It’s getting to the point that when people (clients, colleagues, journalists) reach out to me for my niche expertise, I have to resist asking why they didn’t prompt their way to a likely more comprehensive, more current, and better structured answer than I could possibly give. My impression is that, for now, the acceptability of a human expert’s opinion remains superior, if only because it maintains plausible accountability: if the advice was poor, they can rightfully shift the blame and claim “this expert told me so”. No such excuse if they prompted their way to an answer. In fact, I feel that they don’t even mind me using an LLM to structure my own reply, as long as I take accountability for the output. The second-order implication is that brand matters. People will still seek human expertise from recognized names (corporate or individual) that serve as a proxy for “I’ve done my due diligence”. Once the next generation takes over who trusts Claude / ChatGPT as a universal source of truth, the same way boomers might have looked at Encyclopedia Britanniæ, then human experts will be well and truly cooked. A generation is 20 years. We’re already 3+ years into this transformation. I give it until the end of the next decade at the very latest before all expectations for human expertise have vanished.