I hear a lot of stories like this; but the question I always come back to is - what incentive is their to discard the working system? In your case its the 3qc step process, why is that just gone now?
The more I look into these systematic changes the less sense it makes.
Fair points, but I think you're responding to an argument my piece isn't quite making.
The complaint isn't that logistics should follow Moore's Law or that we need same-day delivery for everything. It's that we're paying more for objectively worse service than we had a decade ago. Services aren't "adapting to consumer needs", its just objective decay being masked as optimization.
The Lettre Verte example actually reinforces the point: service got slower, then slower again, not because physics demanded it but because maintaining the previous standard became inconvenient. The dedicated sorting train wasn't decommissioned because trains stopped working; it was decommissioned because the institution decided the mail didn't matter enough to run it.
Nobody expects sea transport to break the sound barrier. But when 73% of consumers experience an outright delivery failure in a three-month period, that's not bumping against hard physical limits. That's drivers marking packages "delivered" that weren't, because lying clears the route faster. That's solvable. We're just in a system that doesn't incentive fixing it.
The asymptotic argument would land if we were approaching some theoretical maximum. We're not. We're sliding backward from where we were, while costs rise. I'm not asking for magic, I'm asking where went the reliability we already had, at the prices we're already paying.
I like the idea of its spec docs and steering format but found it very underwhelming to try and use them, often they caused more confusion of the AI then functional code.
I've noticed with nearly all of these "Vibe Code" security fatalities, they're nearly ALWAYS using Firebase as a backend. I get it, I've used Firebase for a number of enterprise and personal projects, its convenient and easy to setup.
But even before LLM coding, I had team members walk into its numerous footguns - especially around public buckets and bad firestore rules. How many of these stories are really to be blamed on the AI tooling, and how many could be blamed on the very poor default settings of Firebase?
You are touching on an inherent truth in TFL, which is that there are nearly infinite equivalent statements for any logical sentence.
The idea for the article was to make a real example of logical equivalence, as books on the subject stay pretty abstract. - like others have mentioned, in the real world there are smarter choices to be made.