying attention.

LinkedIn has been pushing the Boost feature hard so I took the bait and boosted a post.  
Here are my sincere thoughts: It’s a trap for lazy companies.

I spent $35.58. Got 29 website visits.  
Total impressions: 12,557  
Boosted impressions: 7,961  
Which means ~4,596 came organically.

You’d think boosting helps you reach a new audience.  
And yes, technically it can - especially if you insert new targeting that is outside your general network.  
But let’s look at the ratio.

With 7,961 Ad Only impressions, I reached 5,344 members.  
That’s a reach-to-impression ratio of about 67%.

Now compare that to a typical organic post.  
It’s very possible sometimes to get 5,000 impressions and only reach 2,000 people. That’s just the same people seeing it more than once.

So, does boost improve that ratio?  
Not really. You can see the math yourself.

Organic impressions: 4,596  
Organic reach: 4,371

That's a 95% ratio.

Conclusion: Boosted posts are actually WORSE for reach.

Now let’s talk contrast.

That boosted post took me 5 minutes to write.  
So did my last two organic posts.

One hit 185,704 impressions.  
The other hit 54,424.  
Total spend: $0.

So on one hand, 5 minutes + $36 = 12.5K impressions  
On the other, 5 minutes + $0 = 185K or 54K impressions

Which one would you pick?

Now sure, the boosted post got 29 clicks.  
Organic posts usually don’t - unless you know how to work the system.

Here’s how:  
Let the post ride for 3–4 hours.  
If it’s getting traction, edit in a link.  
Reach won’t collapse. I’ve done it. It works.

The real issue?

Most companies have 50+ employees and just one person posting on LinkedIn.  
The rest of the team is silent - while ad money gets burned for reach that could’ve been free.

That’s what LiGo fixes.

One person manages content.  
Everyone shows up.  
Nobody else has to log in.  
No training required.

So yeah, if you’ve got money to burn and don’t want to build a content culture, sure - boost away.  
But if you’re trying to scale reach, influence, and authority?  
Start by activating your team.

There are better ways to get in front of new audiences.  
Ask me in the comments - I’ll share them.

ChatGPT wrote 5 of my last 10 LinkedIn posts. Here's what happened:


Zero people called it "AI content."
They got 100K+ impressions & 700+ engagements.

Everyone says AI writing is obvious, and BAD. 
I think they're dead wrong. 

Here's my stats so far:

Post 1: LinkedIn comments strategy → 199 engagements
Post 2: Most effective marketing strategy → 33 engagements
Post 3: AI job disruption reality → 72 engagements 
Post 4: Distribution over content → 15 engagements
Post 5: Elite engineers as bad founders → 301 engagements

I use a GPT that knows ME:

- has my public LinkedIn data
- has my LinkedIn posts
- knows my dense writing style
- knows my contrarian takes on different topics 

You can do the same:

𝗙𝗼𝗿 𝗖𝗵𝗮𝘁𝗚𝗣𝗧 → Search "LiGo" in GPTStore
𝗙𝗼𝗿 𝗖𝗹𝗮𝘂𝗱𝗲 → Google "LiGo integrations" and follow the steps

𝘞𝘳𝘪𝘵𝘦 𝘢𝘯𝘥 𝘱𝘶𝘣𝘭𝘪𝘴𝘩 without leaving the chat.
Schedule posts directly from ChatGPT/Claude. 

AI that knows your context writes like you on your best day.

---
This post is scheduled directly from ChatGPT.

Confession: ChatGPT wrote my last semi-viral post. 


60K+ impressions
300+ engagements.
4 reposts.

The post argued that elite engineers make terrible founders.
People either loved it or hated it.
The debates in comments were intense.


What mattered: People cared about the opinion.
They agreed or disagreed passionately.
Nobody questioned if it was "AI-written".

Behind the scenes: 

- It was a CustomGPT that studied my last 20 posts.
- Learned my voice. Topics. My contrarian angles.
- Wrote in my exact style, and then published it to my linkedin profile. 

You can use this same GPT with YOUR profile:

𝗙𝗼𝗿 𝗖𝗵𝗮𝘁𝗚𝗣𝗧 → Search "LiGo" in GPTStore
𝗙𝗼𝗿 𝗖𝗹𝗮𝘂𝗱𝗲 → Google "LiGo integrations"

One-time setup. Then your AI becomes your co-writer.
Write, publish, schedule - all without leaving the chat.

When AI knows your context deeply, it writes your thoughts better than you do on a tired Tuesday.

No one will blame you for using ChatGPT to frame your opinion. 
They will if the entire thing doesn't "sound" like you at all. 

I had ~5400 followers before I removed nearly 900 connections (and hence followers) in the last 2 weeks. I'm now standing at 5,341.

I’ve talked about this before - that your content dies when your network isn’t your audience.  
Here's the post if you missed it:  
https://www.linkedin.com/posts/muhammad-rassam_im-removing-software-engineers-from-my-connections-activity-7323671198238683136-8nZB

But people still hesitate.

“What if I want to hire?”  
Cool. Use LinkedIn Recruiter. Post a job. Ask your network. But don’t compromise visibility and reach on every single post just so you might maybe hire one person someday.

“What if I don’t want to remove friends?”  
You’re not removing them from your life - just from the algorithm. They won’t get notified. Most won’t even notice. And if they do, they’ll survive.

“What if I want to keep my options open?”  
Sure. But if your content is showing up in front of the wrong crowd, it’s not optional - it’s wasted. No one buys from you. No one shares. You get impressions that don’t matter.

And yeah, “what if I want a marriage proposal too?”  
Look - there were a lot of pretty girls I removed while doing this. And I won’t lie, it felt bad. But business comes first.

LinkedIn’s algorithm today is different. It promotes intentional networking over the blind viral style of old. The ‘suggested post’ feature can break you into new audiences organically - but only if your current audience is aligned with your content.

Your content needs a clean launchpad. Not dead weight.

Trim accordingly.

🎨 Behind the Scenes: Crafting Click-Worthy Thumbnails

People think thumbnails are about flashy fonts and big faces.

But here’s the truth I’ve learned: it's about *predictive storytelling*.

Before I even start designing, I ask:
- What emotion should this trigger?
- What promise does the title make?
- Can I *visually hint* at the video’s climax?

One small tweak I’ve been making lately: reducing clutter.
More whitespace. One focal point. Clean, punchy text.

Crazy how click-through rates spike when you stop overloading the frame.

Thinking of sharing a few before-afters next. Would that be helpful?

👇 Let me know what *you* look for in a thumbnail.

Apparently, if you use an emdash in a LinkedIn post, it means ChatGPT wrote it.

That’s the bar now.

Not the insight.  
Not the nuance.  
Not whether the post actually says anything useful.

Just - “oh no, a long dash… must be AI.”

It’s such a dumb take.

I want to use emdashes.  
They’re useful. They read well. They don’t hurt anyone.

But I don’t - because the second I do, someone in the comments will yell “robot.”

And it’s not worth the energy to explain to them that:  
a) GPT sounds the way it does because we trained it to, and  
b) maybe their post about “hard truths” and “growth mindsets” is the real AI slop.

Suck my emdash.

I write every post with AI.  
Not because I’m lazy.  
But because I like bouncing ideas off something that doesn’t argue back.

The trick is simple:  
Don’t let it write like a copywriter.

No hooks. No threads. No lead-ins.

Just a blunt prompt and a high bar for what feels like me.

And when it works - it really works.  
Not because it’s polished.  
But because it isn’t trying so hard to be.

So yeah, maybe one day I’ll bring back the emdash.  
Until then, you’ll just have to imagine it.

People keep lying that “AI won’t take your job.”

I don’t know why.

I mean.. it’s already happening.

And we’ve seen this movie before.

2022 and 2023 were the “efficiency” years.
Google, Meta, Amazon, Salesforce - tens of thousands of layoffs, all in the name of trimming fat.
Wall Street rewarded them instantly. Stock prices shot up.

Now comes part two.

But this time, it’s because of leverage.
Because the tools are getting good.
Really good.

You don’t need five writers when one knows how to wield Claude.
You don’t need three designers when one knows how to prompt Midjourney or ChatGPT.

The first wave was “AI won’t take your job, someone using AI will.”
The second wave will probably be very unforgiving to that ‘someone using AI’ as well.

Microsoft didn’t lay off 7,000 employees by accident.
They didn’t pour $10B into OpenAI for fun.

They’re doing the math.

And no - I’m not saying all these people will stay unemployed.

Some will upskill. Some will pivot. Some will become solo founders, fractional consultants, or creators.

But let’s stop pretending AI isn’t disrupting jobs.
It is. It will. And the pace is accelerating.

If you're a 10x engineer, you're most likely to be a 0x founder. 


Because the skills that made you great at writing code?
Are the same ones that make you terrible at building businesses.

10x engineers are wired to:

- eliminate edge cases
- automate everything
- think five steps ahead

Which is great…
if you're at Google.
Or building Stripe’s fraud engine.

But if you're in the early stages of your startup?
All of that slows you down.

 You need speed.
 You need to build momentum.

And it gets worse:

When 10x engineers become technical co-founders, they often drag the entire company into over-engineering purgatory.

they’ll spin up a containerized backend, battle over frameworks, and debate infrastructure decisions like they’re building the next AWS.

All before they’ve validated if anyone gives a damn.

When you're at Google, you're rewarded for eliminating 0.1% failure modes.

When you're an early stage startup, you're rewarded for finding just one person who’ll pay.
That’s a different game.

A 10x engineer optimizes systems.
A 1x founder optimizes for momentum. 

If you have to choose early on ...
Always choose the one who’s obsessed with forward motion. 
NOT perfection.

Since April, I've personally lost about 60% of my reach on LinkedIn, but I'm not concerned.



Every platform eventually betrays its creators:
* Facebook pages saw 80% organic reach drops
* Instagram shifted to favor video over photos
* Twitter/X made verified reach a paid feature
* LinkedIn now penalizes external links

Platform dependency is a silent killer.

Your entire business sits on rented land, controlled by companies who don't care about your growth targets.

You're one policy change away from irrelevance.

Here's why I am not concerned:

A year ago, pretty much everyone pushed me to focus on a single channel and master that alone.

I didn't listen.

And now I'm thankful for that decision.

Because I never put all my eggs in the LinkedIn basket:
* Created my newsletter a year ago
* Launched a product-specific newsletter weeks ago
* Ensured LinkedIn only contributes 10-20% of our total growth (reddit, medium, blog, insta, youtube are others)

So if LinkedIn's feeling "moody" right now?

I'll wait out the storm and focus more on other platforms for the time being.

My growth doesn't depend on their algorithm.
Neither should yours.

If your ONE platform disappeared tomorrow, what's your backup?

50 vrais posts LinkedIn Écrit
avec LiGo + ChatGPT et Claude

Vous voulez que votre article soit présenté ici ?

50 vrais posts LinkedIn Écrit avec LiGo + ChatGPT et Claude

Vous voulez que votre article soit présenté ici ?

50 vrais posts LinkedIn Écrit
avec LiGo + ChatGPT et Claude