Soundings
Our Perspectives and Insights
Limited Time Smart: Run a Deliberate AI Inversion
Anthropic’s Fable 5 is now rolled out to most plans, but when it was first introduced, it set in motion an exercise I am now doing routinely. If you remember from a month ago (eons in AI time), Fable 5 was introduced as free for a limited time window. That brought back the Simpsons episode where Homer removes a crayon lodged in his brain and becomes brilliant, or for older Transformers fans, the one where Grimlock gets smart and builds the Technobots. The question I had at the time was, “What should I do with this newfound intelligence?”
Call it limited time smart: a window where you have more intelligence on hand than you did last quarter, and a clock running on it. The reflex in most companies is to flip everything over to the new model and keep doing exactly what they were already doing, only faster. A jump in capability is not a reason to run harder down your current road. It is the rare chance to stop and ask whether the road was ever right.
A Faster Car on the Wrong Map
One of my first AI articles made the point: AI is not a solution, it is an accelerator. It amplifies good and bad decisions with equal enthusiasm. A smarter model is a faster car. Pointed at the wrong destination, all it does is get you lost sooner, and with more confidence.
So the worst possible use of a sudden capability jump is a pure speed upgrade: same strategy, same assumptions, same roadmap, now executed quicker. You will feel productive. You may even hit your numbers. And you will have spent a genuinely rare moment reinforcing a direction you never re-examined.
Point the New Intelligence Backward, Not Just Forward
The better use is deliberate inversion: turning your own world upside down on purpose. Question the assumptions. Make the case against your current path and see if it survives. A smarter model hands you a tool you did not have last quarter, and the highest-value thing to aim it at is not new work, but the work you already committed to. It takes guts, and a strong dose of humility.
Take the strategy you are running, the roadmap you already sold, the assumptions you never wrote down, and run them straight back through the new intelligence. Ask it where you are wrong. Ask it what a smarter version of you would kill. It will surface things your old constraints kept hidden, the shortcuts that made sense when the model was dumber, the plan that only held together because nobody had the bandwidth to pressure-test it. Then you fix them while the window is open.
It takes guts, and a strong dose of humility.
In the Simpsons, Homer decided he did not like being smart and put the crayon back. Grimlock used the borrowed intelligence to build something better that survived after his own faded. Which one are you going to be?
The best example in business history is not about AI at all.
In 1985, with Intel bleeding money on memory chips, Andy Grove turned to Intel’s co-founder Gordon Moore and asked what a brand-new CEO, one with no loyalty to the past, would do if he walked in the door. Moore said, without hesitation, that he would get them out of the memory business.
Grove’s response, as he tells it in Only the Paranoid Survive, was to ask why the two of them should not walk out the door, come back in, and do it themselves.
They did. Intel exited memory and reinvented itself as a microprocessor company. The forcing function that year was a crisis, but the move never required one. It required the nerve to question the map before something else forced the question for them.
How to Run a Deliberate AI Inversion (Before You’re Forced To)
The problem is not that we cannot question our direction. It is that we only do it when a crisis, or a shiny new model, makes us. And when we finally do, we interrogate the engine instead of the map. Here is how to make it a habit instead of an accident:
- Put it on the calendar, and give it a hard stop. Inversion never happens when it is convenient. Schedule it like a board meeting, on a cadence, whether or not anything is on fire. Limited time is a feature: an open-ended review just becomes another thing you postpone.
- Run everything back through the new model. Feed it the strategy, the roadmap, and the assumptions you are already committed to, and ask where they break. The jump in capability is most valuable pointed backward at decisions you already made, not just forward at the next task.
- Argue the opposite, on purpose. Make the strongest possible case to kill your current strategy. If it survives the attack, you have earned your conviction. If it does not, you fix what broke or you abandon it, and either way you just saved a year.
- Decide something before the window closes. An inversion that ends in a document is the Homer move with extra steps. The previous step hands you two outcomes, fix what broke or abandon it, and both are decisions that have to be made while the better tool is still in your hands. Name what you are changing, name what you are killing, and put a date on each.
An inversion that ends in a document is the Homer move with extra steps.
- Build something that outlasts the window. Do not just bank the answers, bank the method: the assumptions you finally wrote down, the questions that broke them, the criteria you used to decide. Next quarter’s inversion starts from that instead of a blank page. Grimlock did not stay smart. He built the Technobots and poured everything he had into them.
The Bottom Line
There is never a good time to run an inversion, and too often we only run one after the crisis has already happened. We have tools to help now, and we can lean on smarter models, but we do not have to wait for the next one to arrive. The limited time window on Fable 5 taught me I did not need it. The strategic value is in running the inversion as a regular practice, not as a reaction.
So do not wait for the jump. Put the inversion on the calendar, run your path against the best thinking you have, and then do it again next quarter whether or not anything is on fire. The leaders who win are not the ones who got smarter for a week. They are the ones who never stopped questioning where they were going.
And yes, I’m a fan of Transformers and Simpsons.
A Pilot & Rutter principle.