In business, particularly in startups, there is an idea that a business must possess a significant enough moat in order to stave off potential competitors in a crowded marketplace. Often attributed to investor Warren Buffett, moat, is a term used to describe a company’s competitive advantage. Like a moat protects a castle, certain advantages help protect companies from their competitors.

The same thinking can be applied to our work world, where there are two kinds of people using our LLM tools. One type uses LLMs like fortune cookies or magic 8-Balls. They ask an average question they would have used search to answer four years ago, and instead of generating a collection of search results that they would have had to parse and critically examine, they get an LLM regurgitated version of a mediocre answer.

These people are coasting along right now, using Chat GPT, Claude, Microsoft Co-Pilot, and on and on for entertainment, planning their next trip, flooding the attention zone with AI slop, or just lurking about, wondering what to do with these tools next.

The second type of person uses LLMs to examine assumptions and to level up what they are already doing at work. They are interested in prompt engineering and search for context inside the LLM’s answers to queries, and when the answer is average–like an answer from a fortune cookie–they push the tool past its limits. These people know and understand how language, persuasion, sales, negotiation, and conflict work at deeply interpersonal levels between people in the real world, and they employ critical thinking because they have mental and emotional discipline.

They are building a bigger and bigger moat, one prompt at a time, that eventually will transform into a delta; that is, a concrete, measurable variable of change between two states: those who have a large moat, and those who don’t.

As the delta between these two groups increases, and one group declines while the other one ascends in the economy of the future we’re building the rails for now, which one of these two groups do you think will have a more defensible moat in their job, their career, their financial situation, and even in their cultural life in America, as we breathlessly outsource more and more of our average outcomes to these new machines we’re building?

And here’s another question while you’re pondering the answer to that one: Which group of people will complain–or advocate if you will–for more fairly distributed outcomes as rewards begin to accrue to one group over the other economically and socially?

My advice to all of you reading: Expand whatever moat you’re working on right now into an ever-growing delta.