Becoming unLLMable

But the more and better access everyone has to tools, the clearer it becomes that the final bottleneck to great work is not knowledge or information. Heck it’s not even intelligence—it’s that elusive, intangible, sublime quality—call it taste, imagination, creativity, courage, judgment, intuition, agency.

To shape the inputs—and to evaluate the outputs—you still need expertise. When you prompt an LLM, every word you input shapes what comes next.

The key thing to take away from this is that the real expertise here isn’t doing the work. It’s knowing how to guide the work. How to evaluate the work. It’s knowing what is worth prompting — which is really just another way of saying: knowing what’s worth doing.

On the right: what I would prompt to give it more pzazzz. A prompt that reflects my understanding of positioning and psychology. I provide the right context, define the tone and personality for the brand, ask the AI to generate multiple variations to choose from (this is key), and keep pushing and pressure-testing and giving feedback to the models ... See more

By which I mean, does this work carry the unmistakable fingerprint of human creativity, perspective, and lived experience. Of course the irony is I use LLMs to aid me in producing the work, but as a whole AI makes average attainable by anyone, and therefore raises the bar for what is considered exceptional.