According to Ethan Mollick on X, he flagged a Nature interview and book review where AI pioneer Luc Julia argues modern AI systems are little more than glorified pocket calculators, prompting debate about how well this view fits 2026 capabilities; according to Nature’s review, Julia emphasizes statistical pattern matching over understanding, cautioning against hype, while many 2026 deployments in copilots and generative search suggest growing practical impact. As reported by Nature, Julia’s position urges businesses to focus on measurable utility and reliability rather than anthropomorphizing models, which in 2026 translates into opportunities in narrow, high-ROI workflows such as code assistance, customer support summarization, and document automation with controllable outputs. According to Nature, the takeaway for enterprises is to invest in evaluation, guardrails, and domain data to convert pattern recognition into dependable products, aligning with current trends toward retrieval-augmented generation, model distillation, and enterprise-safe deployments.