AI's Limitations and the Role of Human Judgement — Stuart McClure Keynote, TIF 2026

Stuart McClure delivers a keynote at TIF 2026 tracing his path from surviving the 1989 United Flight 811 cargo-door failure to a career built on preventing the preventable, spanning Foundstone, McAfee, and Cylance, where he applied predictive AI to stop attacks before they happen. He argues that today's LLMs are "artificial knowledge" rather than intelligence: trained on the same data, they regress to the mean and cannot supply the wisdom, experience, mission, or accountability that only humans provide. He frames WethosAI's work as replicating how individuals uniquely think through Cognitive Twins, measured across known recall, likeness, and unknown recall at over 90 percent accuracy, so teams can move past hierarchical decision-making toward a collaborative "hive mind" where humans stay in the loop and remain accountable. A closing Q&A covers Orange County's shortage of risk-tolerant venture capital, how software engineers should adapt by mastering and explaining AI-generated code, AI cost overruns, burnout, and the rising importance of first-principles critical thinking.