Hey friends,
My feed this week has been one thing. Fable.
The new model everyone's posting about. Screen recordings, hot takes, "this changes everything" threads. And to be fair, it's genuinely impressive. I've used it. It's good.
But I want to say the quiet part out loud, because almost nobody posting those videos will.
If you run a business, the newest model is rarely the thing that matters. And if your business is built around any single model, you have a bigger problem than which one is winning this month.
Let me explain both, because this is the part that actually protects you.
You’re Probably Reaching For Too Much Model
Here's what I see inside real businesses, not demos.
Almost all of the actual work uses a thin slice of what these models can do. Drafting a reply. Pulling a summary. Sorting a list. Answering a common question. Tagging a ticket. This is the bread and butter, and a smaller, cheaper, faster model clears that bar without blinking.
The frontier models are extraordinary. But most business tasks don't need extraordinary. They need reliable, fast, and cheap at volume. Reaching for the most powerful model for everything is paying Ferrari money to drive to the shops. It feels impressive. It just isn't what the job needed.
There are real jobs that earn the top model. Complex reasoning, messy judgement calls, the genuinely hard stuff. Use it there. Just stop making it your default for work a simpler model would do for a fraction of the cost.
Don’t Switch On Hype. Switch On Data
So how do you actually decide?
Not from my feed, and not from a benchmark someone posted. The right model is the one that wins on your work, on your data, on the tasks you actually run. That's the whole test.
When a client asks whether they should jump to the new model, I don't argue for or against it. I say let's test it. Take your real queries, the actual messages your customers and staff send, and run the new model against your current one, side by side, same conditions. Score them. Then decide on the numbers, weighing quality against cost and speed.
Nine times out of ten, that one exercise kills the urge to switch. The tenth time, you switch with proof instead of a feeling.
Don't switch models on hype. Switch on data.
Build So The Model Barely Matters
This is the one that saves you real money and real pain.
If your business runs on top of a model, you're in good shape. If it's built around one, you've turned that model into a single point of failure. And models are the least stable thing in this entire space. Prices move. A better one lands next month. One of them even got pulled from availability for a few weeks earlier this year. If a single release or price change can break your operation, that isn't a model problem. It's an architecture problem.
The fix is to treat the model as a swappable part, not the foundation. Own the things that compound: your data, your workflows, the logic of how work actually gets done. Keep those yours. Let the model plug in behind them, so when a better or cheaper one arrives, you swap it in a day instead of rebuilding for a month.
The businesses I've built this way don't flinch when a new model drops. They run the test, and swap only if it wins. The launch that wrecks everyone else's week is a quiet Tuesday for them.
The Dumb Monkey Move
Rebuilding your entire setup around whatever model is trending, then doing it all again thirty days later when the next one lands. Chasing the model-of-the-week and calling it a strategy.
It feels like progress. It's a treadmill.
A Calm Takeaway
Fable is great. So was the last one. So will be the next one, and it's coming sooner than you think.
The people who win with AI aren't the ones sprinting to the newest model. They're the ones who built so calmly that the model barely matters. Right-size to the job. Test on your own data, not the hype. Keep the model swappable, so every shiny launch is an option for you and never an emergency.
Let everyone else chase the week. You build for the decade.
See you next week,
— Aamir
📲 Resources & Links
🎧 Listen to the Podcast Episode on: Spotify | Apple Podcasts | YouTube
📘 Book: The CEO Who Mocked AI (Until It Made Him Millions) by Aamir Qutub