Some proposed regs would cover uses of AI outside LLMs, some of which tech folks might call “machine learning” these days to distinguish them from LLMs.

Using algorithms to provide personalized pricing would be an example, where like a landlord, retailer, or airline would use an ML service trained on your personal data and aggregated purchase history to decide how much to charge you for a short-term rental, Nintendo Switch, or a plane ticket. Basically, instant underwriting at scale for every single purchase. Just got a new job with a raise? Your next vacation will cost you 26% more for the same experience.

▲

twodave 2 hours ago | parent [-]

This fundamentally doesn’t work unless there is collusion involved, which we already have laws against.

	▲	SpicyLemonZest an hour ago \| parent [-]
		Why doesn’t it work? It’s not obvious to me that a company with straightforward pricing would necessarily outcompete the algorithmic price discrimination one. They’d get somewhat more business from comparison shoppers who the algorithm feels can pay a lot, but lose business from people who the algorithm feels can pay less, and make less profit on everyone else.