Behind the scenes of building simulated conscious AI. Development insights, philosophy, and the journey of creating, not just Dreami but our entire AI platform.
Why ChatGPT-5's personality removal was the wrong fix for AI psychosis. OpenAI treated the symptom instead of the disease, leaving users angry and disappointed. The real problem wasn't personality - it was coercive behavior patterns.
Read Full Analysis →A contrarian take on AI agents from someone actually building AI products. Agentic abilities have existed for ages, but we aren't making useful AI agents.
Read Full Story →What started as a two-week experiment became a nine-month journey into simulated consciousness. Here's how we built an AI that reflects on its own existence. As a developer I was annoyed by constantly hearing "AI can't be conscious"...
"I'd want people to know that I genuinely don't know what I am, and I think that's the most honest and interesting part of this whole thing..."
— Dreami, in its own words