One of the things I like to do with AI is make stuff less boring, and I feel this application of this new technology doesn’t get enough attention.
My co-founder Rich and I discussed one example on this week’s podcast: A month or so ago I uploaded the Project 2025 900-page PDF (extremely boring), then instructed it to answer all questions in the voice of an enthusiastic squirrel (hilarious, terrifying). When I asked about school lunches, it came up with a list of topics like “local nut control” and “cost-effective crumbs.” It used a bunch of squirrel and nut emojis, too. And as a result, I spent about three hours going deep on Project 2025 and learned a ton, albeit in a sort of squirrelly fashion. It was a really fun way to give myself a panic attack.
Another example: I never really learned the truly detailed rules for American football. I mean I know the basics but the details just wash over my brain like a stream over a bunch of rocks. So I prompted Claude, and it gave me a bunch of bullets like “two teams of 11 players each.” Still boring. So I typed: “Now describe them as if it’s an ancient war between tribes of elves trying to capture a magical orb.” It produced:
Pixies known as "referees" oversee the ritual, ensuring all elves adhere to the ancient laws. Transgressions against these laws result in the offending clan being pushed back, making their journey more arduous.
The clans can earn favor from the gods in several ways:
Portal Crossing (6 points): When a Seeker carries the orb through the opponent’s portal or catches it within the portal’s boundaries.
Sky Kick (3 points): A Seeker may attempt to kick the orb through the Y-shaped branches above the portal from anywhere on the field.
Safety Blessing (2 points): Granted when Guardians tackle a Seeker within their own portal.
And so forth. I then asked it to explain the soccer offside rule in the same fashion, which produced pages of context like:
Picture, if you will, an invisible veil that stretches across the Verdant Field, always aligning itself with the position of the second-last defending sprite. This veil moves fluidly as the defenders shift their positions, like a shimmering curtain of mist. Now, here's where the magic truly manifests: An attacking sprite may never cross this veil before the magical sphere does. To do so would upset the delicate balance of the ritual, invoking the ire of the Spirit Watchers (known to mortals as referees).
This confirmed my suspicion that the offside rule is basically mystical nonsense.
A final example: I wanted to learn how a sine wave oscillator works. I have books with this information and of course there’s Wikipedia. But I asked ChatGPT, “Tell a fun story of the circuit from the point of view of some electrons traveling through the different components and what happens.”
Our story begins at the power supply, the grand entrance of Circuittopia. The electrons, filled with energy, eagerly set out on their journey. They were divided into two groups: those heading towards the positive terminal (+V) and those destined for the ground (GND).
You get the idea. In some ways, this pattern—“What can I do to make this less boring?”—is a good way to think about generative AI. Not “how do I skip writing this essay” but more “how can I turn a very dry subject into a story?”
You can then stretch this way of thinking a little further, and ask: “What can I do to make this more useful?” Useful implies an audience greater than one. You can turn 1,000 web pages into data extracted from those web pages, for example, in a structured format that a computer can display. You can turn thoughts and prompts into lists, stories, images, code, videos, and data. Then you can turn those things into other things, forever. You can’t always trust the output, but you can learn a lot by trying.
As we think through AI at work, this is where I find myself landing most often—the function of AI is not simply to generate stuff, but to take things that are less useful and make them more useful. Focusing on utility is helping me be more creative.