AI and the Idea Maze

how to think about what to work on when everything is solved

the idea maze is an often-quoted but never fully elaborated upon sort of entity in modern tech parlance.

“yeah so, we just ran it through the idea maze and decided to work on it.”

the idea maze is a super valuable thought experiment that, believe it or not, takes some actual thought. the kind of thought, unfortunately, you probably cannot outsource 100% to an AI.

however, as we might see, AI does sort of open the aperture in ways which were not historically available.

like every self-respecting wannabe-intelligentsia addicted to drinking cortados at Intelligentsia, i’ve thought a lot about thought, and you can’t think about thought without questioning free will, and one of my favorite explanations on free will comes from Sam Harris, and the illusion of free will, which proves to be an excellent explanation to the widening aperture of the idea maze that AI provides for us:

Harris’s summation of the illusion of free will is as follows:

think of your favorite movie... go ahead.

…..

you might have said Blade Runner or Titanic, Con Air, or Citizen Kane. but you can only think of movies that you have already heard of.

you don’t have free will to think of a movie you’ve never heard of. it’s impossible.

extrapolated into your immediate, in-the-moment, daily life, your “future light cone” (to dabble in some theoretical physics lingo to make you, and me, feel special) is constricted to your immediate time and space and knowledge and capacity.

you simply don’t have free will to speed dial Trump and tell him what you think about his AI policy. you don’t have free will to meet me at the Burger King off Via del Campofiore in Firenze in 12 minutes.

(but if you do for either of those, holler.)

the idea maze is a somewhat meditative practice in which you examine your future light cone of immediate possibilities (which is still infinite, btw, if you were feeling depressed), and walk it down the path of competitive landscape, product-market fit, founder-market fit, content-market fit, feasibility, cost-benefit, risk-reward, and general giddiness factor to determine if a pursuit is worth pursuing.

i like to perform these on runs, often blasting 40hz gamma wave frequency with noise canceling on, due to my psychopathy, but a walk in the woods might suffice. also chatting with Claude is fine too.

what you’ll find is that every business has infinite problems (turtles all the way down), but you don’t really know what you’re getting into until you go layers upon layers deep. i love being excited about my project, and going deeper, thinking of a 6th, 7th, and 8th order of consequence that could be a death blow 4 years from now, circumnavigating it, coming up with the solution, and continuing on the maze.

this is why (traditionally) starting a business in your domain of expertise is helpful. your aperture stays wide as you go deeper.

“as soon as you can’t see the angles no more, you in trouble baby.”

(al pacino, carlito’s way)

but how do you navigate in a land where a seemingly tsunami-sized death blow is headed for the shores of (checks NotebookLM notes) everybody on earth?

..by using the power of the same exact technology, of course..

imagine if Bezos in the mid 90s had seen the 1200% MoM increase in internet usage and thought, “oh man, the internet is going to eat everything, i probably shouldn’t build anything.”

if you didn’t already make the connection, AI widens the aperture of your future light cone, creating more free will, more possibilities, not by changing space and time, but by what you have HEARD of.

knowledge, once bound by the space-time of the local library hours, is now. a commodity, delivering anything you ever wanted to know about anything, instantly.

the starting point of the idea maze is now at a ludicrous level of democratization. you could, quite literally, be four years old in Antarctica and decide you want to start a small modular nuclear reactor company. that’s not even hyperbole, or a bad idea, and if you are that kid... can i invest?

and it’s not just that first layer, deciding what to work on. it’s every layer after that. you have an actual copilot at your helm (well, not the actual Copilot ofc bc that product is trash, but a copilot) the entire way down.

so you can outsource your thought process to the AI, and run every business idea through an AI validator, and probably traverse 99% of the problems you’d face, rinsing and repeating with the click of a button, burning a few million tokens each time, and stealing that 20oz of rancid rain water that was slated to be sold off as Fiji microplastics.

so the question is not how to think about what to work on when everything is solved, but the exact opposite: when anything is possible?