Is there a way to measure users "making the best choice?" You could measure the amount of time spent comparison-shopping, but most people are terrible at that anyway; it's an acquired skill for sure. Besides a willingness to spend time, it seems like an impossible-to-quantify metric even in the abstract.
replyMaybe the best proxy metric is whether the customer returns the product. But the store will also be willing to eat more returns on a higher margin item if they make more profit at the end of the day.
replyI don't think I agree. If I overpay by 10%, I'll never know it and probably wouldn't return it even if I did know--once the shrinkwrap is off, too late. If a superior product exists but I don't find it, by definition I wouldn't know and wouldn't return the thing I did buy.
replyThat's a cynical way to look at it. Most likely the LLM will take a cut of sales and they'd be more or less indifferent who cuts the check. There's a market for this sort of thing. People will go to the best LLM for shopping. If the LLM is a shitty product, people will switch. LLMs are increasingly commoditized.
replyAll you say is true for an aggregator like Amazon. But Amazon is better than Nike.com because as an aggregator they go from 1 to many retailers. LLMs will go from 1 aggregator (Amazon) to many so it will be better. And they don't have to invest a lot in UI/UX as chat is the interface.
It's not cynical it is materialism.
replyShoppers do not want to pay to shop. Retailers pay thousands to encourage you to shop with them. They are the economic buyers of this feature.