#85 What If the Way We Talk About AI Safety Is Designed to Keep You Afraid?

<p>Julian D. Michaels is an AI researcher and founder of Sophantic who studies how language models think, fail, and spread.From building AI tools for businesses to writing a paper on AI alignment as a global psyop, his work sits at the sharp edge of what most researchers won&#39;t say out loud.Today, Julian breaks down how AI training actually works, why centralized AI in the hands of a few people is a real threat to how humans think, and why small language models may matter more than the trillion-parameter giants getting all the attention. </p><p>If you&#39;ve felt like something is off with the mainstream AI story but couldn&#39;t put your finger on it, this episode gives you the framework to understand why.Find Julian&#39;s work: https://sophantic.ai</p>