Published Work
Articles published on Substack and other platforms. Each piece documents specific experiments, failures, and findings from the field — the public record of what the CogniCuisine methodology looks like when applied to real problems with real AI systems.
When the Machine Blinked
Gemini’s safety filter flagged professional culinary language during a food photography session. What followed was a cross-platform comparison of AI accountability behaviors across seven models — and a documented record of how differently each system responds when challenged on its own outputs.
Read on Substack ↗The Meatball That Broke Two AIs
A real account of pushing ChatGPT and Claude past their failure modes on a cooking problem neither would solve at first. The piece documents the pattern Meyers-Lussier calls “the third quit” — the point at which AI systems either break through or collapse entirely — and what happens when a human refuses to accept the surrender.
Read on Substack ↗Can AI Recipe Generators Be Trusted?
An investigative look at the food safety and nutritional accuracy of AI recipe applications, drawing on peer-reviewed research and direct platform testing. The answer is more complicated than the marketing suggests, and the methodology for evaluating AI recipe output is more important than the output itself.
Read on Substack ↗Sweet Potato–Gochujang Meatballs with Coconut–Miso Sauce
A fully developed CogniCuisine recipe combining Korean and Southeast Asian flavor profiles, developed through systematic human-AI collaboration. The piece documents the development process alongside the final recipe — a working example of how the methodology produces results a human and AI working independently would not have reached alone.
Read on Substack ↗Stay Updated
New articles are published regularly on Substack. Follow CogniCuisine there to receive each piece as it publishes.
Follow on Substack ↗