RCRyan Cinmyprompts2production.hashnode.dev·3d ago · 10 min readI Spent a Day Building an AI with a Soul. Here's What It Taught Me About the Difference Between Using AI and Mastering It.This past Friday I was scrolling LinkedIn and came across a short video — someone had shared a clip of Andrej Karpathy, co-founder of OpenAI and the person who built Tesla's self-driving programme, sp00
RCRyan Cinmyprompts2production.hashnode.dev·5d ago · 8 min readPart 3: How I Built an Enterprise Sales "Lie Detector" by Pitting Two AIs Against Each OtherSeries Intro: Domain First, AI Second In my previous post, Part 2: When the Sandbox Goes Blank—What BigQuery Taught Me About AI, Sales Forecasting, and the Limits of "Vibe Coding", I explored what hap00
RCRyan Cinmyprompts2production.hashnode.dev·Jul 2 · 9 min readPart 2: When the Sandbox Goes Blank — What BigQuery Taught Me About AI, Sales Forecasting, and the Limits of Vibe-CodingThis is Part 2 of an ongoing series. Part 1 covered how a four-week crash course in Tableau and BigQuery — using AI as an active dialogue partner rather than a passive tool — shifted my understanding 01M
RCRyan Cinmyprompts2production.hashnode.dev·Jul 1 · 6 min readPhase 3 — Implementing AI with Security - where theory meets the reality of getting things done The simulations pointed clearly in one direction: relying entirely on traditional, single-pass security checks to catch the logical flaws an AI can introduce is a losing game. The five-stage loop — Bu00
RCRyan Cinmyprompts2production.hashnode.dev·Jun 28 · 8 min readA Safer AI Agent Pipeline, Revisited — Adding the Constraint That Makes It HoldPhase 1 — where this started A few weeks of reading about AI agent security turned into a small, sandboxed setup — no real accounts, no real risk — built to test one question: if an AI agent is going 00
