{"version":"1.0","type":"rich","provider_name":"Acast","provider_url":"https://acast.com","height":250,"width":700,"html":"<iframe src=\"https://embed.acast.com/$/68357ec21b846c88bdcd7480/69baec926071e3bfbde371b4?\" frameBorder=\"0\" width=\"700\" height=\"250\"></iframe>","title":"139: Sturgeon's Vibe Code","description":"<p>Everyone’s got an opinion on vibe coding. Half the internet says it will save you £150,000 a year and you will never need a developer again. The other half says it is a dumpster fire of security holes and spaghetti code. Both camps are loud. Both camps are partially right. And lots of us watching the bunfight are just confused.</p><p><br></p><p>In this one, Tom and Corissa try to make sense of it — not by picking a side, but by honing some questions. Their lens is: bounded applicability. Some tools are good for some things. Nothing is good for everything. The hard part is knowing which is which.</p><p><br></p><p>Including-but-not-limited-to:</p><p><br></p><ul><li>Tom’s first proper foray into Claude Code, a surprisingly useful Python script it produced in 15 minutes</li><li>A practical rubric: the questions to ask before you let an AI mess with your stuff</li><li>The developer who reviewed and approved every chunk of code and then came back a week later to find chaos</li><li>The architect’s clever marketing stunt that was already broken by the time anyone tried it</li><li>Why the surface sheen of coherence stops you thinking critically</li><li>Dave Snowden’s quietly devastating observation about what we’re actually doing when we accept AI outputs</li><li>Vibe-coded accounting software ... a brief thought experiment</li></ul><p><br></p><p>This one's for anyone trying to navigate the hype — in any direction — without losing their critical faculties in the process.</p><p><br></p><p>Links &amp; references</p><p><br></p><ul><li>Theodore Sturgeon — science fiction writer; originator of Sturgeon’s Law (90% of everything is crap)</li><li>Children of Time by Adrian Tchaikovsky</li><li>Dave Snowden, cited via Corissa’s recollection of a LinkedIn post about AI</li><li>Episode 039: Bounded Applicability — the concept underpinning this episode’s diagnostic framework </li></ul><p><br></p>","author_name":"Tom Kerwin"}