🚀 Den Delimarsky: The Model Context Protocol (MCP) Explained — Why It’s the "USB-C for AI"

A few weeks ago, I took April Zhu's Colorfigures skill, tweaked it to draw 8-bit pixel art, and rebuilt a little world called kirupaVille. By hand, that would have taken me about ten hours. With AI, it took minutes. That speed is thrilling, and it leaves me with a question I keep coming back to: if AI can make almost anything, which part of the work is still ours?In this conversation, I chat with April Zhu, a Singapore-based product designer trained in architecture and the creator of the Colorfigures and Tinycharts AI skills. We talk about why she left architecture for the faster feedback loops of tech, how a hand-drawn washing machine turned into Colorfigures, why she recently unsubscribed from Photoshop, and why taste and judgment are the parts of design she doesn't want to outsource.We also talk about personalized interfaces that April finds equally exciting and scary, what happens to work and purpose as AI takes on more of it, WALL-E and the hedonistic treadmill, brain implants and the last fully biological humans, third places, and why being more productive keeps meaning more work. We close by wondering whether we're running out of imagination, with stops at the World's Fair, Star Trek, moon bases, and a new world order.🔗 Links & Mentions---------------------------------------📖 Full transcript and show notes: https://www.kirupa.com/podcast/april_zhu_design_taste.htmApril's Links:--💼 LinkedIn: https://www.linkedin.com/in/aprilzhu423/🌐 Website: https://april-zhu.com/🎨 Colorfigures (her hand-drawn figures skill): https://april-zhu.com/colorfigures📊 Tinycharts (her hand-drawn charts skill): https://april-zhu.com/tinycharts💬 X: https://x.com/Aprilzz423Kirupa's Links:--💬 Follow Kirupa: https://twitter.com/kirupa📚 Articles & blog posts: https://www.kirupa.com🗞️ Kirupa's newsletter: https://www.kirupa.chat❓ Discuss this further: https://forum.kirupa.com🕒 Chapters--------------0:00 – From architecture to faster feedback loops9:08 – How Colorfigures began14:37 – Unsubscribing from Photoshop16:54 – Don’t outsource your taste22:53 – Personalized design and the future of UX28:44 – Work, purpose, and the hedonistic treadmill40:49 – Cyborgs and living in the future47:57 – Third places and social creatures56:41 – Why more productivity means more work1:00:15 – Predicting the future💡 Key Takeaways--------------1. Architecture's feedback loop runs in years, while tech can put your work in front of people within weeks. That faster loop is what pulled April from one to the other.2. Colorfigures started with a simple question: can AI make something clumsy and hand-drawn, something warm that has temperature?3. For a design-trained person, the output matters more than the medium. April recently unsubscribed from Photoshop and still uses Figma for UI work.4. Taste and judgment are a designer's biggest advantage. Let AI brainstorm the variations, then make the judgment call yourself.5. Being more productive with AI doesn't automatically buy you more free time, and April suspects things will get worse before they get better.If you're a designer or developer wondering which parts of your craft to hand to AI and which to keep for yourself, this conversation is for you.#AI #Design #Colorfigures #UXDesign #FutureOfWork #Creativity #Architecture #Kirupa #AprilZhu

What does a fulfilling life look like once the job that filled your calendar, and a good chunk of your identity, is gone? My hunch is that the answer has less to do with your next title and more to do with structure, curiosity, and having problems you actually want to solve.In this conversation, I chat with Amit Chopra, a technology leader and author who spent three decades building products at companies like Microsoft, Google, Amazon, and Meta. We start in a school computer room in Delhi and end up with AI agents, flying cars, and advice for students about to enter an AI-shaped job market.We also talk about Easter eggs in early Microsoft products, why Amit moved from code to product, the ups and downs of an empty calendar, bucket lists, corporate conditioning and the Shawshank Redemption grocery-store scene, whether a computer science degree is still worth it, the jobs that only existed to patch inefficiencies, and the moments that made Amit sit up and say "aha" about AI. Amit closes with why anyone starting out today is arriving at a remarkable time, and he's available for seminars and guest lectures at universities and startup accelerators.🔗 Links & Mentions---------------------------------------📖 Full transcript and show notes: https://www.kirupa.com/podcast/amit_chopra_life_after_nine_to_five.htmAmit's Links:--💼 LinkedIn: https://www.linkedin.com/in/amitchopra🌐 Website: https://www.chopra.dev/🎓 Guest lectures and seminars (for universities and startup accelerators): https://www.chopra.dev/guest-lectures📖 Without a Badge. What Next? (his book): https://www.amazon.com/dp/B0H857DBKQ📘 Portfolio of Happiness (his book): https://www.amazon.com/dp/B0DYF4NRHXKirupa's Links:--💬 Follow Kirupa: https://twitter.com/kirupa📚 Articles & blog posts: https://www.kirupa.com🗞️ Kirupa's newsletter: https://www.kirupa.chat❓ Discuss this further: https://forum.kirupa.com🕒 Chapters--------------0:00 – From a Delhi computer room to a career in tech10:05 – Psion handhelds, 8051 drills, and Snake in Excel15:58 – Easter eggs and the move from code to product23:10 – Leaving the nine-to-five and facing an empty calendar29:31 – Bucket lists, conditioning, and motivation37:13 – Mentoring students in a fast-moving AI landscape43:44 – Is a computer science degree still worth it?49:46 – Standing out when AI does the middle59:10 – Jobs born from inefficiency and the case for AI literacy1:10:00 – Aha moments, flying cars, and a great time to start out💡 Key Takeaways--------------1. Technology was always a means to an end. Coding was never the point, so thinking about the problem you're solving matters more than the stack you know.2. An empty calendar can feel like freedom or a tailspin. It helps to have structure, like a bucket list, and motivation that comes from within instead of from other people's validation.3. Education can't fill the gap when the goalpost keeps moving, so build a builder's mindset: pick a problem that bothers you and prototype a solution with AI.4. Many jobs existed to patch inefficiencies in the system, and those are the ones AI absorbs first. Knowing how to prompt and validate AI is now table stakes.5. A computer science degree still matters, but it won't have a recruiter waiting at the door. People starting out today are arriving at a remarkable time.If you're wondering what comes after the nine-to-five, or how to prepare for a career in an AI-shaped world, this conversation is for you.#Career #AI #Mentoring #Retirement #Layoffs #ProductManagement #Education #Kirupa #AmitChopra

When AI can do all the hard things we once had to learn the hard way, what is left for education to teach our kids? My hunch is critical thinking. Knowing how to learn and being able to think critically still feel durable, even when every other answer is a fingertip away.In this conversation, I chat with Dave Thackeray, a career journalist who runs the strategic storytelling studio Word And Mouth. We set out to talk about making sure AI behaves in ways we understand, but the conversation quickly went in many directions, from Dave's first steps on a ZX Spectrum 48K in 1982 to authenticity, curiosity, critical thinking, and purpose.We also talk about whether a perfect barometer of truth would be worth what it costs in privacy, scoping a project before AI agents start building it, The Return (the book Dave wrote with AI's help about Adam Smith), longevity, loneliness, and why new societies only become real experiments when there's no easy way back. Dave closes with the book he feels sums up the conversation: Black Box Thinking by Matthew Syed.🔗 Links & Mentions---------------------------------------📖 Full transcript and show notes: https://www.kirupa.com/podcast/dave_thackeray_staying_human.htmDave's Links:--💼 LinkedIn: https://www.linkedin.com/in/davethackeray/🌐 Word And Mouth (his storytelling studio): https://wordandmouth.com/🗞️ All Mouth (his newsletter): https://thacknology.substack.com/Kirupa's Links:--💬 Follow Kirupa: https://twitter.com/kirupa📚 Articles & blog posts: https://www.kirupa.com🗞️ Kirupa's newsletter: https://www.kirupa.chat❓ Discuss this further: https://forum.kirupa.com🕒 Chapters--------------0:00 – A Spectrum 48K and a lifelong curiosity5:37 – Authenticity, incentives, and a barometer of truth15:41 – Lazier or more curious with AI24:23 – Stability, uncertainty, and critical thinking38:44 – Think it through before you build44:57 – Purpose, struggle, and hidden superpowers52:35 – A simpler life before the factory1:00:20 – Longevity, loneliness, and autonomy1:05:01 – Wealth of Nations among the stars1:12:45 – Pioneers, new states, and no way back💡 Key Takeaways--------------1. In a world of polished personas and AI avatars, being yourself is a form of respect for the people you're talking to.2. AI can make you lazier or more curious. The difference is whether you stop at the first answer or keep asking.3. Knowing how to learn and thinking critically stay durable, even when AI moves too fast for a new normal to settle in.4. Agents and skills can speed up the build, but they can't replace a clear understanding of what you're making before you start dancing on the keyboard.5. Purpose often grows out of struggle, failure, and chance conversations, which makes it a harder question as AI takes on more of the work.If you're wondering what stays human as AI takes on more of the work, or how to keep thinking for yourself while it does, this conversation is for you.#AI #CriticalThinking #Authenticity #Education #Purpose #Longevity #Journalism #Kirupa #DaveThackeray

What happens when one person can run six to eight Claude Code sessions plus another six to eight Codex sessions at the same time? Coding stops being the only challenge. Coordinating the work starts to look a lot like leading a team.In this conversation, I chat with Francis Ma, an entrepreneur, tech executive, investor, and former Google product leader who helped shape Firebase. We trace his path from an IBM 8086 and QuickBASIC to Amazon, startups, and Firebase, then get into the systems behind effective AI-agent work: clear goals, shared design rules, coding standards, model fit, and feedback loops.We also talk about friction-free onboarding, the accidental paths that lead people into product management, autonomous forum personas, cheaper experiments, and why the boundaries between product managers, engineers, and designers are getting blurrier.🔗 Links & Mentions---------------------------------------📖 Full transcript and show notes: https://www.kirupa.com/podcast/francis_ma_systems_thinking.htmFrancis's Links:--💼 LinkedIn: https://www.linkedin.com/in/francisma/🌐 JoyCrew (his startup): https://joycrew.ai/💬 Blog: https://francisma.com/Kirupa's Links:--💬 Follow Kirupa: https://twitter.com/kirupa📚 Articles & blog posts: https://www.kirupa.com🗞️ Kirupa's newsletter: https://www.kirupa.chat❓ Discuss this further: https://forum.kirupa.com🕒 Chapters--------------0:00 – Early computers, games, and BASIC6:03 – From engineer to founder and product leader9:54 – Friction-free onboarding and accidental PMs20:29 – Systems thinking for fleets of agents25:29 – A parallel-agent working setup29:12 – Design the system before the prompt31:27 – What an autonomous forum learned39:46 – Goals, standards, feedback, and model fit47:26 – Cheaper experiments and first-principles work58:24 – Creating luck and rethinking product work💡 Key Takeaways--------------1. A friction-free first step builds the confidence people need to keep exploring.2. More agents don't remove the need for coordination; they make shared goals and standards more important.3. A strong design system lets independent agents build features that still feel like parts of one product.4. Claude Code, Codex, and Gemini behave like different teammates, each better suited to certain kinds of work.5. Cheaper prototypes let teams test strategic directions instead of debating them only on paper.If you're building with coding agents or thinking about how product work changes when everyone can build, this conversation is for you.#AI #AIAgents #SystemsThinking #ProductManagement #Firebase #ClaudeCode #Codex #Kirupa #FrancisMa

Artificial intelligence is changing how we learn, teach, and even think. So what happens to computer science education when students can ask ChatGPT to do their homework — or even write perfect code?In this conversation, I chat with Elisa Cundiff, computer science instructor at Colorado State University and one of the National Science Foundation’s 100 “superstar educators.” We talk about teaching in the era of AI, why struggle still matters, and how critical thinking might just be the last human advantage.🔗 Links & Mentions---------------------------------------🌐 Elisa's LinkedIn: https://www.linkedin.com/in/elisacundiff/--💬 Follow Kirupa: https://twitter.com/kirupa📚 Articles & blog posts: https://www.kirupa.com

I’m back with Den Delimarsky to talk about the Model Context Protocol (MCP) — the open standard that’s quietly transforming how AI assistants connect to data, tools, and services.We unpack what MCP is, why it matters, and how it’s becoming the foundation for agentic workflows, developer tools, and authentic AI experiences.🔗 Links & Mentions---------------------------------------🌐 MCP Registry: https://registry.modelcontextprotocol.io🧠 Learn more about MCP: https://modelcontextprotocol.io---🎙️ Follow Den: https://www.linkedin.com/in/dendeli/📝 Den's blog: https://den.dev/--💬 Follow Kirupa: https://twitter.com/kirupa📚 Articles & blog posts: https://www.kirupa.com

Among the many cool things Cameron Balahan has done, in the mid-1990’s, he created one of the best sites for tech enthusiasts. He created winshareware.com, a site where you could go and get the latest details on things you can download and run on your Windows machine.This chat is a trip down memory lane for those of us who started off with the internet in the 1990s. We talk about classic shareware (or nagware?) apps like WinZip and WinRar. We talk about OpenGL and DirectX and how that influenced whether you played Unreal Tournament or Quake. We talk about the Windows registry, connecting to the internet via dial-up, hunting for drivers, AOL CDs, classic video games like Math Blaster, and a whole lot more.This is going to be a great chat, so invite your friends, get some popcorn, and let’s get started!Cameron's Links:---------------------🌎 Blog: https://cameronb.dev/💼 LinkedIn: https://www.linkedin.com/in/cameronbalahan/🐦 Twitter / X: https://x.com/cameronbalahanKIRUPA:---------------------🗞️ Sign-up to Kirupa's newsletter: https://www.kirupa.chat❓ Discuss this further: https://forum.kirupa.com-------

If you look up "serial entrepreneur" in the dictionary, you'll likely see Darius's profile!Darius is someone who I've known for decades both online and offline, and I've gotten a chance to see first hand his ability to create products and companies that people love. In this fun chat, we cover everything from how he got started in tech, his early stint in building companies, parenting, life, giving back to the community, and a whole lot more. This is going to be a fun one!Darius's Links:---------------------🌎 Blog: https://www.bubs.co/🚗 Rides (his latest venture): https://www.rides.co/💼 LinkedIn: https://www.linkedin.com/in/dariusmonsef/KIRUPA:---------------------🗞️ Sign-up to Kirupa's newsletter: https://www.kirupa.chat❓ Discuss this further: https://forum.kirupa.com-------

While technology trends evolve and change, the need to connect with others and build a community has remained a constant. In this fun chat, Zach Hawtof, Co-Founder and CEO of Tightknit, shares his experience with community building and how organizations who use Slack can benefit from a deeper story here.Along the way, we touch upon a bunch of topics from NYC vs. SF, product management, the importance of SEO, authenticity (and AI-based authenticity), and a whole lot more.Zach's Links---------------------🌎 Tightknit: https://tightknit.ai/💼 LinkedIn: https://www.linkedin.com/in/zhawtof/KIRUPA:---------------------🗞️ Sign-up to Kirupa's newsletter: https://www.kirupa.chat❓ Discuss this further: https://forum.kirupa.com-------

While public speaking consistently ranks as one of the most terrifying things people will ever need to do, some people like Dan Chuparkoff make it their full-time profession. In this fun chat, we talk about Dan's career progression towards becoming a keynote speaker and dive into all the ins/outs of what life as a professional speaker is like.Dan's Links---------------------🌎 Website: https://danchuparkoff.com/🎥 YouTube Channel: https://www.youtube.com/c/DanChuparkoff💼 LinkedIn: https://www.linkedin.com/in/chuparkoff/📚 Instagram: https://www.instagram.com/danchuparkoff/-------🗞️ Sign-up to Kirupa's newsletter: https://www.kirupa.chat❓ Discuss this further: https://forum.kirupa.com-------