Csongor Vogel

There is no limit to learning—that’s why I included AZ Premium’s new feature teaching students how to learn faster using a framework inspired by Elon Musk’s “first-principles thinking.” If you learn how to learn, switching fields or mastering new skills becomes much easier. Why AZ Premium’s Learning Framework Works: - Helps you identify and deepen core concepts first. - Enables faster absorption of advanced topics later. - Makes you more problem solver. #AZ #AZPremium

122

3 Comments

🚀 “Apps Are Losing Face — AI is gaining it" At DevDay 2025, #OpenAI announced that #ChatGPT can now directly host apps like Spotify, Canva, Zillow, Coursera, Booking, Expedia, and Figma through its new Apps SDK. This isn’t incremental — it’s disruptive in two dimensions: 1️⃣ AI becomes the new human-centric interface We’re shifting from “open app X, then app Y” to “talk to ChatGPT, and it orchestrates apps behind the scenes.” - Apps fade into the background — functional modules delivering capability — while AI becomes the digital front end. In short: AI = the new digital. 2️⃣ The convergence of two developer ecosystems This move fuses app developers (~30M) and AI developers (~5M) into one powerful innovation loop at massive scale, through enabling access to 800M ChatGPT users.... This will accelerate #Agentic AI where intelligent agents can reason, plan, and act across tools like calendar, CRM, design, and commerce — seamlessly. 🔮 We’re stepping into an era where interfaces default to #AI and apps become modular capabilities under the hood. The winners will be those who position themselves as trusted, scalable capabilities inside the AI stack, not standalone screens fighting for attention. If you’re building in this space — I am keen to learn from your insights.... The future of “apps behind the scenes” is unfolding fast. https://lnkd.in/ekDUE_5K Steve Jarrett Mathieu Ducrot Miguel Angel Alvarez Domínguez Philippe Ensarguet Aliette Mousnier-Lompré Julien Sicart Hriday Ravindranath

101

14 Comments

🧱Building LangGraph: Designing an Agent Runtime from first principles Nuno Campos wrote a detailed blog on how he designed and built LangGraph, going deeper than any other piece of content we've put out. Includes: 1. what’s different about building agents compared to traditional software 2. how we turned these differences into required features. 3. how we designed and tested our framework for these requirements. https://lnkd.in/gMdnRh9s

538

15 Comments

💡 The Paradox of Vibe-Coding: Build Fast, Ship Slow I recently tested the limits of “vibe-coding” — building an app in just 1 week. The surprising part wasn’t development speed, but everything that followed: ⏳ 3 weeks of App Store submission cycles ❌ Rejections due to quirks in the process 🧑‍💻 Countless moments where tools still assumed I “think like a developer” This highlights a bigger industry truth: today’s vibe-coding stack (Cursor + Claude Sonnet, Expo builds, app store workflows) is fast enough for creation — but still too technical for true no-code scale. The opportunity ahead isn’t in writing code faster. It’s in: * Frictionless publishing → cutting approval delays & rejections * Abstracted build systems → removing the need to wrestle with config/tools * Lowering cognitive overhead → enabling creators to build without a developer’s mindset Whoever solves these challenges at scale will unlock a much larger TAM for no-code apps — and bring a new wave of builders into the ecosystem. 🚀 We’re close. But the real breakthrough isn’t coding speed. It’s everything that happens after.

107

10 Comments

Two weeks ago, 100 builders moved 30 projects from concept to prototype in less than a day at the Center for Digital Technology and Management (CDTM) Hackathon in Munich. What stood out to me is that beginners are building right next to seasoned developers, thanks to AI tools with simple UIs like Lovable, ElevenLabs, and n8n. These products give anyone the agency to solve problems that once felt out of reach. This is my call action to non-technical builders, to come join these events. But come prepared, play with the tools beforehand, every one of them has free trials, great docs, and video tutorials. If you spend just a few hours experimenting, you’ll walk into the hackathon ready to turn your idea into a demo. 👉 Next up: the Anthropic Hackathon (CDTM x TUM.ai, feat. Lovable + ElevenLabs) on Sept 24–26. Apply here: https://luma.com/m9lr6lux #AIHackathon #AWS #Innovation #CDTM #Munich #TechForGood #Entrepreneurship #GenerativeAI #BuildersGonnaBuild #FutureOfWork

62

1 Comment

�� Startups are hard. Not just because of the long hours, shifting priorities, or constant uncertainty. They're hard because they demand everything — especially when you're building one alongside a full-time role. At mentor.sh, Dima Maleev and I are learning this firsthand while I work full-time as Staff Engineer at Netflix. There's no playbook for this. We code when we can. Answer customers between meetings. Ship features at midnight. Debug on weekends. No elaborate processes, no code reviews – just two people building what's right, when it needs to be built. The reality? Something always gives. Maybe it's the gym session you skip. The dinner you cancel. The Netflix series ironically left unwatched. Your spouse gets used to you typing during breakfast. The pressure compounds. Full day at Netflix, then straight into startup mode. Context switching becomes your superpower and your curse. But here's what it teaches you: How to operate with radical efficiency. How to identify what truly matters. How to build without overthinking. These constraints force innovation in ways comfort never could. Building a startup while working full-time doesn't just test your limits. It fundamentally changes how you approach everything. For those juggling similar challenges — what sacrifices have been worth it? What weren't?

41

6 Comments

The appetite for learning right now is really high, and I find it inspiring to see so many people wanting to learn. So we're sharing as much as we can on how we're building Fin. Designing and building successful AI products is hard. It requires a new set of skills, very different to the last decade of Saas and consumer CRUD apps. Our AI group has a new podcast, and we've two really great episodes live. They are very approachable for everyone keen to learn the details of making AI products that really work well. First one is Fergal and Fedor discussing out AI Insights, and the second is Fergal and Pratik discussing how we get Fin to execute complex customer issues that need procedures. Link to each one in the comments!

60

4 Comments

I landed in the world of AI & Product Management by accident, and understanding machine learning without any previous knowledge was a tough uphill battle - So I'm launching the 𝗔𝗜 𝗣𝗿𝗼𝗱𝘂𝗰𝘁 𝗣𝗹𝗮𝘆𝗯𝗼𝗼𝗸!, The book I wish I had when I started my journey into AI Product Management. Written together with Marily Nika, Ph.D and published by Wiley, this book is all about how to understand AI&ML and the role of Product Managers in AI without having to be technical, knowing how to code or needing any previous experience in ML. 💎 AI Product Management is not about learning how to use tools to automate PRDs or keeping up with which vibe coding tools is best. So what is this book about? We divided it in three parts: ⭐️ Part I: Speak The Language Of AI I remember the first time the Data Scientists in my team told me "The model has a recall of 83%, what do you think?" I felt like a dear in headlights. Part I is all about understanding the difference between AI, ML, Deep Learning, GenAI and why it matters to you - the Product Manager - and to your Product Strategy. ⭐️ Part II: Chart Your Career in AI There's three types of AI Product Managers - AI-Experiences PM: You integrate AI&ML experiences into your product - AI-Builder PM: You build ML Models and ML Platforms - AI-Enhanced PM: You use AI to be a more efficient PM We talk about the difference and similarities between the roles, what their day-to-day looks like and how you can navigate your career to find your own path in the world of AI Product Management. ⭐️ Part III: Go From Theory to Impact Beyond learning frameworks, metrics and how LLMs work, this is all about how you apply AI&ML as a Product Manager to find opportunities for AI, understanding the impact and, very importantly, how to think about Responsible AI in all of this. I'm very excited about this launch and you can preorder now! https://lnkd.in/gNiaUQeg

122

30 Comments

Shipping fast is good. Shipping safely is non-negotiable. Every startup wants to “move fast and fix in prod” until what breaks is users trust and results into less usage. Fast shipping gets you to market. Safe shipping keeps you in the market. I’ve seen both sides: → Engineers under pressure, merging code at 5AM → Compliance skipped - “just for now, this is urgent” → Rollbacks on Friday night → Incident reports that could’ve been prevented or peer-reviewed Speed isn’t the villain. But speed without guardrails? That’s how you leak PHI, miss audit windows, or worse — find your brand on a Twitter thread at midnight. Here’s the approach we followed that actually worked: ✅ Build security into the CI/CD. Almost doing a small VAPT in each push. ✅ Automate every checkpoint possible. ✅ Shifted on secure infra with compliance ✅ Give devs confidence, not friction Shipping fast is about momentum. Shipping safely is about compounding trust. As Naval Ravikant says if its not compounding, its not wealth. And Secure Process takes longer to build but faster to use. #CTO #EngineeringLeadership #DevSecOps #StartupInfra #DhavalWrites #BuildBetter

97

4 Comments

If you’re a product manager learning about APIs, you need to know the 3 main types. We’re taking the coffee order analogy a step further... by explaining how each type “orders” differently. ☕ Here’s a preview: 👇 💬 REST — the needy customer 📝 GraphQL — the prepped customer 😴 Webhook — the chill customer You’ll learn: ↳ What makes each type unique ↳ How they handle requests differently ↳ When to use which No CS degree required. Just caffeine and curiosity. 📌 Save this for your next dev sync or PM interview. ——— 👋🏻 Hi, I’m Jackie! I break down technical concepts for new and aspiring product managers.

612

63 Comments

When we started writing blogs at nekuda, our chief writing officer :) Idan Levin, pushed to write clearly and authentically. Initially, we leaned heavily on ChatGPT for drafts. But over time, we realized readers value our real thoughts, not masked in AI language. Now, we write from scratch. You’ll probably notice grammar mistakes or occasional typos, but that’s intentional. It signals authenticity. It’s human. Today, I’m sharing a new blog about headless commerce and the future commerce stack we see emerging from a merchant’s perspective: /catalog, /cart, /checkout, and /postorder. These end-points will be key pillars of the next era in commerce, especially with the rise of agentic shopping. We’ve been getting many questions about the future of agentic commerce stacks. Here’s our take, straightforward and transparent. We also discuss the delicate balance between brand representation and AI driven interfaces, highlighting emerging standards like MCP-UI (https://mcpui.dev/). Read our full thoughts here, and let us know what you think: https://lnkd.in/erevTHEn #HeadlessCommerce #AgenticCommerce #MCP #Ecommerce #FutureOfRetail

59

5 Comments

I was just hired for another C level role at Zell: Chief AI Ops Officer My role is to run a very simple yet effective workshop Every two weeks, we ask: what blocks us from going 10X faster? Just two simple questions: → How do we ship 10X more features? → How do we book 10X more meetings? We put every blocker on the table. Then? We pick one target per question. Two weeks to crush it. Here you have some successful experiments: - Meeting prep: before every meeting an agent looks up LinkedIn the person I am talking too, checks what we discussed last time and sends a recap on Slack. - Post call analysis: after every call a sales expert agent sends on our sales channel the pain points of the client, risk factors that could kill the deal, some coaching tips on what I could do better next time and a role plays I can do to improve. - Minor bugs and improvement: when customers report bugs or ask for some change we use Devin from Cognition, just tag him on Linear and he goes ahead and fixes things. Next two weeks' focus: - Improve tickets quality: trying out ChatPRD and Raycast AI - Automate testing: looking for AI that can test all the features we ship. For years you could only scale a company through people, now you can also do it through processes. When we will hire, and it will happen soon we don't expect people do execute tasks but to orchestrate AI systems. Any major process you automated with AI? 🚀

37

3 Comments

We pay our engineers commission not just for building software, but for driving revenue. Every month, I share four numbers with the whole team: New bookings revenue Total MRR Churn Profit We are growing by 20% per month, and we pay commission to engineers based on monthly profit. There is a clear path to an additional $5k per month for engineers just in commission. Why? Because after 20 years in SaaS, I’ve seen the same pattern at every company: Sales and engineering operate in silos Engineers are told to “just build” Sales signs bad-fit deals CS puts out fires Nobody is aligned on outcomes And it’s always pissed me off. So we changed it. Since we rolled this out, engineers now: Flag ICP issues before we close deals, because bad-fit customers = churn = less commission Sit with CS to understand recurring pain points, and quietly build product to solve them Prioritise features that reduce churn or increase conversion, because it impacts their pay Question infra decisions, because cost hits profit, and profit drives bonus Proactively review onboarding and activation flows, because time-to-value matters A few real outcomes: Onboarding time dropped 40% after engineers built internal tools (unsolicited) A single feature update cut churn by 19% We walked from a big deal flagged by engineering as a churn risk. They were right. It’s simple: When engineers think like owners, the whole company wins. So we show them the numbers. We give them real incentive. And they show up in ways we never had to ask for.

60

14 Comments