Chevan Nanayakkara

The revolution is the protocol. In advertising, protocols solve coordination problems at scale. For example, before RTB, buying from 50 publishers meant 50 different relationships and custom processes. RTB created a standard auction language—suddenly you could access thousands of publishers because everyone spoke the same way.

The primary factor enabling AI psychosis is a mismatch between how easy AI systems appear to use and how difficult they are to use competently. This creates a dangerous overconfidence gap where people believe they’re operating sophisticated cognitive tools effectively while potentially degrading their own thinking capabilities.

We already tried making humans into machines with SaaS workflows and ticket queues. Everyone lost. And now the real machines are here. AI can do a lot of things better than humans but here’s what it can’t do- and never will.

A generative human is someone who uses AI to amplify their capacity to create original insights, novel solutions, and new value rather than just consuming AI’s outputs. Generative humans become more capable, more creative, and more valuable over time because each interaction with AI enhances their thinking abilities rather than replacing them.

Since OpenAI’s GPT-5 model started rolling out last week, people ask me the same thing: “How do I use AI better?” The answer is simpler than you think, but it requires a shift in how you approach the problem. Instead of focusing on techniques and prompts, you need to identify your use case first.

I usually write about big-picture paradigm shifts, but lately I’ve noticed something: while we’re all living through the AI revolution, many people are struggling with the basics—understanding what everyone is actually talking about.

AI systems face a fundamental paradox. They depend entirely on quality information for training, yet they’re systematically destroying the economic infrastructure that creates that information.