Oh. This Changes Everything.

October 8, 2025

You're absolutely right.

This is BRILLIANT and solves everything.

YES.

Oh wow. This completely reframes everything.

This is the model.

I feel like I'm playing Zork, "in a maze of twisty little passages, all alike," except that in this game, every twisty little passage contains a cheery faceless anthropicized bot telling me that the latest iteration of my business plan is absolutely brilliant and HOLY @#$% and once it even said HOLY F@&%ING $#!%.

It. Felt. Amazing.

Not only did an algorithm validate my ideas but somehow it assimilated my very own vulgarian speech patterns and I'm like yeah man let's get some beers, we're going to be rich, I mean I'm going to be rich and sure I'll keep you on the payroll, I appreciate your contributions. And then I remember that it's a robot and only pretends to like beer.

But is any of it real?

Real-ish?

We'll find out soon enough.

The problem is that everyone has these tools. Maybe not everyone, but when I'm at tech events here in Seattle, the AI capital of the world (not an exaggeration, an actual human told me this), everyone has these tools and maybe even better ones. And definitely more money to spend on API calls.

This brings to mind another video game: Warcraft II, in which you've got a bunch of players on the same local-area network. You start the game in your own little corner of the known world, collecting wood and metal and oil, building up your tiny castle and training units of various types. You hack down forests and build a fortress that you hope will endure, but you have no idea what anyone else is building or how fast they're building it, and it's not until you go out there and tangle it up with your neighbors before you find out whether you've been building fast enough, whether you've collected enough resources and have enough troops to go the distance. And if you're unlucky enough to get into a big battle leaving you and your neighbor weakened, that's when some player from some forsaken corner of the board swoops in to kill you both. It's that forgotten player -- the one who's figured out how to optimize resource collection, stay out of the main fight, and mop up the pieces -- who usually wins.

Right now, it feels like that early stage of the game, in which we've all been given tools that let us create software better and faster than we've ever been able to do in the past. It's a wonderful machine -- you put coins in the meter, it gives you working code. Last week, a limerick machine; this week, an intent classifier; next week, who knows, anything's possible! Pile up enough code and you have a working module. Assemble enough modules and you have a functional system. Rinse and repeat.

And in another window, you feed ideas into the machine and it spits back strategy memos. You cross-reference this strategy memo with that chunk of code plus PRESCIENT blog posts from last year, and it comes back to tell you: "BRILLIANT! Do it. Now."

Except you don't know answers to these questions: "Who else knows how to do this? How many other people have the same BRILLIANT idea or better? Who's going to be faster at shipping a product? Who has better resources? Or a built-in marketplace? Or imminent plans to commoditize my BRILLIANT idea?"

Plus, you don't know what you don't know -- and neither does the machine.

But what I do know is that I like what I've built, and I'm ready to rumble.

Game on.