AI coding hype vs. the reality of worsening software quality
Nothing works and everyone is euphoric
Piotr argues that despite ever-improving models and the agentic coding hype, everyday software keeps getting worse. He cites recent personal bugs—banking app FaceID loops, Slack stealing focus, a crashing car infotainment system—and pins the blame on KPI-driven teams that never prioritize stability. The post doesn't offer quantitative data, but the core claim is clear: until orgs dedicate time to fixing bugs over shipping features, the quality decay will continue.
Why it matters: The author lines up a week of personal failures, bank face scans, a mis-sent Slack message, a frozen car screen, against rising model capability and falling software quality, and blames KPI-driven team culture.