Microsoft spent the past year telling Americans that their next laptop had to be an "AI PC." The pitch was simple: buy a machine with a Copilot Plus badge, get a dedicated neural processing unit, and unlock a future of on-device intelligence that ordinary computers supposedly can't touch.
The marketing worked well enough that shoppers started assuming anything without the badge was already obsolete.
The flagship feature, Recall, was supposed to be the reason to upgrade.
It quietly screenshots nearly everything you do, then lets you search your own history like a personal surveillance archive.
Security researchers found the data poorly protected almost immediately, and Microsoft yanked it from launch, delayed it, and rebuilt it behind Windows Hello and encryption.
It finally shipped months late, opt-in, and buried in settings.
A headline feature that arrives apologetically is not a feature.
Meanwhile, the Copilot key that Microsoft strong-armed onto new keyboards has become a running joke.
Pressing it often just opens a chatbot that can't reliably control your settings, summarize your files, or do much beyond what a browser tab already did.
The NPU inside these chips sits idle for most tasks, because the software that supposedly needs it barely exists.
You paid a premium for silicon that mostly waits.
This is the pattern that should worry anyone paying attention.
Companies no longer sell computers on what they do today.
They sell them on a roadmap, a promise, a keynote slide.
Then the roadmap slips, the features get watered down for safety and legal reasons, and the customer is left holding hardware designed around software that never fully arrived.
It's the same playbook as 3D TVs, VR headsets, and the metaverse: manufacture urgency, then let the buyer absorb the disappointment.
The societal cost is subtler than a bad gadget.
When "AI" becomes a sticker slapped on everything from laptops to washing machines, people stop believing any of it.
Genuine advances get drowned out by noise.
And the companies doing the loudest promising are training an entire generation of consumers to assume that new technology is mostly hype with a financing plan attached.
The everyday impact lands hardest on ordinary households.
A family stretching to spend $900 on a laptop because they were told it was future-proof just bought a machine that runs the same spreadsheets as a $500 one.
The extra money bought a badge, a button nobody presses, and a promise that keeps getting postponed.
None of this means on-device AI is worthless.
Running models locally has real privacy and speed advantages, and the hardware will eventually matter.
But "eventually" is doing an enormous amount of work in that sentence, and you're the one financing it upfront.
Our take: skip the badge and buy the laptop that's actually good right now.
Final Thoughts
If a machine needs a marketing campaign to justify its price, the campaign is the product.