You’ve probably noticed my repeated ridicule of companies who ONLY talk about their use of AI. And my concurrent wonder about how car companies never talk about “tire intelligence,” or “TI.”
After all, tires are critically important to car companies, just as AI is critically important to tech companies. But car companies don’t hammer their prospects by interjecting “TI” into every conversation and advertisement.
So it’s time to delve into tires…and one important aspect of the tire industry.
Sailun and retreads
“Retreading has been part of the commercial tire industry for decades, yet for some fleets, hesitation still remains.”
And for individuals. I am personally reluctant to buy a retread.
But Sailun, a tire manufacturer, believes my reluctance is outmoded.
“Concerns around safety, consistency, and durability are often rooted in outdated perceptions of how today’s retread process works. In reality, modern programs follow controlled, standardized procedures shaped by engineering, inspection protocols, and widespread industry adoption.
“Today, nearly 90% of large trucking fleets rely on retreads, and an estimated 44% of commercial truck tires currently in service are in their second life or beyond.”
For details on how some tires (cough, cough, Sailun) are ideally suited for retreading, read the article.
Now let’s delve into AI.
Your favorite LLM…and retreads
Is there an AI equivalent to retreads?
The question initially sounds silly. Every AI prompt is brand new.
But there’s a cost to always getting new responses. A huge cost.
“An anonymous AI consultant told Axios that one of its clients accidentally spent half a billion dollars in a single month because it never bothered to put a usage limit on employee access to Anthropic’s Claude. That is… a lot. Like to the point of straining credulity.”
Even big companies don’t have half a billion dollars to waste on tokens. They would have to fire EVERYBODY, including all their executives and their boards of directors, to fund that level of AI use.
But before a company does something as drastic as actually firing an executive, perhaps it could look at a “retread” strategy.
Take the knowledge gained from AI prompts and store it for future reference without requiring tokens to be burned.
(Bredemarket does this all the time, even though I’m not charged by the token.)
Admittedly the re-use (retreading) of old information lessens exposure to new, updated information. But if AI is getting dumber anyway, that’s a good thing.
Bredemarket 