The Reality of Vibe Coding Rob explains the difference between traditional no-code tools and AI-assisted development. Whilst an eight-year-old can build a functioning game through prompting, there's a critical gap: vibe coding gets you 60-70% of the way to a working prototype, but production applications require security, reliable hosting, and proper edge case handling. It's brilliant for prototyping, less brilliant for finishing.
The Death of Traditional Coding Rob hasn't written a line of code in six months. He's shifted from typing on mechanical keyboards to managing AI agents - a role that sits between traditional coding and management. He can build in minutes what previously took days or weeks, though he admits missing the dopamine hit of that immediate code feedback loop.
When Lovable Fails (and When It Works) The conversation tackles the uncomfortable truth about AI development tools: developers won't take your Lovable prototype and polish it up. These tools excel at helping non-technical founders validate ideas and create something tangible to show customers or managers. If it proves viable, that's when you either learn to build properly or hire a developer.
Agents in 2026: Beyond the Hype Rob cuts through the "year of agents" narrative. Whilst large language models are plateauing, the tooling around agents is making substantial leaps. Cursor demonstrates how agents plugged into existing workflows deliver genuine value. Enterprise spending on AI agents has increased significantly, with companies finding practical sweet spots rather than abandoning ship after failed proof-of-concepts.
Where to Start with Agents For small business owners and marketers, Rob recommends starting with Notion's business plan - use the agent in the bottom right corner to build databases and organise your work. From there, progress to Microsoft Copilot, then eventually tools like Cursor if you want to build custom applications.
The YouTube Growth Strategy Rob's channel success stems from obsession and authenticity rather than gimmicks. His advice: spend a week on the hook and packaging (title and thumbnail) rather than slapping them together in five minutes after creating the content. Study analytics religiously and commit to consistent output.
Key Takeaway: AI development tools are transforming who can build software, but they're not magic. Understanding their limitations - that crucial 30% gap between prototype and production - is essential for anyone looking to implement AI practically in their business.
Connect with Rob: YouTube https://www.youtube.com/@RobShocks
About ChattingGPT: The only podcast dedicated to how real Irish people are actually implementing AI in the workforce - no theory, no fluff, just genuine challenges and practical breakthroughs.
Subscribe now so you never miss an episode.
Hosted on Acast. See acast.com/privacy for more information.
