DN #19: Vibe Coding Slop, Full-Stack Designers & Distribution as Product (w/ Jacob Counsell)

DN #19: Vibe Coding Slop, Full-Stack Designers & Distribution as Product (w/ Jacob Counsell)

"Distribution is a nightmare for everybody. But secretly, it's a product problem." I talk to Jacob Counsell, product designer of 15 years in Silicon Valley tech and founder of LaunchChair.io [http://LaunchChair.io], about why most vibe-coded apps feel broken after four months, why designers will touch code again after a decade of being told not to, and why founders complaining about distribution usually have a product problem they're afraid to admit. Jacob breaks down the wedge LaunchChair plays in the vibe-coding space, the spec-aware prompt engine behind it, and why he thinks a team of three with agents now ships what fifty people shipped three years ago. In this episode: - Vibe Coding Slop: Why most Lovable and Bolt apps feel broken after four months of building. - Distribution as Product: Why founders blaming distribution usually have a product problem they're afraid to name. - The Full-Stack Designer Returns: Why designers will touch a lot more code in the AI era. - The HCI Overcorrection: How we stopped hiring weirdos from art school and the internet got boring. - The 5-Person Team Thesis: How a team of three with agents now ships what fifty people shipped three years ago. - LaunchChair's Wedge: A spec-aware prompt engine that forces functioning features instead of broken Lovable mockups. - The Silofication Problem: Why big tech ships bugs that sit unfixed for two-plus years. - Baseline + A/B Testing: Why endless user research before launch is a trap, and what to do instead. 🎧 Full episode on all podcast platforms 💬 Should designers touch code in the AI era? Where do you land? Let us know in the comments! 🔔 Please like and subscribe! Every subscriber helps our channel grow. #DN #VibeCoding #ProductDesign #JacobCounsell #LaunchChair #Founders #AI #StartupBuilding #Distribution #Designers Timestamps: 0:00 Intro 0:29 Why Jacob built LaunchChair: a wedge into the vibe-coding slop problem 2:08 Being technical + design: the full-stack designer advantage 3:30 How AI tools changed product design in the last year 4:30 The HCI overcorrection: why we stopped hiring weirdos from art school 6:38 Distribution is a nightmare — but secretly, it's a product problem 9:00 Agent orchestration for LinkedIn (without becoming a Claude Slop Cannon) 11:55 The full-stack designer is back: designers will touch a lot more code 15:23 Smaller teams, fewer silos: why a team of 3 ships what 50 used to 18:35 When a fast-and-loose team works (and when it doesn't) 22:58 Baseline + A/B testing beats endless user research 24:09 LaunchChair's wedge: spec-aware prompts that build functioning features 28:59 BuildHop and PromptJoy: two side products dogfooded with LaunchChair 33:50 Why most vibe-coded apps look broken (the load-more example) 35:30 Acceptance criteria + remediation prompts: how LaunchChair fixes hallucinations 35:57 The future of design: designers shipping production-ready features

DN #18: Dirty Jobs, AI Robotics & Why Humanoids Are a Pipe Dream (w/ Jay Kapoor)

"Why would you give a robot human weaknesses?" I talk to Jay Kapoor, co-founder and general partner of VSC Ventures, about the "dirty, dusty, and dangerous" industries that VCs keep skipping over and why he thinks humanoid robots are the wrong bet. Jay started his career at the NFL and Madison Square Garden before spending 11 years investing in seed-stage companies, and now backs founders building AI and robotics for the deskless workers powering $80T of global GDP. Jay walks through why 90% of US factories still don't have a single robot in 2026, why the death of systems integrators is opening a massive deployment gap, and why the next great founders no longer need a decade in their industry to build in it. In this episode: - The DDD Thesis: Why dirty, dusty, and dangerous industries are the most underpriced sector in venture today. - The 90% Factory Gap: Why 9 out of 10 US factories still don't have a single robot in 2026. - The Death of Systems Integrators: Who's responsible for actually deploying robots when the consultants disappear. - Why Humanoids Are a Pipe Dream: The case for purpose-built robotics over Optimus and Figure. - The 500% Turnover Job: Recycling sorters quit before they find their parking spot, and why automation is the only fix. - The 4-to-1 Electrician Shortage: Why Jensen Huang says plumbers and electricians are the bottleneck on AI compute. - Founder-Led Storytelling in VC: Why VSC Ventures runs a 30-person PR agency for portfolio companies. - Founder-Market Fit Is Dead: Why a decade in an industry is no longer required to build in it. 🎧 Full episode on all podcast platforms 💬 Humanoid robots: pipe dream or future of labor? Let us know in the comments! 🔔 Please like and subscribe! Every subscriber helps our channel grow. #DN #Robotics #AI #VC #VSCVentures #DataCenters #Manufacturing #Tradespeople #Startups #Founders Timestamps: 0:00 Intro 0:35 The DDD thesis: dirty, dusty, dangerous industries explained 4:04 Why labor shortages + generational ownership shifts + AI converge into the opportunity 4:58 From the NFL to backing robots: Jay's path through media into venture 6:52 Building VSC Ventures around founder-led storytelling 9:35 Why founder-CEOs becoming industry champions is the new playbook 12:35 Why most VCs miss DDD opportunities (and how that creates pricing alpha) 16:13 Why fundraises are now won by speed of execution mid-process 19:28 Where capital is still underfunded: implementation, not robots 21:30 The 90% gap: why most US factories still have zero robots 23:39 The data center bottleneck and SoftBank's $40B automation bet 24:22 Why Jay is skeptical about data centers in space 32:42 Humanoid robots: pipe dream or the future of labor? 35:45 Purpose-built vs general-purpose robots in industrial settings 38:45 Where the worst labor shortages will hit next: electricians and plumbers 41:49 Where to find Jay: Twitter, LinkedIn, and the Climb show

DN #17: AI Sentence DNA, Voice as the Last Moat & Audience-First Building (w/ Sadok Hasan)

"Voice is the last moat." I talk to Sadok, growth and AI operator and founder of Bloomberry, about why the next defensible thing for solo founders isn't a feature but their authentic voice. Sadok spent the last decade running paid growth at Air Wallex, Pure Storage, Procore and Google before bootstrapping Bloomberry, an end-to-end content distribution system that learns your voice over time and writes in it. Bloomberry's research has identified 7,000+ cadences, structures, and phrases that AI consistently uses, and Saduk thinks killing those patterns is the difference between content that converts and content that gets scrolled past. In this episode: - AI Sentence DNA: The 7,000+ cadences and phrases that flag AI-generated content (and how to strip them out). - Voice as the Last Moat: Why authentic founder voice is the one thing AI cannot replicate at scale. - The 1-Comment First Customer: How Sadok's first ever LinkedIn comment, on day 2 of launch, converted into a paying user. - Bootstrap vs Venture: Why he is deliberately avoiding VC for now, and what burnt-out venture-backed founders are signaling. - Audience Before Product: Why content distribution should start before the product exists, not after. - Founder-Led Marketing: Why podcast appearances and personal posting beat scaled outreach for early-stage founders. - Protect the Mid-Funnel Leak: The unglamorous fix most founders skip while chasing top-of-funnel. - When to Raise: The $2K to $5K MRR threshold below which VCs will not take a meeting. 🎧 Full episode on all podcast platforms 💬 Bootstrap or raise: where do you land for an early-stage SaaS in 2026? Let us know in the comments! 🔔 Please like and subscribe! Every subscriber helps our channel grow. #DN #Bloomberry #AISlop #VoiceMoat #Bootstrapping #FounderLedMarketing #ContentMarketing #PerformanceMarketing #SaaS #Startups Timestamps: 0:00 Intro 0:18 Killing AI slop: the thesis behind Bloomberry 1:21 What Bloomberry actually does and why ChatGPT can't replace it 4:30 AI Sentence DNA: 7,000+ cadences and phrases that flag AI writing 6:17 Why he's bootstrapping instead of raising VC 7:52 The 1-comment first customer story 9:47 From Fry's Electronics retail floor to growth operator 13:30 When to start with paid ads (Google and Bing, not Facebook) 17:45 What "what sticks" actually means in early iteration 24:13 Why FinTech is the wrong industry for solo founders 27:41 The $2K-$5K MRR threshold for raising venture 31:56 Why B2B beats B2C for monetization 37:19 Building an audience in an AI-saturated world 39:05 Audience first, product second 41:24 Where to find Bloomberry

DN #16: Solving the Cold Start Problem & The Brutal Math of Marketplace VC (w/ Colin Gardiner)

"If you take the year Airbnb did its seed round... if you invested in every single one of the 2,500 deals that year, and everything went to zero except Airbnb, you'd still return a couple thousand X." In this episode of DN, Niklas talks to Colin Gardiner, the founder of Yonder VC (a pre-seed fund exclusively focused on marketplaces and network effects) and the former Chief Revenue Officer of Outdoorsy, where he helped scale the business to billions in rentals. Colin breaks down the absolute hardest challenge in tech: the "chicken or the egg" cold start problem. He explains why founders who buy ads to acquire demand early on are creating a "false positive" for product-market fit, and why building a "SaaS-enabled marketplace" is the ultimate cheat code to onboarding supply. The conversation also dives into the brutal mathematics of Venture Capital, the massive difference between "Market Making" and "Market Taking," and why Colin expects every single founder he invests in today to be completely AI-native. In this episode: • The Cold Start Problem: How to successfully launch a 2-sided marketplace from zero using the "SaaS-enabled" playbook (used by OpenTable and Outdoorsy). • The False Positive of PMF: Why paying for demand and running early ads is the most dangerous trap a marketplace founder can fall into. • Market Making vs. Market Taking: Why startups that create new markets (Airbnb, Polymarket) reach unicorn status, while commodity markets (Freight) struggle to survive. • The VC Power Law: The staggering math behind Airbnb's seed round, and why Yonder VC prioritizes a wide "surface area" over trying to pick a single winner. • The AI Amplifier: How Colin coded his own AI automations to parse through 900+ pitch emails, and why VCs will pass on you if you aren't an AI-first founder. • Pre-Seed Fundraising in 2024: Why you should delay raising venture capital as long as humanly possible to prevent massive equity dilution. 🎧 Full episode on all podcast platforms 💬 Are you building a SaaS product or a 2-sided marketplace? Let us know in the comments! 🔔 Please like and subscribe! Every subscriber helps our channel grow. #DN #Marketplaces #StartupFounder #ColinGardiner #YonderVC #VentureCapital #SaaS #NetworkEffects #ColdStartProblem #SiliconValley Timestamps: 0:00 Intro: Niklas meets Yonder VC founder Colin Gardiner 0:47 Studying the US Labor Market with Janet Yellen to scaling Outdoorsy 4:30 Solving the "Cold Start" problem: How to bootstrap supply 7:38 The biggest mistake founders make: Buying ads to force product-market fit 9:10 The "SaaS-Enabled Marketplace" framework (Outdoorsy, OpenTable, Google) 11:19 Market Making vs. Market Taking: Why commodity marketplaces fail 15:15 The VC Power Law: The insane math behind Airbnb's seed round 17:47 What Yonder VC looks for in pre-seed founders (High agency & stubbornness) 20:17 How AI is fundamentally changing the way VCs underwrite startup investments 23:08 Why AI is a pure amplifier that is widening the skill gap between founders 24:06 The "Retrenchment Phase": Why engineers are focusing on optimizing token costs 25:58 Final advice: Do this before raising a pre-seed marketplace check

DN #15: The AI Launch Playbook - Build Communities, Not Just Products (w/ Tim Haldorsson)

"You're not competing against AI, but you're competing against other people with AI." In this episode of DN, Niklas sits down with AI and Web3 founder Tim Haldorsson, who runs a 30-person growth studio out of Lisbon, Portugal. Tim breaks down the massive shift happening across the digital service industry, where solo operators armed with specialized AI agents are effectively replacing traditional 5-person marketing departments. He explains his exact framework for building hyper-personalized AI tools in Claude—including how he uploaded a database of 32,000 of his personal X posts and a library of 700 agency press releases to flawlessly clone his writing style. Niklas and Tim also discuss the vital importance of IRL (in real life) community building, citing how billion-dollar companies like Cursor and Anthropic are leveraging local cafe meetups to build impenetrable moats. In this episode: • The 1-Person Marketing Department: How a single strategist managing a PR agent, a social media agent, and a content agent can outcompete a legacy team of 5. • Cloning Your Brain in Claude: Tim’s exact method of feeding 32,000 historical tweets and 700 press releases into Claude to create the ultimate custom writing skill. • AI vs. Human Journalists: The shocking blind study by the New York Times where 80,000 readers rated AI-generated articles higher than human-written pieces. • The Offline AI Movement: Why massive AI labs are suddenly borrowing Web3's playbook and investing heavily in IRL community events and hackathons. • The Widening "AI Gap": Why consulting giants like McKinsey and ultra-adapted small teams will completely crush legacy service businesses that refuse to adopt AI. • Product vs. Distribution: Why early-stage founders should actually prioritize building an audience and a community before stressing over their product roadmap. • Lisbon's Tech Renaissance: Why 300 days of sun and a massive influx of international builders have turned Lisbon into Europe's ultimate hub for AI and Web3. 🎧 Full episode on all podcast platforms 💬 Do you think AI writes better than humans? Let us know in the comments! 🔔 Please like and subscribe! Every subscriber helps our channel grow. #DN #ArtificialIntelligence #Marketing #AIAgents #Claude #TimHaldorsson #Web3 #CommunityBuilding #Startups #Lisbon Timestamps: 0:00 Intro: Niklas meets AI & Web3 founder Tim Haldorsson 1:44 From solo freelancer to running a 30-person international agency 3:32 What the AI industry is actively borrowing from Web3 marketing 7:58 Cafe Cursor & Anthropic: Why billion-dollar tech labs are pushing IRL meetups 11:43 The Founder's Dilemma: Should you focus on product or community first? 17:46 The NYT Study: Why 80,000 readers preferred AI writing over human journalists 18:53 Using 32,000 old tweets to perfectly clone your writing style in Claude 23:52 The AI Gap: Why consulting giants will crush smaller legacy agencies 28:40 The divide between the "AI-pilled" tech hub and the rest of the world 32:44 How one person + four AI agents can replace a full marketing department 34:53 Automating competitive research & building an AI skill with 700 press releases 37:45 Advice for solo founders: Free GitHub agent repositories vs. hiring an agency 40:36 Why Lisbon has explicitly become the ultimate European hub for AI builders