If Part 1 delivered the simple truth — that AI is a tool, not magic — Part 2 delivers the harder one: that the people perpetuating the hype are not always cynical actors lying for profit. Many genuinely believe what they are saying. And that makes the problem considerably more difficult to solve.
Stuart McClure's second essay in the AI Exposed series examines the structural incentives that keep the hype machine running. Venture capital cycles reward bold narratives. Media attention follows superlatives. Career advancement in technology has long correlated with being associated with the Next Big Thing. These forces create pressure to overstate capability, understate risk, and paper over failure.
For security in particular, the stakes are acute. When a security product is overhyped, the customer may feel falsely protected. When an attacker uses AI tools that defenders have dismissed as "not real AI," the gap between narrative and reality becomes a vulnerability. Stuart has spent his career closing exactly those gaps — at Foundstone, at McAfee, at Cylance, and now at Qwiet AI.
The hard truth is also institutional: organizations that have invested heavily in AI narratives find it difficult to course-correct without admitting prior error. The result is a kind of collective fiction that persists long past the point when individual participants privately know better. Breaking this pattern requires leaders willing to say uncomfortable things clearly — which is precisely what this series sets out to do.
The essay is a call for epistemic honesty in an industry that often rewards its opposite, and for the kind of rigorous critical thinking that makes AI deployments actually work rather than merely sound impressive in a board presentation.