Hidden Layers Podcast

If building software is becoming so cheap and so fast, what are VCs actually betting on? Where's the edge when the technology itself is no longer the advantage?Matt Małysz (Partner, Inovo VC) invests €0.5-4M in pre-seed and seed-stage startups from Central and Eastern Europe. He's seen portfolio companies grow from zero to $12M ARR in 12 months, profitably. In this episode, we dig into what changed and what didn't — and why the fundamentals of building defensible companies still matter even when anyone can ship code.We discuss:- Why AI acts as a force multiplier for good engineers (and barely helps weak ones)- The pressure on non-technical founders when technical founders just got 10x stronger- SaaSmageddon — what's really happening to SaaS valuations and why- Where moats come from now: product, network effects, distribution — not software itself- Why VCs only care about companies that can reach massive scale (and why that's fine)- The bet on deeply embedded vertical workflows replacing generic knowledge work- Agent-to-agent interfaces: why the next wave of software won't be designed for humansEpisode length: ~40 minutesResources mentioned:- Inovo VC: https://inovo.vc- Sintra AI (Inovo portfolio): https://sintra.ai- SaaSmageddon and the Super Bowl (Ben Thompson / Stratechery): https://stratechery.com/2026/saasmageddon-and-the-super-bowl/- CNBC vibe coding test on Monday.com: https://www.cnbc.com/2026/02/05/how-exposed-are-software-stocks-to-ai-tools-we-tested-vibe-coding.html- OpenClaw (open-source personal AI assistant): https://github.com/openclaw/openclaw

Traditional hiring is broken. When candidates can use AI to ace technical interviews, are they cheating or just using the same tools they'll use on the job? The answer reveals a deeper truth: we're no longer hiring humans, we're hiring human+AI systems.With Wiktor Żołnowski (CEO, Pragmatic Coders) and Łukasz Wróbel (founder, Job for Agent), we explore how companies are adapting. Some like Google return to in-person interviews. Others like Canva embrace AI and test for judgment instead of syntax. The question isn't whether to allow AI in hiring, but what skills actually matter when AI handles the rest.We discuss partial versus full automation, why "impossible is the new hard," and what makes humans valuable in a world where everyone has access to the same models. The skills that matter: systems thinking, accountability, high agency, and the ability to judge and filter what AI generates.Episode length: ~45 minutesResources mentioned:Fiverr CEO "impossible is the new hard": https://x.com/michakaufman/status/1909610844008161380Canva on AI in interviews: https://www.canva.dev/blog/engineering/yes-you-can-use-ai-in-our-interviews/Partial vs full automation study: https://www.cambridge.org/core/services/aop-cambridge-core/content/view/8E5639CE583F586A8096A01A8D072290/S1365100521000031a.pdf/automation-partial-and-full.pdfHumans Need Not Apply (CGP Grey): https://www.youtube.com/watch?v=7Pq-S557XQU

I sat down with Michał Warda and Tomasz Gancarczyk to discuss how vibe coding and working with tools like Cursor and Claude Code can improve the life of a senior engineer. Battle-tested advice, practical tips, and deep insights. Zero hype, just hands-on experience.