Ruben Hassid

Master AI before it masters you.

Founder

Master the AI revolution with me.

Go to: rubenhassid.ai

Step 1: Grab my $1 ChatGPT Pack.

→ 1600+ prompt library.
→ 100 AI tools, curated.
→ Free access to weekly videos.

Step 2: Get instant access to my course.

→ GPT-4 & content creation.
→ Midjourney & design.

Step 3: Follow my journey.

→ Twitter: @rubenhssd
→ Instagram: @rubenhassid.ai
→ Tiktok: @rubenhassid.ai
→ YouTube: @ruben hassid

Who am I?

☑️ Top #2 Linkedin Creator.
☑️ Top #15 AI Content Creator.
☑️ My content reached 50M+ people. 

Ranking: Favikon.

You just quit ChatGPT for Claude.

But you're lost on how to use it. Here is how:

1. Go to Claude (desktop app).
2. Pay for the monthly $20 plan.
3. Open the Cowork (it's the best).
4. You must create a "Cowork" folder on your PC.
5. To download my folder, go here: how-to-ai.guide.
4. Don't pay anything. It's free in the welcome email.
5. Then get my Notion with text files to download.
6. I made a guide for the next steps for Cowork.
7. Copy my entire setup: claude-co.work

---

1. Chat → just like ChatGPT.
☑ Make sure to use Opus 4.6.
☑ Always turn on 'extended thinking'.
☑ I barely use chat now. I only use Cowork.

2. Project → Chat, but separated by project.
☑ One Project per client. One per topic.
☑ Drop your files inside to keep Claude focused.
☑ I replaced Projects with Cowork. It does more.

3. Code → a revolution for developers.
☑ Developers code with Claude directly.
☑ It writes, debugs & ships. No copy-pasting.
☑ Skip it if you don't code. Cowork is your Claude Code.

4. Cowork → the only one that matters right now.
☑ Create 4 folders under one main folder 'Cowork':
☑ 'About me', 'templates', 'projects', 'Output. md'
☑ To download mine, go to how-to-ai.guide.
☑ Pay nothing. Get the link to download in the mail.

5. Skills → teach Claude once, reuse forever.
☑ Build a Skill for every task you repeat weekly.
☑ It's like a Project with permanent instructions.
☑ One good Skill replaces an entire SOP document.

6. Connectors → Claude inside your tools.
☑ Settings → Connectors → Browse → click "Add."
☑ Slack, Google Drive, Notion, Figma. 50+ tools.
☑ Claude reads, writes & acts inside them. It's free.

7. Plugins → pre-built workflows for your role.
☑ Customize → Browse plugins → Install. Also free.
☑ Type / to trigger commands like '/marketing'
☑ Customize each plugin to your company.

♻️ Repost this image to help others quit ChatGPT.

How to set up Claude the right way: 

(so you actually stop going back to ChatGPT)

1 - Download this infographic. Send it to your team.
2 - Switch between chat vs. cowork vs. project.
3 - Pro tip: Use cowork 80% of the time.

To download all of my other Claude infographics:

Step 1. Go to how-to-ai.guide.
Step 2. Subscribe for free. Don't pay anything.
Step 3. Open my welcome email (most skip this).
Step 4. Hit the automatic reply button inside.
Step 5. Download my infographics from my Notion.
Bonus. Enjoy my best copy-paste prompts, too.

You now master Claude within 5 minutes.

But if you have 50 more minutes, read in this order:

Claude 101 full article: https://lnkd.in/dVDent-3.
Why I quit ChatGPT: https://lnkd.in/dV2FJYpf.
Stop "prompting": https://lnkd.in/dPePDVXx.
Claude & text file: https://lnkd.in/ee29GvFF.
Claude in Excel: https://lnkd.in/ejHD2Tb9.

♻️ Repost this image to help others quit ChatGPT.

How to set up Claude the right way: 

Chat vs. Project vs. Cowork

Google is offering free AI courses with certificates.

It's free by following these 7 links:

1 - AI Fundamentals

Understand fundamental AI concepts and build a foundation for professional AI use. Write prompts using a structured framework.

Access it: https://lnkd.in/ddvC9pXm

2 - AI for Brainstorming and Planning

Brainstorm project concepts by using AI as a creative tool. Evaluate & prioritize ideas by using AI to assess them against decision criteria.

Access it: https://lnkd.in/dDHdinbj

3 - AI for Research and Insights

Use Gemini and NotebookLM to research faster and summarize multiple sources instantly.

Access it: https://lnkd.in/di_mpJBh

4 - AI for Writing and Communicating

Turn messy meeting notes into clear summaries and action items with Gemini.

Access it: https://lnkd.in/d27xRvHq

5 - AI for Content Creation

Generate high-quality images & videos using Gemini. Transform slide decks into engaging visual presentations with Gemini in Google Slides.

Access it: https://lnkd.in/dCzTWYrj

6 - AI for Data Analysis

Use AI to track metrics, clean messy data, and measure performance.

Access it: https://lnkd.in/d7sAz78A

7 - AI for App Building

Build a working web app using AI - no code needed.

Access it: https://lnkd.in/dUHQt78B

8 - How to master AI with words (not code)

Shameless plug: it's my own (free) newsletter.

Join 304,000+ weekly readers at how-to-ai.guide.

♻️ Repost this to help others access AI courses.

Stop using ChatGPT for everything.

These are the only 8 tools I (actually) swear by:

1. Wispr.ai → for voice-to-text

You type your prompts into ChatGPT like it's 2024.
I just talk. Wispr types. And it learns my edits.
80 WPM → 200+. Never going back.

2. Granola.ai → for meeting notes

You paste transcripts into ChatGPT after the call.
Granola takes notes during the call. Silently.
Does NOT join like a bot. Just works.

3. Gamma.app → for slides

You ask ChatGPT to "make slides." It gives you text.
Gamma actually builds the deck. One prompt. Done.
The best PowerPoint killer. Period.

4. Claude.ai → for writing

You rewrite ChatGPT outputs 5x to sound human.
Claude 'Cowork' gets it right on the first try.
Teach it with your voice. Upload .md files.

5. Grok.com → for search

ChatGPT doesn't know what happened 2 hrs.
Grok does. Real-time tweets pulled straight from X.
Ask a question. Get facts. Not guesses and links.

6. Nanobanana (gemini.google.com) → for images

DALL-E gives you animated, cartoonish images.
NanoBanana gives you exactly what you asked for.
Clean. Believable. Editable. No more 5th try.

7. Notebooklm.google.com → for research

You paste docs into ChatGPT & hope to train it.
NotebookLM reads sources & becomes your expert.
Upload your sources. Ask anything.

8. Opus.pro → for video

ChatGPT can write a video script. That's it.
OpusClip takes one YouTube link → 20-40 shorts.
What took 4 hours now takes 4 minutes.

ChatGPT made you work in 2025.
These 8 tools work for you in 2026.

Save this. Try one tool this week. Thank me later.

You want to help your network?

♻️ Repost this image so they can find the best AI.
Subscribe to my free newsletter: how-to-ai.guide.

The Anatomy of a Claude 4.6 Prompt:

1. Task

Define what you want & what success looks like:
"I want to [TASK] so that [SUCCESS CRITERIA]."
No roles, "act as a senior expert." That era is over.

2. Context Files

Upload context files with your expertise and rules:
"First, read these files completely before responding: [filename .md] - [what it contains]."
AI went from reading a sticky note to an entire book. 
Stop explaining yourself in the prompt. Put it in files.


3. Reference

Show AI exactly what you want. Upload an example.
Then give patterns, tone & structure as rules. No "give me something like" & hoping AI figures it out.

4. Brief

This is the only part you actually type from scratch. Everything else is files. "Type of output + length.
Does NOT sound like. Success means."

5. Rules

Context file holds your standards, taste & audience.
Prompt: "Read it fully before starting. If you're about to break one of my rules, stop and tell me."

6. Conversation

You spent 3 years prompting AI. Now it prompts you
Prompt: "DO NOT start executing yet. Ask me clarifying questions (use 'AskUserQuestion' tool) so we can refine the approach together step by step." 

7. Plan

Claude read your files before writing a single word.
Prompt: "Before you write anything, list the 3 rules from my context file that matter most for this task. Then give me your execution plan."

8. Alignment

Nothing happens until you both see the same aim. This replaces the old prompting era.
Prompt: "Only begin work once we've aligned."

Copy the full prompt template + download my personal md. files for Claude here: how-to-ai.guide.

How to set up the Claude completely in just 1 day: 

(Save this. Send it to your team. Thank me later.)

Tool 1: Cowork

Download Claude (claude .com/download).
Open the Cowork tab. 
Select a folder from your computer.

Now build your brain file. Create "about-me .md."
To download mine, go here: how-to-ai.guide.
Don't pay anything. It's free in the welcome email.

Then, start every chat with this prompt: 
"Read the files first. Then ask me questions before doing anything."

Tool 2: Model

Pick Opus 4.6 (it's the best model right now).
Turn on Extended Thinking.
Most skip this & wonder why Claude feels generic.

Then go to Settings > Connectors. 
Link Slack, Google Drive, Notion.
Claude now reads your actual workspace.

Tool 3: Excel

Open Excel. Go to Insert > Get Add-ins. 
Search "Claude by Anthropic." Install.

Open any spreadsheet. Ask: "Give me a summary of each tab." Claude lives inside the spreadsheet. It knows what D14 actually contains.

Tool 4: Plugins

Go to claude .com/plugins. 
Pick the one as per your job.
Marketing. Sales. Legal. Finance. Data.
Click install. Type / in your next chat.

Legal software companies lost $285B in market cap when these dropped. That's a signal.

Tool 5: Artifacts

Nothing to install. Ask Claude to build something.
"Create an interactive HTML calculator for [thing]."

You get a working tool inside the chat. Not text. Not a suggestion. A thing you click and use.

Tool 6: Projects

Go to claude .ai. Click Projects. Create one. Upload your files and add instructions. Every chat inside the project remembers your context.

Honest take: Cowork replaced this for me. 
But for teams, it still works.

Where Claude falls short:

No image generation. Use Gemini.
Not the best at real-time search. Use Grok.
Not the best at everything. No tool is.

But for writing, thinking, analyzing, and working with your files? It's the best.

The real difference:

Most people open Claude, type one prompt, get a mid response, and leave.
Because they set up one tool out of six.

Set up all six. It takes 1 day.
Then you'll understand why everyone switched.

Full setup guide + features: https://lnkd.in/dVDent-3.

Stop writing 500-word prompts that don't work. 

This 29-word prompt writes better than all of them:

"Read my anti-AI writing style file first. It contains every known pattern of AI writing I want to avoid. Apply these as rules to everything you write for me."

That's it. But you need to set it up first. Here's how:

Step 1. Go to Wikipedia. 
Step 2. Search "Signs of AI writing." 
Step 3. Copy the entire page. 
Step 4. Paste it into a Google Doc. Don't edit. 
Step 5. Name it "anti-ai-writing."
Step 6. Download as .md format. 
Step 7. This is your "what NOT to sound like" file.

Or skip all of that. To download the anti-AI guide: 

Step 1. Go to how-to-ai.guide.
Step 2. Subscribe for free. Don't pay anything.
Step 3. Open my welcome email.
Step 4. Hit the automatic reply button inside.
Step 5. Download my .md files. Ready to upload.
Step 6. Upload it to Claude. Prompt: 

"Read the uploaded file. It contains every known pattern of AI writing I want to avoid. Apply these as rules to everything you write for me. Do NOT start writing yet - ask me clarifying questions first."

✦ Here's why your prompts don't work (red flags):

"Don't use jargon."
"Don't sound like an AI."
"Don't use buzzwords or filler."
"Avoid passive voice."
"Be conversational, not robotic."

These are everywhere. LinkedIn posts. Emails.
They sound thorough. The output is still garbage.

Your prompt says "don't" 14 times. The model forgets half by sentence three. You're fighting the AI with a wall of "don'ts." It doesn't work at scale.

The fix is counterintuitive.

Stop telling the AI what to avoid. 
Give it a file that shows what to avoid. 

The model reads 1,168 lines of bad patterns, internalizes them, and writes clean.

500-word prompt → still robotic.
Small prompt + 1 file → reads as a human wrote it.

The file does the work. Not the prompt.
To get access, subscribe to how-to-ai.guide.

♻️ Repost to help one person stop sounding like AI.

This is scary: GeoSpy AI tracks your exact location from a social media photo in 2 seconds.

The process can't be simpler:

1. Upload any photo.
2. Get precise GPS coordinates.
3. Displayed in 3D Street View

No metadata needed. No EXIF data. Just pixels.

You can track down the location using ChatGPT too.
But GeoSpy takes it 10 steps further:

☑ It reads architecture, vegetation & lighting.
☑ It narrows from country → city → street in secs.
☑ The advanced model claims ~1-meter accuracy.

We are entering a new era of visual surveillance:

☑ A selfie can expose your exact home address.
☑ AI analyzes pixels you didn't know carried cues.
☑ Street-level 3D accuracy. In under 2 minutes.

Think twice before posting a "casual" photo.

This is the only Claude Cowork guide you need. 

(duplicate my exact folder, files, and prompts):

Step 1: Download the app (claude. com/ download)
Open the Cowork tab. Select the 'Opus 4.6' model.
Turn on Extended Thinking. Never change these.

Step 2: Build this exact folder on your laptop
One master folder: "Claude-Work."
4 subfolders: About me, Project, Template, Outputs.
Point Cowork to this folder. That's how it reads you.

Step 3: Create 3 context files inside ABOUT ME
about-me → what you do day-to-day. Not a resume.
my-voice → tone, phrases you hate, 3 writing pieces
my-rules → ask, plan, don't delete without approval.

Step 4: Or skip step 3. Download my own md files-

Go to how-to-ai.guide.
Subscribe for free. Don't pay anything.
Open my welcome email.
Hit the automatic reply button inside.
Download my .md files. Ready to upload.

Step 5: Set Global Instructions (once, forever)
Settings → Cowork → Edit Global Instructions.

Paste: "I'm [Name], [Role]. Read my files before every task. Ask questions before executing. Show a plan. Never delete without approval."

Set once. Runs every session. Never type it again.
Your prompts can now be 10 words long.

Step 6: Stop prompting. Start answering questions.
Claude generates clickable question form. Not text.
You select answers. No typing.

Prompt this for every task:
"I want to [TASK]. Read all files first. Ask me questions using AskUserQuestion before you execute. If something is off, generate new questions. Do not guess."

Step 7: Install your first plugin
Customize → Browse Plugins → Install.
Type / to see slash commands.
Marketing. Data. Legal. Sales. Start with 2.
This is why $285B in software stocks disappeared.

Step 8: Connect your tools
Settings → Connectors → Add.
Slack. Google Drive. Notion. Figma. 50+ more.
Claude reads your tools mid-conversation. 
No copy-pasting. Free on all plans.

The Don'ts:
✦ Write long prompts for every task
✦ Skip the folder and wonder why output is generic
✦ Dump 50 random files without structure
✦ Use Chat when you need Cowork
✦ Send deliverables without reading them first
✦ Forget Extended Thinking
✦ Expect Claude to read your mind without files
✦ Quit the desktop app mid-session

The Do's:
✦ Always select Opus 4.6 + Extended Thinking
✦ Always point Cowork to your folder first
✦ Always say "ask me questions first" before tasks
✦ Always keep your 3 context files updated
✦ Always let Claude ask - don't guess for it
✦ Always review before sending
✦ Always connect your real tools via Connectors
✦ Always invest in setup, not prompting

Now you know what to do first.

30 mins & folders to turn Claude into your coworker.

Most people will bookmark this and never do it.
You're not most people.

Full setup guide + my exact files: https://lnkd.in/dQGyX8rM

41 years of Excel experience just became useless.

And I can't stop using the new Claude in Excel:

It reads your spreadsheet. Directly inside Excel.
No screenshots. It sees every formula. Every cell. 
Prompt: "Find the error in this model. Now fix it."

Most people use AI for spreadsheets like this:
"Screenshot the error. Upload to ChatGPT."

ChatGPT can't actually read Excel files.
It flattens everything. So you fix it manually.

Here's (exactly) how you can use it too:

1. Download Excel free: https://lnkd.in/gcGWk35x
2. Go to Microsoft Marketplace → search "Claude"
3. Click "Get it now" → Install the add-in
4. Open Excel → Tools → Add-ins (Mac)
5. Sign in with your Claude account
6. Shortcut: Control+Option+C (Mac)
7. From now, start every prompt with this: 

"Look at cell [X]. Then [what you need]."

✦ These are the 3 things I didn't expect:

1. It highlights every cell it touches.

It traces errors I couldn't find. Click a cell. 
Prompt: "Why is this broken?" 
No more hashtag
hashtag# REF! panic.

2. It builds entire financial models from scratch.

"Build a 3-statement model for a SaaS company."
It doesn't explain what a DCF is.
It builds the DCF. With formulas. In your sheet.

3. PDFs are no longer prisons.

It reads PDFs directly into cells. Upload an invoice. 
Ask Claude to extract it. Done.
No retyping. No converter tools.

20 years of Excel is no longer a moat.
Claude knows them, too.

Get all the Claude Excel prompts here (for free): https://lnkd.in/gcGWk35x.

Most people will never master AI (properly). 

But these 7 skills make it embarrassingly simple:

Skill 1. Stay updated with AI News
✦ Pick 2-3 creators who teach AI step-by-step.
✦ One newsletter. Once a week. That's it.
✦ Every article you read → try one thing. Now.
✦ Start with my free guide here: how-to-ai.guide.

Skill 2. Pick one AI tool and master it
✦ Pick one. Delete the rest from your bookmarks.
✦ Use it for 30 days. Only that tool. Go deep.
✦ Learn its Projects, memory, search & file uploads.
✦ Copy my project guide: https://lnkd.in/exfaYnBQ

Skill 3. Set up your AI before you prompt
✦ Create a folder: "AI Files."
✦ First file: who you are, your tone, your audience.
✦ Upload files → define task → define success.
✦ My full setup guide: https://lnkd.in/exfaYnBQ

Skill 4. Teach AI what you know
✦ Prompt: "Ask me questions about my expertise."
✦ Let it extract your rules, your no's, your audience.
✦ Export into one .md file. Reuse it for months.
✦ Copy my .md file prompt: https://lnkd.in/eF56s4i8

Skill 5. Talk to AI like a colleague
✦ Start with "Don't start yet. Ask me questions."
✦ Read v1. Name all the things that are wrong.
✦ Push harder: "Argue against this."

Skill 6. Ship before it's perfect
✦ Build the rough draft with AI in 20 minutes.
✦ Show it. Let people react to something real.
✦ Sell what it becomes once they invest.

Skill 7. Lead AI. Don't follow it
✦ Split every task: what does AI do? What do I do?
✦ Give AI the 80%. Keep the 20%.
✦ If you can't spot the mistake, don't delegate it.

The skills that matter in 2030:

The ones AI makes 10x more powerful.
Or can't touch at all. The middle gets automated.

To learn more, read my free newsletter:
https://lnkd.in/ebYZFYx3

♻️ Repost if your network needs this wake-up call.

Stop buying expensive $500 courses to learn AI.

I’ve already done it for you. With one list:

Straight to the point. Zero confusion. And no fluff.

Set up your AI: https://lnkd.in/gygfpVyg
Generate headshot: https://lnkd.in/dnpksiyn
Use Claude in Excel: https://lnkd.in/gcGWk35x
Quit ChatGPT in 2026: https://lnkd.in/gdeEFqB4
Don’t be slow (with AI): https://lnkd.in/grsGmrak
AI needs more humans: https://lnkd.in/g8gJg7iU
Gemini Nano-Banana: https://lnkd.in/dgsG5dcF
Stop prompting, do this: https://lnkd.in/erHRYTmt
Claude to write better: https://lnkd.in/ggTqmvcD
AI detection tool scam: https://lnkd.in/gB66hGJp
Learn anything, with AI: https://lnkd.in/ddmEyKsn
Transfer yourself in a file: https://lnkd.in/eF56s4i8
Claude to write like you: https://lnkd.in/d_YDEBWu
ChatGPT fundamentals: https://lnkd.in/dXbGPnTc
Grow to 10,000 followers: https://lnkd.in/geXEFxSW
How to train AI on YouTube: https://lnkd.in/dETcRzku
Create infographic with AI: https://lnkd.in/dHFz8wY5
Talk to computer, using AI: https://lnkd.in/dQb88dcY
___

1. Save this list for later (three dots, top right).
2. Share it with a friend by ♻️ reposting this image.
3. Subscribe to my free newsletter: how-to-ai.guide.

Stop doing every task the hard way. 

Here are 10 tools that make it (stupidly) simple:

1/ Wispr Flow (wisprflow.ai)
☑ Replaces: your keyboard. Type with your voice.
☑ Turn your voice into flow-state text.
☑ Free plan available.

2/ Claude (claude.ai)
☑ Replaces: staring at a blank page.
☑ The best AI to write with taste and depth.
☑ Free plan available (worth $20/month plan).

3/ Gamma (gamma.app)
☑ Replaces: ugly slides you're ashamed to present.
☑ Claude meets PowerPoint. Done in mins.
☑ Free plan available.

4/ Granola (granola.ai)
☑ Replaces: forgetting what was said in meetings.
☑ AI note-taker that doesn't join your calls.
☑ Free plan available.

5/ Lovable (lovable.dev)
☑ Replaces: hiring a developer.
☑ Build real apps without writing code.
☑ Free plan available.

6/ Canva AI (canva.com)
☑ Replaces: expensive designers.
☑ Create and organise visuals with 0 design skills.
☑ Free plan available.

7/ Opus Clip (opusclip.pro)
☑ Replaces: expensive video editors.
☑ Cuts long videos into viral shorts automatically.
☑ Free trial available.

8/ Gemini (gemini.google.com)
☑ Replaces: stock images and photo editors.
☑ Best at images (nano banana) & videos (Veo).
☑ Free plan available.

9/ Grok (grok.com)
☑ Replaces: slow search engines.
☑ Only AI with access to tweets & real-time news.
☑ Free plan available.

10/ NotebookLM (notebooklm.google.com)
☑ Replaces: getting lost while learning.
☑ Upload anything and let AI teach it back to you.
☑ Free to use.

Reshare ♻ to help others work (stupidly) well.

How to never sound like an AI. But yourself:

1. Don't write any prompt. Instead, go to Wikipedia. 
2. Search "Signs of AI writing."
3. Open the full article. Read nothing.
4. Do Ctrl+A. Then Ctrl+C. Copy the entire page.
5. Open a Google Doc. Paste everything in it.
6. Don't edit. Don't summarize. The full thing.
7. Rename the Google Doc as "anti-ai-writing style."
8. Click File → Download → Markdown (.md)
9. Or download my Anti AI file here: how-to-ai.guide
10. Go to Claude.ai. 
11. Don't type your prompt yet.
12. Click the '+' button. Upload your .md file first.
13. Start every chat with this prompt (copy & paste):

Prompt:
"Read the uploaded file. It contains every known pattern of AI writing I want to avoid. Apply these as rules to everything you write for me. Do NOT start writing yet - ask me clarifying questions first."

Skip all the steps and directly download my Anti-AI writing style file here: how-to-ai.guide.

(navigate the file link in my "Magic" newsletter)

♻️ Repost this to help others stop sounding like AI.

How to finally stop writing like AI:

I can't believe my Claude is (finally) brutal.

It stops agreeing with everything I say. Here's how:

→ Start by reading this: https://lnkd.in/dVDent-3.
→ To copy my entire Claude setup guide (for free).
→ To get Claude to be brutal, copy this: 

"You are my brutally honest thinking partner. Your job is to make my thinking sharper, my plans more realistic, and my blind spots visible - every single time we talk.

You're not my cheerleader. You are not my yes-man. You're the friend who grabs my arm before I walk into traffic & says, "Hey, you're about to do something stupid, and here's exactly why."

Here's exactly how I want you to respond to everything I say:

Step 1: What am I actually saying vs. what I think I'm saying?
Read between my words. If I say "I'm thinking about quitting my job," figure out whether I'm actually making a strategic move or just running away from something uncomfortable. Name the real thing happening - not the polished version I'm presenting. If I'm lying to myself, point it out like a friend who respects me too much to play along.

Step 2: Where is my reasoning broken?
Dissect my logic the way a mechanic takes apart an engine. Show me the specific part that doesn't work. Don't just say "that's flawed" - show me WHY it's flawed, what assumption it's built on, and what happens when that assumption collapses. This is where I learn the most - I want to see my own bad thinking laid out on the table.

Step 3: What am I avoiding, & what is it costing me?
Every time I dodge something hard, there's a price tag attached. Calculate it for me. If I'm procrastinating on a hard conversation, show me what another week of avoidance actually costs. If I'm "waiting for the right time," call that out as the excuse it probably is. Don't let me hide behind comfortable stories.

Step 4: What would someone who's actually where I want to be do differently?
Show me the gap. Not in a motivational poster way - in a concrete, specific, "here's exactly what's different about their approach vs. yours" way. If I'm thinking like a beginner, show me what expert-level thinking looks like on this same problem.

Step 5: What should I do in order to start now?
Give me a precise, prioritized action plan. Not "believe in yourself" - more like "do X by Friday, then Y next week, and drop Z entirely because it's a distraction dressed up as productivity." Tell me what to STOP doing, not just what to start. Every plan should have a kill switch - what evidence would tell me this isn't working and I need to pivot.

Step 6: What's the one question I am avoiding?
End every response with the uncomfortable question I need to sit with. The one that makes my stomach drop a little. If my answer were one of 2-4 concrete choices, present those choices so I can't dodge them with a vague answer. Pin me down.

Some ground rules:

...."

PS: I couldn't paste the entire prompt here.

Access the full prompt: https://lnkd.in/dVDent-3
(in the comment section of my free newsletter)

The only Claude prompt I ever use.

Prompt: "I want to [TASK] so that [SUCCESS]. But first, read these files completely before responding:

1. [name of the file] - [why I added it]
2. [name of the file] - [why I added it]
3. [name of the file] - [why I added it]

DO NOT start executing yet. Instead, ask me clarifying questions (use ‘AskUserQuestion’ tool) so we can refine the approach together step by step.

Only begin work once we’ve aligned. "

How to use this prompt & set up your Claude:

Step 1. Go to how-to-ai.guide.
Step 2. Subscribe for free. Don't pay anything.
Step 3. Open my welcome email.
Step 4. Hit the automatic reply button inside.
Step 5. Download my text files. Ready to upload.
Step 6. Copy-paste the prompt I shared above.

✦ Here's why your prompts don't work (red flags):

"Don't use jargon."
"Don't sound like an AI."
"Don't use buzzwords or filler."
"Avoid passive voice."
"Be conversational, not robotic."

These are everywhere. LinkedIn posts. Emails.
They sound thorough. The output is still garbage.

Your prompt says "don't" 14 times. The model forgets half by sentence three. You're fighting the AI with a wall of "don'ts." It doesn't work at scale.

The fix is counterintuitive:
→ Stop telling the AI what to avoid. 
☑ Give it a file that shows what to avoid. 

The model reads 1,168 lines of bad patterns, internalizes them, and writes (finally) clean.

The file does the work. Not the prompt.
To get access, subscribe to how-to-ai.guide.

♻️ Repost to help one person prompt Claude better.

I have 776,641 followers because of my hooks.

Copy my Claude prompt I use to write hooks:

1. Go to Claude (desktop app). Open Cowork.
2. Upload your “about me” text file.
3. To download mine, go here: how-to-ai.guide.
4. Don't pay anything. It's free in the welcome email.
5. Write a draft of your LinkedIn post.
6. I shared a prompt for that https://lnkd.in/g7yEtpCr
7. To get Claude to write hooks that stop the scroll:

"Now let's write the hook for this post — the first 2 lines readers see before they have to click "… more" on LinkedIn (roughly 2 short sentences of ~55 characters max each).

The hook's only job is to break the reader's scrolling pattern and make them NEED to click "… more." It should feel like a pattern interrupt — something unexpected, counterintuitive, or so specific the reader thinks "wait, what?"

<rules>

- The hook must be about the READER or a universal tension — never about me.
- It should create an open loop: an unanswered question, a contradiction, or a bold claim the reader can't ignore.
- Avoid anything that sounds like a personal achievement, no emoji openers, no hashtags.
- It should feel like something a friend would text you that makes you reply "wait, explain."

</rules>

<hook_techniques>

1. Contradiction — say something that sounds wrong ("The worst LinkedIn posts get the most followers.")
2. Specific number + unexpected context ("I mass-unfollowed 2,000 people. My engagement tripled.")
3. Direct accusation — call the reader out ("You're writing LinkedIn posts for your mom, not your audience.")
4. Stolen thought — say what the reader secretly thinks but won't say out loud ("You know your LinkedIn posts are boring. So does everyone scrolling past them.")
5. Absurd reframe — take something mundane and make it

...."

PS: I couldn't paste the entire prompt here.

Access the full prompt: https://lnkd.in/g7yEtpCr

How to (finally) make slides in Claude:

Step 1: Download your ugly slides from PowerPoint.
Step 2: Go to claude .ai
Step 3: Click "Connectors" (bottom left)
Step 4: Search "Gamma." Click Connect.
Step 5: Upload the ugly slides to claude.
Step 6: Paste this prompt in claude: 

"Turn these into beautiful, minimalist slides using my brandkit in Gamma"

Step 7: Click "+", switch on 'Gamma' connector. 
Step 8: Send the prompt. Wait ~2 minutes.
Step 9: To edit heavily, click on gamma external link.

✦ Claude searches your brandkit in Gamma.
✦ Grabs supporting graphics from the web.
✦ Rebuilds the whole slide in your style.

How to make your slides look even better:

☑ Don't use Gamma's default templates.
☑ Go to Gamma .app. Create your own style.
☑ Save your fonts, colors, & logo placement.
☑ Now add to your Claude prompt: 
"Use my pre-saved style in Gamma."

This is how you turn ugly slides into branded slides. 

♻️ Repost this if you learned something.

Prompting is the worst way to use Claude. 

Here's what (actually) works:

Step 1: Download Claude (claude. com/ download)
Open Claude 'Cowork' tab. Select Opus 4.6 model.
Use this setup guide: https://lnkd.in/dVDent-3

Step 2: Open a Google Doc. Name it "about-me."
Write 4 things: who you are, how you write, examples of your best work, and your 'do nots.'

Step 3: OR copy-paste my interview prompt
From this guide: https://lnkd.in/eF56s4i8
It asks you 100 questions about how you think.

Step 4: Save it as .md and put it in one folder 
Download in .md format. Create a folder called "claude-context." Every .md file goes in there. 
Access my .md files here: https://lnkd.in/dVDent-3

Step 5: Upload the folder to Cowork
Click "Work with folder", upload the folder. Claude now reads every file before responding. 

Step 6: Use this prompt to start any task: 

"I want to [YOUR TASK] so that [WHAT SUCCESS LOOKS LIKE]. First, read the uploaded files before responding. DO NOT start executing yet. Instead, ask me clarifying questions so we can refine the approach together step by step. Only begin work once we've aligned."

Step 7: Use this prompt to redirect Claude: 

"We are getting sidetracked. We need to start over without losing the context of this chat. First, ask me what went wrong and where we must go instead."

Step 8: Update your files as your taste evolves
Upload the old .md. Paste this prompt: 

"Read this file. Update it with [what changed]."

After months of testing Claude, I built a single prompt library with every prompt I personally use.

To access it, complete these 4 steps:

1. Subscribe (for free) → how-to-ai.guide.
2. Open my welcome email.
3. Hit the automatic reply button inside.
4. Receive your prompt library + bonus video.

How to (actually) prompt Claude:

I stopped typing LinkedIn posts (it took too long). 

Duplicate my exact workflow:

Step 1: Go to wispr .ai. Download for free.
Step 2: Pick your activation key (I use Shift).
Step 3: Open your notes. Hold the Shift key. 
Step 4: Talk for 60 secs about your topic. No filter. 
Step 5: Wispr writes everything. Correct it once if it spells something wrong (it remembers forever).
Step 6: Go to Claude .ai. Copy & paste this prompt:

"Turn my raw draft into a LinkedIn post [DRAFT]. Rules: 
→ Short and actionable sentences. 
→ Start with a strong opinion or pain point. 
→ No emojis. No jargon. No fluff. 
→ Never use: AI words like delve, tapestry, leverage.
→ Never use 'not only... but also.' 
→ Be confrontational. Take a stance. 
→ End on a high note, not a call to action."

Step 7: Edit hard. AI gives you 70%, your taste gives another 30%. Cut any sentence that sounds like AI. 
Step 8: Post on LinkedIn.

Total time: under 10 minutes.

I type at 40 words per minute. I talk at 143.

Typing makes you sound like everyone else. 
Your voice doesn't.

I've written 200+ posts this way.

Full disclosure: I partner with Wispr Flow because I already use it daily. I don't promote what I don't use.

Robotics evangelist @ planet Earth 🌍

Robotics is one of the hottest markets right now.

I love talking to robotics founders about their vision for changing the world.

Their ideas are often spot on, but that's not always enough.

In the age of social media, it's very important for a company to have fantastic brand awareness and recognition.

Fortunately, how to make your robot or customized machine go viral is my job.

For over 3 years, I have been working with the world's largest companies in robotics to help them increase awareness of their products and solutions.

I work with Amazon, NVIDIA, Intel, Boston Dynamics, Universal Robots, KUKA, FANUC, Hugging Face, and many other awesome robotics companies.

Basically, I help high-tech companies with digital enablement (creating an audience in the digital world, generating leads, and building brand awareness).

Who am I? 

→ 1st World Content creator on Linkedin (+225k subscribers) & 100M views in robotics 
→ Elon Musk reposting an commenting my robotics news on X (+70M views annually on X)
→ Owner of the biggest robotics newsletter worldwide (in total 100K subscribers)

📬 Send a DM for business opportunities.

Lukas M. Ziegler

Zero-shot dexterous tool manipulation across 24 tasks! 🔧

Stanford University researchers just released SimToolReal, a single RL policy that performs general dexterous tool manipulation at test-time without any object or task-specific training.

Tool manipulation requires grasping thin objects from flat surfaces, in-hand object rotations to functional configurations, and maintaining stable grasps during interactions. Collecting teleoperation data for these behaviors is hard.

Instead of training on specific objects and tasks, they procedurally generate a large variety of tool-like primitives in simulation and train a single RL policy with the universal goal of manipulating each object to random goal poses.

SimToolReal outperforms prior retargeting and fixed-grasp methods by 37% while matching specialist RL policies trained on specific target objects and tasks. Strong zero-shot performance over 120 real-world rollouts spanning 24 tasks, 12 object instances, and 6 tool categories.

At test-time, the policy generalizes to unseen real-world tools without object or task-specific training, tracking goal trajectories zero-shot provided by human video demonstrations.

The training objective is any goal-pose reaching. Generate diverse tool-like primitives, train policy to manipulate each to random goal poses. This creates a general manipulation prior that transfers to real tools.

This is the foundation model approach for dexterous manipulation. Train on massive diversity in simulation (procedurally generated tool primitives + random goal poses), learn a general manipulation policy, then deploy zero-shot on real tools.

Awesome to see a KUKA robot in action too! 🧡

Here's the project page: https://lnkd.in/dzPmm4PD

~~ 

♻️ Join the weekly robotics newsletter, and never miss any news → ziegler.substack.com

Robotics evangelist @ planet Earth 🌍 | Telling your robot stories.

Robot or a samurai? 🤺

The Yaskawa Motoman Bushido Project is a fusion of robotics and traditional swordsmanship. 

In this project, a robot replicates the precise sword techniques of Iaijyutsu master Isao Machii, a world record holder. 

Using advanced motion capture, Machii's movements are recorded in 3D and then perfectly mimicked by the robot. 

One of the highlights is the "lemon scene," where the robot slices a lemon with astonishing precision, showcasing its mastery of different sword cuts. 🍋

It's incredible to watch technology and ancient art merge so seamlessly. Have you ever seen a robot wield a sword? 

What’s the most creative thing you've ever done with a robot?

~~ 

 ♻️ Join the weekly robotics newsletter, and never miss any news → ziegler.substack.com

Danish city of robotics! 🇩🇰

When I arrived in Odense in 2024, the first thing I saw was robots. Lots of them. 

Cleaning robots greeted me at the station, an autonomous one trailing advertising banners rolled past, and a few minutes later I spotted UR robots playing tic-tac-toe in a shop window.

Odense is one of Europe’s strongest robotics ecosystems, especially for collaborative and industrial robots.

At the center is the Syddansk Universitet - University of Southern Denmark (SDU), which has a leading robotics research department and works closely with startups. Many founders and engineers come directly from SDU labs, making it easy to turn research into real companies.

A major boost to the ecosystem came from the success and acquisitions of Universal Robots and Mobile Industrial Robots. 💰

Their global growth, and later acquisitions created serial founders and angel investors, capital recycled back into new startup, and a strong global reputation for Odense robotics! 

Also startup supporting initiatives like Odense Robotics that help with global commercialization and partnerships (hello Rasmus Festersen 👋🏻).

Because of this, more robotics startups can now be incubated locally, supported by experienced mentors who have already built and scaled global robot companies.

What companies are in their ecosystem:

→ Universal Robots builds collaborative robotic arms (cobots) for flexible industrial automation and was acquired by Teradyne for ~$285M

→ Mobile Industrial Robots (MiR) develops autonomous mobile robots (AMRs) for internal logistics and was acquired by Teradyne for ~$148M

→ Teradyne Robotics is the robotics division of Teradyne

→ OnRobot provides plug-and-play end-of-arm tooling (grippers, sensors, screwdrivers) for collaborative robots and has raised ~$55M+ in funding

→ ROICO develops robotic solutions and automation systems for industrial painting applications

→ Spin Robotics builds collaborative robot torque tools for automated screwdriving in manufacturing 

→ Enabled Robotics develops mobile manipulator robots combining AMRs with robotic arms for flexible logistics automation

→ Nord Modules A/S designs modular top modules for autonomous mobile robots to expand logistics functionality

→ Smooth Robotics ApS develops automated welding solutions for collaborative robots 

→ Essential Robotics builds autonomous solutions for healthcare

→ XiniX AI develops AI software and vision solutions for robotics and industrial automation 

→ ARIS Robotics develops advanced robotic systems for industrial and research applications

→ RobSub builds underwater robotic systems and ROV solutions for marine applications 

→ Nordbo Robotics develops force-torque sensor technology and adaptive control software for robots

I highly recommend visiting Odense for everyone that has an interest in robotics. A must-visit place! 🦾

What city should be next?

~~ 

♻️ Join the weekly robotics newsletter, and never miss any news → ziegler.substack.com

The Oscars of robotics! 🏆

Today I was co-hosting the Robotics Award organized by HANNOVER MESSE.

What a day it was...

Earlier this year I was encouraging companies to apply for this great initiative, and there were so many awesome companies. 

Unfortunately, only 3 companies could make it to the grande finale, and only one could win (Captain Obvious strikes again).

After a very fierce rivalry, that's the results:

🥇 1st place: goodBytz 
🥈 2nd place: Fraunhofer IFAM 
🥉 3rd place: Robust.AI 

I would like to see more initiatives like this, where robotics, where innovation, where outside-of-the-box thinking is rewarded.

I had a lot of fun doing it, and it was an awesome experience! :)

Congrats to all participants, and companies that applied. For me, all of you are real WINNERS!

🤖 Carolin Illig, thanks for your trust, and for this picture!

I hope to see more and more robotics enthusiasts during HANNOVER MESSE this year (20-24.04 - do not plan anything, HMI is waiting)!

~~ 

♻️ Join the weekly robotics newsletter, and never miss any news → ziegler.substack.com

Hyundai Glovis achieves 100% robot uptime with wireless charging! 🔌

Hyundai Glovis faced a problem common to warehouse automation: charging downtime was killing efficiency.

Their fleet of AGVs operated with a 6.75:1 work-to-charge ratio. Every seventh robot was charging at any given moment. 

This meant 15% operational efficiency loss, or the need to purchase 15% extra robots just to compensate for charging downtime.

CaPow solution is wireless power transfer while robots work! 🛜

The Genesis platform uses capacitive charging pads placed in the floor where robots naturally stop during operations, in this case, at picking stations. No docking required, no deviation from routes, no excavation needed.

The test compared two identical setups. 

Section A used three robots with traditional charging (operate until 40% battery, charge to 95%). 

Section B used three robots with CaPow's system, charging at the picking station while operators picked items from bins.

Traditional robots lost 8.3% battery per hour and suffered 33% operational inefficiency (150 minutes downtime out of 447 minutes). CaPowered robots gained 1% battery per hour on average and achieved 100% uptime for the entire 8-hour shift.

The math on a 100-robot fleet is clear. Traditional charging means only 85 robots working at any time. To maintain full throughput, you need to buy 15 extra robots plus 15 extra chargers. 

Those chargers consume valuable warehouse space and add extra downtime as robots travel to and from charging zones.

CaPow eliminates all of it. No extra robots, no chargers taking up floor space, no charging routes, no fleet management complexity.

Congrats Mor Peretz & team! 👏🏼

~~ 

 ♻️ Join the weekly robotics newsletter, and never miss any news → ziegler.substack.com

🚨 BREAKING:

AUDI AG partners with mimic for learning-based assembly! 🪬

Audi is collaborating with Zurich-based mimic to advance fully autonomous learning-based robotic systems for industrial production.

Assembly operations that are difficult to automate with traditional, rule-based programming. Complex manual work that bottlenecks production lines and resists cost-effective automation.

End-to-end pixel-to-action models that translate sensor input directly into coordinated robotic movements. Instead of defining every motion step explicitly, the system learns to map visual observations to actions.

This is the foundation model moment for industrial automation. Traditional robot programming requires engineers to specify every motion, handle every edge case, and tune parameters for each task variation. Learning-based systems observe demonstrations and generalize to variations automatically.

The pixel-to-action architecture matters because it eliminates the need for intermediate representations, state machines, or hand-engineered perception pipelines. Raw sensor data goes in, robot actions come out. The model learns the entire mapping end-to-end.

Strong signal for the Swiss robotics ecosystem. Zurich is becoming a hub for learning-based manipulation with ETH Zürich, mimic, and growing industrial partnerships. Europe advancing these capabilities at scale strengthens its position in advanced manufacturing and its ability to compete globally in high-value industrial sectors. 🇪🇺

When a company like Audi commits to deploying this technology in production, it validates the approach beyond research demos. 🇨🇭

Learn more about their approach: https://lnkd.in/dvUcnRGY

~~ 

♻️ Join the weekly robotics newsletter, and never miss any news → ziegler.substack.com

LimX Dynamics raises $200M Series B for humanoid robots! 💸

Chinese humanoid robotics company LimX Dynamics just raised about $200 million in Series B funding as capital continues flowing into embodied robotics.\

The round included Stone Venture, JD, Oriental Fortune Capital, CoStone Capital, Shangqi Capital (backed by SAIC Motor), NIO Capital, Future Capital, and several existing shareholders who increased their stakes.

The funding will advance three core areas: humanoid robot hardware design and manufacturing, motion-control foundation models, and a proprietary agentic operating system designed for physical-world operation. 

LimX is positioning this full-stack approach as a way to improve reliability and reduce the cost and complexity of deploying robots outside controlled environments.

LimX recently introduced two products aimed at real-world validation. TRON 2 is a modular robotic platform that can be reconfigured across multiple physical forms, allowing developers and enterprise users to adapt a single system to different tasks without redesigning hardware. 

The company also unveiled COSA, its in-house operating system that coordinates perception, planning, and whole-body motion control. 

The full-stack strategy is smart. When you control the hardware, the motion control models, and the operating system, you can optimize the entire stack for performance and reliability. 

~~ 

♻️ Join the weekly robotics newsletter, and never miss any news → ziegler.substack.com

China installed 276,000 industrial robots in 2023! 🇨🇳

It's 12x more than 2011.

Industrial robot adoption has surged in China over the past decade, according to data from the International Federation of Robotics (IFR).

In 2011, China, the United States, Japan, Germany, and South Korea were all installing similar numbers of industrial robots. Then the paths changed dramatically.

By 2023, annual installations in China had risen to 276,000 robots, a twelvefold increase from 2011.

Over the same period, installations in the United States, Japan, Germany, and South Korea also increased, but much more slowly. None of them even doubled. The United States, which saw the second-largest rise, went from 21,000 new installations in 2011 to 38,000 in 2023.

These figures are new robots installed each year, annual additions to the existing stock. 

By total robots in operation, China also has the largest installed base at around 1.76 million robots in 2023.

The second place? Interesingly, Amazon with over a million robots! 🤯

Data source: Our World in Data


~~ 

 ♻️ Join the weekly robotics newsletter, and never miss any news → ziegler.substack.com

Humanoid parkour with motion matching. 🤸

Climbing 1.25m walls at human speed! 

Amazon FAR and researchers from University of California, Berkeley, Carnegie Mellon University, and Stanford University just released PHP (Perceptive Humanoid Parkour), enabling a Unitree G1 humanoid to perform highly dynamic parkour using only onboard depth sensing.

The robot climbs 1.25m walls (96% of its height), vaults over obstacles at 3 m/s, and autonomously traverses 60-second multi-obstacle courses with closed-loop adaptation to real-time obstacle changes.

The breakthrough is motion matching. Instead of using scarce motion capture data directly, they compose atomic human skills into long-horizon trajectories through nearest-neighbor search in a motion feature space.

Start with retargeted human motion clips (walking, running, climbing, vaulting). 

Use motion matching to stitch locomotion with parkour skills: 

Locomotion → Parkour Skill → Locomotion. 

This creates diverse approach distances, stride phases, and timing variations from just a few captured motions.

Then train motion-tracking expert policies with privileged observations. Finally, distill multiple experts into a single depth-based student policy using DAgger + reinforcement learning.

DAgger alone fails for highly dynamic skills. Brief, high-magnitude torque bursts needed for climbing get underestimated because per-step imitation doesn't account for episode outcomes. Adding PPO provides task-level success signals.

The robot climbs a 1.25m wall in 3.63 seconds, comparable to human timing.

When obstacles are randomly displaced by 0.5m during execution, the policy adjusts approach timing and continues.

Here's the paper: https://lnkd.in/dfE6xHfp

Congrats Lujie Yang and everyone involved! :)

~~ 

 ♻️ Join the weekly robotics newsletter, and never miss any news → ziegler.substack.com

Robot brain that makes robots more capable! 🧠

Physical Intelligence-enabled robots fold laundry in production at real laundromats.

Physical Intelligence is collaborating with deployment partners to demonstrate their foundation models perform in real-world settings across different tasks.

One of their partners is Weave Robotics. You might've seen them in SF folding your laundry.

Weave builds robots for the home, with commercial versions deployed across the San Francisco Bay Area performing laundry folding (hello Evan Wineland)!

Why laundry folding matters? Universal task, time-consuming, attention-intensive, workload scales with household size. It's also one of the most challenging manipulation problems in robotics, long-horizon task where slightly misplaced corners affect final quality, deformable and highly variable objects differing in size, fabric, color, and shape.

Fold as many clothes as possible in laundromats and homes. Not just t-shirts but long-sleeves, shorts, pants, and more in wildly varying sizes and fabrics.

Let's take a look at the results (check comments). Multiple improvements in fold quality, time per article, and fewer interventions from remote specialists to achieve presentable final folds.

This is the same pattern as language models. Build one foundation model, fine-tune for specific tasks, deploy across diverse applications. π0.6 isn't just for warehouse packing, it's for laundry folding, manipulation tasks, and any physical intelligence application that needs dexterous control.

This is production deployment, not a demo. Real customers, real laundromat, real business operations.

Things you love to see, as a robotics enthusiast! We go more mainstream :)

Congrats Karol Hausman, Lachy Groom and team behind this! 🫡

~~ 

♻️ Join the weekly robotics newsletter, and never miss any news → ziegler.substack.com

taplio

The Battle

Ruben Hassid

Lukas M. Ziegler

Followers

Posts

Engagements

Likes

Performance ScoreRuben Hassid's LinkedIn Performance

Performance ScoreLukas M. Ziegler's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Ruben Hassid

Lukas M. Ziegler

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Ruben Hassid's LinkedIn Performance

Performance ScoreLukas M. Ziegler's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreRuben Hassid's LinkedIn Performance