Canva is the most recent to experience higher-than-expected costs from leaning hard into AI. Founders seem especially cursed by seeing the quantum of what will come later, today. They overestimate the short-term effects while the wider market has yet to grasp what the addition of a reasoning machine to the toolkit is really going to mean.
These visionaries, along with the resources businesses hoping to attract the attention and capital flows now following AI (Rio Tinto being the current test case), have advocated widespread adoption across every team.
You have to admire the pitch they bought.
I want you to pay for this product with a unit I invented and, in every way, defined. I'm calling it a token. A token is about two-thirds of a word, if that helps. You'll use them in what we call a model. The very best models will consume HEAPS of tokens. You won't be able to see or understand what happens inside any of them. Even our smartest people can't explain how the technology works. The only way to find out what this thing can do is to try it. We recommend trying everything. Here is a gigantic bill.
The new tools provide reach, most of all. Without learning or struggle. You are instantly further than you have ever been before.
Strategies that ask for broad adaptation of tools for any use end up with a smouldering slag of tokens where there used to be a mountain of cash.
Behaviourally, we are all suggestible monkeys. Distracted maybe, curious at best. Undirected, it's normal to reach for the darker shelves. The ones we've wished for and know little about. Where our finesse is least. Sometimes it even feels like you've been lured beyond your depth.
Maybe these bills are the measured cost of sating human curiosity. Or at least the fraction of it business is willing to fund before getting back to business.
Direction is the discount.
Chinese open-weight models too, probably.
Plan B helps businesses direct AI toward the specific work that compounds. Get in touch.
