Alexandre Kantjas's Linkedin Analytics

New to Airtable? Go from zero to hero in this 25 𝘮𝘪𝘯𝘶𝘵𝘦𝘴 𝘧𝘳𝘦𝘦 𝘤𝘳𝘢𝘴𝘩 𝘤𝘰𝘶𝘳𝘴𝘦: Airtable is a powerful no-code platform that can automate your workflows and replace entire tools. The problem with it: 𝘮𝘰𝘴𝘵 𝘱𝘦𝘰𝘱𝘭𝘦 𝘩𝘢𝘷𝘦 𝘯𝘰 𝘪𝘥𝘦𝘢 𝘩𝘰𝘸 𝘵𝘰 𝘶𝘴𝘦 𝘪𝘵 effectively or 𝘸𝘩𝘢𝘵 𝘵𝘰 𝘶𝘴𝘦 𝘪𝘵 𝘧𝘰𝘳. So I made this free Airtable crash course - and made sure to keep it under the 30 min mark to make sure it's worth your time. If you are a complete beginner, this video takes you from zero to confidence so you can start using it like the pros. Work will never look the same once you're on the other side. Want the course? Let me know in the comments and I will share it with you.

While everyone is « making » Ghibli art, Hayao Miyazaki must be shaking his head in disappointment. I’m torn between how amazing AI capabilities are getting - and wondering how it must feel to a creator seeing your life’s work being so easily adapted by machines. I don’t have the answer but it does feel wrong for creators and we will have to find solutions for them. Without creators, there is no magic. No spark. No raw, messy, human chaos that makes art actually mean something. Creation needs compensation, acknowledgement, respect.

Claude is finally getting web browsing. If you have never tried Claude, now‘s the time. Hands down my favorite model.

Join us for an exciting nocodeklub Munich meetup 🍻 🥨 We're thrilled to welcome Maximilian Mayr, Head of Operations at Peter Park, as our next speaker. Maximilian will share insights on how Peter Park is revolutionizing the parking industry with innovative solutions and operational excellence. Here's the details: 📍 Location: Make Celonis Office 📅 Date: Tuesday, 18th March 2025 - UTC Don't miss out on this opportunity to learn and network! Link to register in the comments 👇

Make + nocodeklub + Munich =

ChatGPT didn't "kill" marketers and designers last week. Why? Because 𝘈𝘐 𝘤𝘢𝘯𝘯𝘰𝘵 𝘮𝘢𝘯𝘶𝘧𝘢𝘤𝘵𝘶𝘳𝘦 𝘫𝘶𝘥𝘨𝘮𝘦𝘯𝘵 (yet) Don't get me wrong: I'm very excited about vibe marketing, vibe designing and the next vibe for "x" trends to come. Yes, AI dramatically lowers the capability barrier - anyone can generate something decent quickly: 📺 Landing pages 🌠 Visuals 📢 Ads And the results people are able to generate with ChatGPT 4o are pretty decent. But AI doesn't give you judgment... Judgment is knowing what's good, what's not, and why. It's the taste, discernment, and decision-making that comes from experience. Without judgment, you can't tell if what the AI made is actually good. Without judgment, you can't direct the AI effectively. The best professionals in any field have both capabilities and judgment - which they've developed through practice and experience. This is why a seasoned marketer using AI will still outperform a novice using the same tool. The pro has the judgment to know what to ask for, recognize quality, iterate effectively, and understand context. In summary: 1. Skills = capabilities + judgment. 2. “Good judgement comes from experience, which comes from bad judgement” - Bill Gurley So learn AI - but don't forget to master your craft, whatever that it is.

Build an AI agent live with us in 90 minutes ⏳ Don't miss next week's 9𝘹 𝘓𝘪𝘷𝘦 𝘸𝘰𝘳𝘬𝘴𝘩𝘰𝘱 - where Jan will showcase how we built our most beloved AI agent - a meeting notetaker that made our meeting life 9x easier 😌 Here are the details: --- 𝐁𝐮𝐢𝐥𝐝 𝐚𝐧 𝐀𝐈 𝐀𝐠𝐞𝐧𝐭 𝐥𝐢𝐯𝐞 𝐰𝐢𝐭𝐡 𝐮𝐬 👤 𝐇𝐨𝐬𝐭𝐞𝐝 𝐛𝐲: Jan Meinecke, Co-founder, 9x 🕒 𝐖𝐡𝐞𝐧: Tuesday, March 18, 2025, 🇪🇺 5:00pm CET - 🇺🇸 12:00pm EDT 📍 𝐖𝐡𝐞𝐫𝐞: Online You'll learn how to: → Design an AI agent that records and summarizes meeting information → Connect your agent to meeting platforms and Slack → Configure your agent to respond to follow-up questions based on transcript → Implement advanced features like email drafting based on meeting content If you'd like to attend, let me know, and I'll send you the information.

Join our next 9x Live workshop, get 30.000 credits to play with Airtop and ask your questions directly to their CEO. Browser agents are the next frontier in AI automation - able to do anything humans do on the web → Browse pages → View page content → Scroll down pages → Log in accounts → Click buttons → Type text → Extract data Airtop is a great platform to build browser agents - and tomorrow their CEO Amir Ashkenazi will demo in person its capabilities. So you get the best of all worlds: · Discover Airtop? Check ✅ · Ask any question about the platform? Check ✅ · Get free credits to play with the platform after the event? Check ✅ All you need to do is to 𝘢𝘵𝘵𝘦𝘯𝘥 𝘵𝘩𝘦 𝘦𝘷𝘦𝘯𝘵 tomorrow. I'll paste the link to register once again 👇 See you there!

Looks like OpenAI's Sora is now accessible in Europe too. Looks pretty cool - will play with it a bit. Kudos to Europe for helping us getting yesterday's innovation tomorrow 🇪🇺

𝘉𝘳𝘰𝘸𝘴𝘦𝘳 𝘢𝘨𝘦𝘯𝘵𝘴 will soon be able to do 𝘢𝘯𝘺𝘵𝘩𝘪𝘯𝘨 𝘩𝘶𝘮𝘢𝘯𝘴 𝘥𝘰 on a web browser. Learn how in next week's 𝘸𝘰𝘳𝘬𝘴𝘩𝘰𝘱: Browser agents are the next hot trend in AI (next to vibe coding obviously). They can: → Navigate to a new page → View page content → Scroll down the page → Log in the account → Click a specified button → Enter text in a textbox → Extract data Why this is exciting: for the first time, a technology will allow us to automate 𝘢𝘭𝘭 𝘵𝘩𝘦 𝘵𝘩𝘪𝘯𝘨𝘴 𝘵𝘩𝘢𝘵 𝘤𝘢𝘯𝘯𝘰𝘵 𝘣𝘦 𝘢𝘶𝘵𝘰𝘮𝘢𝘵𝘦𝘥 𝘶𝘴𝘪𝘯𝘨 𝘈𝘗𝘐𝘴. Massive! And next week's workshop will give you the opportunity to see them in action: 𝐁𝐮𝐢𝐥𝐝 𝐁𝐫𝐨𝐰𝐬𝐞𝐫 𝐀𝐠𝐞𝐧𝐭𝐬 𝐰𝐢𝐭𝐡 𝐀𝐢𝐫𝐭𝐨𝐩 👤 𝐆𝐮𝐞𝐬𝐭: Amir Ashkenazi, Founder and CEO, Airtop 🕒 𝐓𝐢𝐦𝐞: Thursday, March 13th 5:00pm CET 📍 𝐋𝐨𝐜𝐚𝐭𝐢𝐨𝐧: Online Pierre-Yves will be hosting this session, and will explore with Amir how browser agents are the next frontier in web automation - and how they're about to transform the way you interact with technology. I highly encourage you to join this session - reply to this post if you'd like to attend!

Build an AI agent that researches new leads and updates your CRM - in 60 minutes - live with us: Here's your chance to learn hands-on with Jacob Bank, Founder & CEO of Relay.app In this free 60-minute workshop, Jan Meinecke and Jacob will show you how to create your own AI Sales Research Agent using Relay.app - no coding needed. Jacob will guide you step-by-step through the entire process, making it immediately applicable to your workflow. Sounds interesting? Let me know and I'll share the workshop details.

1830 people registered to our 9𝘹 𝘭𝘪𝘷𝘦 𝘸𝘦𝘣𝘪𝘯𝘢𝘳𝘴 in March vs. 206 in February - a 800% month-on-month progression (9x 🙃). So, we crushed it? Not really: First: generating leads is hard. We're all fighting to get each other's attention and the bar is getting higher every quarter. For a long time 9x has been struggling with awareness - so towards the end of last year creating content became our number one priority. Seeing the traction we got in March for our live webinars shows our efforts are starting to pay off. A growing audience is following the content we share here on LinkedIn, checking our step-by-step tutorials, attending our live events and watching our YouTube channel - and if anything we will continue doubling down our efforts. But the other side of this story is not so flattering... conversions for our course have been really low in March - with absolute results comparable to February. Translation: we got nine times the amount of webinar sign-ups but a conversion rate that dropped by a factor of nine over the same period 💀 Question: where does the problem come from? A few probable reasons come to mind: 🔄 𝐃𝐨𝐮𝐛𝐥𝐞 𝐜𝐨𝐮𝐧𝐭𝐢𝐧𝐠 leads - many people attended multiple workshops 💰 Course 𝐩𝐫𝐢𝐜𝐞 might be too high for some of the leads we attracted ⬇️ 𝐋𝐞𝐚𝐝 𝐪𝐮𝐚𝐥𝐢𝐭𝐲 dropped or lead fit for our courses went down ✉️ 𝐍𝐨 𝐧𝐮𝐫𝐭𝐮𝐫𝐢𝐧𝐠 𝐩𝐫𝐨𝐜𝐞𝐬s in place with people who attended ⏳ 𝐋𝐞𝐚𝐝𝐬 𝐧𝐞𝐞𝐝 𝐦𝐨𝐫𝐞 𝐭𝐢𝐦𝐞 or touch points to convert 🥶 ...LinkedIn may be the 𝐰𝐫𝐨𝐧𝐠 𝐜𝐡𝐚𝐧𝐧𝐞𝐥 Many people recently reached out mentioning we must be "killing it on LinkedIn" and swimming in qualified leads - this post is to warn you that getting impressions is only one side of the equation. Anyways... Our next steps to get a clearer picture and answer this question: → Run a thorough analysis of our past conversions → Enrich and score the leads we generated in March → Build up our presence outside LinkedIn in Q2: YouTube, X, and more... Am I missing anything? Drop me your suggestions in the comments. --- P.S: will keep you posted of the progress we make in future posts P.P.S: our course price is currently €999 and provides lifetime access to all the course content we've made so far (100+hours) + upcoming content, Slack community and all live events. Dashboard made with Attio #buildinpublic

Dear vibe coders: because AI generates your app's code for you doesn't mean you can skip understanding it. Many posts around vibe coding with Lovable, Cursor or Replit. Yes, this tech is mind-blowing. Yes, the future of coding is probably AI generating 99% of the code. But: this certainly doesn't mean you can skip understanding what the code does and how it works. Not understanding your code is like building a house without knowing what the foundation is made of: it might look stable for a while, but one gust of wind (aka a bug, scaling issue, or security threat) and the whole thing crumbles. Many vibe coders are non-coders who do not understand what’s happening under the hood. To them, fast results feel great. But when things inevitably break (and they will), they will be left staring at a wall of errors with no clue how to fix it. So next to vibe coding...start understanding: ✅ Know what your code is doing. ✅ Learn the core concepts — data structures, APIs, and architecture. ✅ Ask why something works, not just if it works. When things inevitably go wrong - and they will - the vibe won't save you. Knowledge will.

130.000+ people tuned on X to watch a 𝘏𝘶𝘮𝘢𝘯 𝘷𝘴. 𝘈𝘐 𝘴𝘩𝘰𝘸𝘥𝘰𝘸𝘯 in landing page design. Can you guess who won? More specifically: a marketer with little to no design experience challenged a top designer to a landing page speed build contest. Why this challenge got so much attention? We'd see for the first time whether a person with no design background or experience using AI can outplay a seasoned designer. The contenders: Henrik Westerlund from Lovable: - No design experience - zero money made building websites - 4 months of using Lovable Brett Williams from Designjoy: - More than 10 years of professional design experience - $145K/month building websites - 5+ years using Webflow Brett won the challenge overall - but it wasn't an easy challenge for him: "𝘐 𝘥𝘰𝘯’𝘵 𝘵𝘩𝘪𝘯𝘬 𝘱𝘦𝘰𝘱𝘭𝘦 𝘤𝘰𝘮𝘱𝘳𝘦𝘩𝘦𝘯𝘥 𝘩𝘰𝘸 𝘩𝘢𝘳𝘥 𝘪𝘵 𝘪𝘴 𝘵𝘰 𝘥𝘦𝘴𝘪𝘨𝘯 𝘢𝘯𝘥 𝘣𝘶𝘪𝘭𝘥 𝘢 𝘭𝘢𝘯𝘥𝘪𝘯𝘨 𝘱𝘢𝘨𝘦 𝘧𝘳𝘰𝘮 𝘴𝘤𝘳𝘢𝘵𝘤𝘩 𝘪𝘯 45 𝘮𝘪𝘯𝘶𝘵𝘦𝘴 𝘸𝘪𝘵𝘩 𝘩𝘢𝘳𝘥𝘭𝘺 𝘢𝘯𝘺 𝘱𝘳𝘦𝘱 𝘪𝘯 𝘧𝘳𝘰𝘯𝘵 𝘰𝘧 𝘢 𝘴𝘵𝘢𝘥𝘪𝘶𝘮-𝘴𝘪𝘻𝘦𝘥 𝘢𝘶𝘥𝘪𝘦𝘯𝘤𝘦. 𝘐𝘵 𝘥𝘰𝘦𝘴𝘯’𝘵 𝘮𝘢𝘵𝘵𝘦𝘳 𝘪𝘧 𝘺𝘰𝘶’𝘳𝘦 𝘶𝘴𝘪𝘯𝘨 𝘞𝘦𝘣𝘧𝘭𝘰𝘸, 𝘓𝘰𝘷𝘢𝘣𝘭𝘦, 𝘰𝘳 𝘍𝘳𝘢𝘮𝘦𝘳 𝘧𝘰𝘳 𝘵𝘩𝘢𝘵 𝘮𝘢𝘵𝘵𝘦𝘳. 𝘠𝘰𝘶𝘳 𝘧𝘪𝘯𝘨𝘦𝘳𝘴 𝘴𝘩𝘢𝘬𝘦 𝘴𝘰 𝘪𝘯𝘵𝘦𝘯𝘴𝘦𝘭𝘺" My takes and impressions - in no particular order → This was highly entertaining - especially as other design OGs like Hunter Hammonds were commenting and judging the builds → This challenge proves how much ground AI is gaining. Just a few years ago, the idea of a total beginner competing with a seasoned pro in web design would have been laughable. → Taste won. Because taste—an intuitive, human ability to recognize what feels right—is still something AI struggles to master. If anything, this challenge was a preview of what’s coming: while AI will keep getting better, we will need to appreciate where human creativity, taste, and instinct will need to remain.

𝘠𝘰𝘶 𝘱𝘳𝘰𝘣𝘢𝘣𝘭𝘺 𝘥𝘰𝘯'𝘵 𝘯𝘦𝘦𝘥 𝘢𝘯 𝘈𝘐 𝘢𝘨𝘦𝘯𝘵 to automate your process. Shoutout to Nick Saraev for expressing this idea so rightly in his recent YouTube video. We are still right in the middle of the 𝘈𝘐 𝘢𝘨𝘦𝘯𝘵 𝘩𝘺𝘱𝘦 𝘱𝘩𝘢𝘴𝘦 where these are expected to solve everything - from automating manual processes to running the whole company. The immense majority of AI and automation use cases does not need agents. 𝘈𝘶𝘵𝘰𝘮𝘢𝘵𝘪𝘰𝘯𝘴 and 𝘈𝘐 𝘸𝘰𝘳𝘬𝘧𝘭𝘰𝘸𝘴 will be just "enough", cheaper and way more reliable. Highly recommend the AI automation fanboys watch and internalise what Nick states here...

Notion's API is 𝘩𝘦𝘭𝘭. This hack lets you retrieve 𝘢𝘭𝘭 𝘵𝘩𝘦 𝘤𝘰𝘯𝘵𝘦𝘯𝘵 𝘧𝘳𝘰𝘮 𝘢 𝘴𝘪𝘯𝘨𝘭𝘦 𝘕𝘰𝘵𝘪𝘰𝘯 𝘱𝘢𝘨𝘦 — no matter what's in it: Extracting content from Notion is notoriously painful and frustrating because the API delivers data block by block — text, images, lists — all separate. But with this trick, you'll get a clean block of text ready to use in your AI workflows. In this tutorial, we're using Make and ChatGPT to go from list of blocks to one content block - from there you will be able to do pretty much anything you need: → Summarize meeting notes → Repurpose the content of a blog article → Extract insights from a desk research Just to mention a few. The Make scenario is quite easy to build - once you've figured out the trick: Step 1️⃣: List Notion page contents Step 2️⃣: Aggregate the blocks Step 3️⃣: Convert to JSON Step 4️⃣: Process with ChatGPT ✨ And to make things easier - we've written a detailed step-by-step tutorial with screenshots explaining the process. Want it? Let me know in the comments and I'll share it with you.

After DeepSeek last month, China strikes again with the launch of the first general AI agent: Manus Manus is an AI agent that combines: 1️⃣ Advanced reasoning 2️⃣ Deep research abilities 3️⃣ Task execution capabilities Put differently: Manus is o1, Deep Research and Operator combined 😅 Manus is the first of its kind - but it's where all models are headed. From knowledge -> to action. Check the three demos in this video to see how wild this combination is. Sam Altman: your turn.

Building an AI agent? Then make sure to always have a 𝘩𝘶𝘮𝘢𝘯 𝘪𝘯 𝘵𝘩𝘦 𝘭𝘰𝘰𝘱. Here’s why: This is easily the most important tip I give to our students looking to build agents. The greatest strength of AI agents is their ability to work autonomously and improvise when facing a new situation Turns out this is also their biggest 𝘸𝘦𝘢𝘬𝘯𝘦𝘴𝘴. When left unchecked, agents lack the judgment to recognize when their outputs don’t make sense in context. Human in the loop solves this. Human in the loop (HITL) refers to involving a 𝘩𝘶𝘮𝘢𝘯 𝘵𝘰 𝘰𝘷𝘦𝘳𝘴𝘦𝘦, 𝘨𝘶𝘪𝘥𝘦, 𝘢𝘯𝘥 𝘷𝘢𝘭𝘪𝘥𝘢𝘵𝘦 𝘵𝘩𝘦 𝘰𝘶𝘵𝘱𝘶𝘵𝘴 𝘰𝘧 𝘢𝘯 𝘈𝘐 𝘢𝘨𝘦𝘯𝘵. Without a human in the loop: → Agent generates results → No error detection Outcome: partially valid or invalid results With human in the loop: → Agent generates results → Humans catch and correct errors   Outcome: consistently valid results​​​​​​​​​​​​​​​​ There are tons of way to build a HITL - Slack and email are my go-to interfaces for now. I’ll expand on that topic in future posts. For now, if you’re building an AI agent, remember this: autonomy is powerful, but oversight is priceless. Keep a human in the loop — it’s the secret sauce that turns “almost right” into “spot on.”