In episode four of Train.Brain.Daily, hosts Morry Morgan and Michael “Hewi” Hewitt-Gleeson explore the difference between language, intelligence and genuine thinking. They begin with Louis Pasteur’s idea that chance favours the prepared mind, linking preparation to military training, business performance and the rapid development of their podcast. Hewi argues that organisations cannot expect better results from occasional seminars or instructions; minds become capable through repeated training.

The discussion then turns to large language models. ChatGPT, Claude and similar systems can produce fluent, polite and convincing language because they have been trained on enormous amounts of data and predict what word should come next. However, Hewi stresses that “languaging” is not the same as thinking. Humans often behave similarly when they repeat clichés, media talking points or opinions reinforced inside social-media bubbles. Real thinking requires escape: curiosity, lateral thinking, experimentation and the creation of new value.

George Gallup becomes a key example. Gallup, who examined Hewi’s PhD, measured human language by asking carefully structured questions and analysing representative samples. Hewi compares this process to an early large language model: Gallup “prompted” people, collected their verbal output, identified patterns and used those patterns to predict something much larger, including presidential-election outcomes. Modern AI predicts the next word; Gallup sought to predict the next president.

The hosts connect these ideas to EBNE—Excellent But Not Enough. An answer may be correct, useful or impressive, yet accepting it as final closes thought. EBNE keeps thinking open by asking, “What else?” Morry links this principle to improv comedy’s “yes, and,” while Hewi connects it to X10 thinking: ask AI for ten options, then examine the best few rather than defending the first answer. This turns AI from a replacement for human thought into an additional mental resource.

Their broader warning is that AI may make people more knowledgeable while also trapping them inside increasingly sophisticated versions of their existing beliefs. Used passively, an agreeable assistant can become a digital yes-man. Used deliberately—with requests for alternatives, criticism, simpler solutions and better possibilities—it can expand human intelligence. The episode concludes that AI should handle repetitive “basement” work, freeing people to move toward the “penthouse” activities of discernment, innovation, humour, curiosity and value creation. In every case, excellence is acknowledged, but the search for what comes next must continue.

Enjoying this podcast? You might also enjoy:

• This Week in Comedy
• Fool Me Twice

Hosted on Acast. See acast.com/privacy for more information.