Where AI hits a wall #ai #tech #learning

The speaker argues that AI hits a fundamental wall when it encounters domain-specific knowledge that experts hold implicitly but rarely articulate as formal rules. The core tension is between AI output that appears correct on the surface and output that is genuinely correct in practice — and the speaker contends that experienced professionals are uniquely positioned to identify that gap.

The first example involves a strategy partner reviewing an AI-generated competitive analysis. The partner pushes back by asking where the firm's proprietary insight on customer switching costs is, pointing out that any firm with access to the same AI model could have produced the same framing. This critique highlights that AI tends to generate generic, replicable outputs, and that true differentiation comes from proprietary knowledge and perspective that the AI simply doesn't possess.

The second example features a loan officer rejecting a covenant tracking prototype. The officer explains that a debt service coverage ratio cannot be treated the same as a minimum net worth requirement, because the two metrics have entirely different monitoring triggers. This is a case of implicit business logic — the kind of nuanced operational knowledge that never makes it into a requirements document but is critical for building accurate financial tools.

Taken together, the speaker's argument is that domain experts add value not just by using AI, but by identifying and articulating constraints that AI cannot infer on its own. In doing so, they convert tacit professional knowledge into explicit rules, which is itself a form of intellectual contribution that AI cannot replicate.