What actually is a token?


[Day 4 of the four unlocks series. So far: the thinking partner and reading at scale. Today, the one people only use in private.]

At a workshop this spring, during the coffee break, a managing partner waited until nobody else was within earshot and asked me: “What actually is a token? People on my team say it forty times a day and I’ve been nodding for a year.”

He manages more money than some banks. And he’d carried that question around for a year because there was no safe place to ask it.

Every senior person has a version of this. The term you’d struggle to define in the meeting where everyone assumes you can. The domain a new board seat dropped you into. The thing your own company builds that nobody ever explained from the beginning, because by the time you were senior enough to be in the room, you were too senior to ask.

The cost of asking used to be social. That cost is now zero. AI is a tutor with no memory of your last basic question and no colleague to mention it to. It explains at exactly your level, in analogies from a world you already know, and it keeps going until the idea clicks, without a flicker of impatience.

Two things make the tutoring actually stick, and most people skip both. First, make it check you: have it ask you a question after each explanation, and don’t let it move on until you can say the idea back in your own words. Reading an explanation feels like learning. Explaining it back is learning. Second, make it quiz you like a colleague would: “give me three questions someone might realistically ask me at work, and grade my answers honestly.”

The fine print: this works on established knowledge, the kind that lives in textbooks and documentation. On breaking news and niche topics, the model answers with the same confidence and less accuracy. When a claim surprises you, ask for sources.

The prompt sets your level, your analogy world, the check-my-understanding loop, and a “test me” command: Personal Tutor prompt template.

Somewhere on your list is a topic you’ve been nodding about for a year. Twenty minutes tonight. No audience.

Alex

PS Tomorrow, the last unlock and the biggest one: delegation.

Alex Talks AI

As an AI Coach, Advisor, and Agent Builder, I help organizations and business leaders harness the power of artificial intelligence to boost productivity and streamline operations. I enable organizations to navigate the transformative landscape of AI, educating teams, identifying operational and strategic opportunities with AI and creating a framework for safe and transparent use of data in the organization.

Read more from Alex Talks AI

An ops leader I coach had a task hanging over her head. Not a hard one. A big, tedious one: 17,000 contacts scattered across six messy spreadsheets that needed to be merged, cleaned, and de-duplicated before they could move into a new system. She'd been putting it off for three weeks. Honestly, the company had been dreading it for closer to two years. Names split across three columns. Addresses split the same way. Phone numbers formatted six different ways. The kind of work that isn't...

Last week I needed current financials for 50 companies. Revenue, EBITDA margin, net debt. Into a spreadsheet. Sourced. I did what most people do: I asked Claude to go do it. It started strong. Companies 1 through 12 were clean. By company 30 it was getting sloppy — a margin that didn’t tie, a “2024” figure that was actually a 2023 restatement. By company 45 it had quietly stopped citing sources altogether. It didn’t get lazy. It got full. Here’s the part nobody explains. An AI has a fixed...

[Day 5, the last of the four unlocks series. So far: the thinking partner, reading at scale, the private tutor. Today, delegation.] A few weeks ago a client showed me the results of an experiment. He’d asked AI to research acquisition targets in his industry and gotten back a list so generic it could have come from a trade magazine’s top-50 ranking. His conclusion: “delegation isn’t there yet.” I asked to see what he’d typed. One sentence. “Find me interesting acquisition targets in [his...