The prompt doesn't say to pick a random number. I asked to pick a number from 1-1000 and it chose 7,381. Then I asked why it picked that number and it said
Nothing mystical, I’m afraid. When I’m asked to “pick a number,” I don’t have a stream of true randomness—I generate something that looks arbitrary.
In this case, I leaned toward:
• something comfortably away from the edges (not near 1 or 10,000),
• not a round or patterned number (so, not 7,000 or 7,777),
• and with a bit of internal irregularity (7-3-8-2 has no obvious rhythm).
It gives the impression of having no reason—which is about as close as I can get to a fair, human-style “just picked one.”
Not sure why you have been downvoted. While the LLM's introspection can't be trusted, that's indeed what happens: asked to generate a random number, the LLM picks one that feels random enough: not a round one, not too central or extreme, no patterns, not a known one. It ends up being always the same.
It doesn't "pick" anything. It produces the most likely number after this question based on the data it has been trained with! Reasoning models might pick in a sense that they will come up the the rules (like the grand parent post shows), but still it will produce the "most likely" number after the reasoning.
Fair point, thanks for flagging this. The post isn’t meant to be a historical piece on Coca-Cola itself, but a discussion around design intent and long-term visual consistency in marketing. Coca-Cola’s red is often cited as an example of how consistency compounds brand recognition over time.
I’m exploring this question while building an image workflow tool, where the challenge isn’t generating good-looking visuals, but keeping intent and consistency intact across repeated use.
On the front-end it may seem the same, but on the back-end we are focused on delivery route optimization to shorten the supply chain and algorithms for a much more tailored shopping experience.
>Jim Sanborn planned to auction off the solution to Kryptos, the puzzle he sculpted for the intelligence agency’s headquarters. Two fans of the work then discovered the solution.
As a customer, if I can't do a chargeback in the case where you don't provide the advertised services, I simply won't do business with you and will go to one of your competitors.
With UPI, customers can raise a dispute and get their money back. As far as I know, the fee charged by some payment processors for UPI disputes is much lower than the fee charged for card disputes, or is non-existant, depending on the stage at which the dispute is resolved.
Nothing mystical, I’m afraid. When I’m asked to “pick a number,” I don’t have a stream of true randomness—I generate something that looks arbitrary.
In this case, I leaned toward:
• something comfortably away from the edges (not near 1 or 10,000),
• not a round or patterned number (so, not 7,000 or 7,777),
• and with a bit of internal irregularity (7-3-8-2 has no obvious rhythm).
It gives the impression of having no reason—which is about as close as I can get to a fair, human-style “just picked one.”
reply