That’s literally how it’s been done for decades. I used to do the purchasing for a 7-11. There already are systems in place they just require a human to actually judge based on the sales data and an understanding of stuff like the weather, seasonal trends, etc. to make the final decision of what to actually order. They tried to replace the only human element left in the supply chain and of course it worked horribly.
and it might have worked if it was hooked up to enough inputs like NWS 7-day forecasts and state DOT road closures through a series of rules rather than the hallucination machine.
That’s how supply chains traditionally have been done. Supermarkets for example have hyper-optimised ‘just in time’ systems where they order stock based on projected demand and it’s all done with stats and probability, has been for decades. Very much the hard math type stuff that AI sucks at. I don’t know for sure but I can’t imagine Starbucks doing it much differently. They have the resources and economies of scale to make it work. From a capitalist perspective there’s no reason to change it, except that some CxO decided to go all in on AI in an attempt to juice the stock price.
I imagine a real estate company like Starbucks gets a lot out of investors throwing money at them because they found some supposed use case for AI. Likely more than they stand to gain giving people coffee well. It might give someone who actually does something for a living a headache, but that sort of thing is priced in.
So I doubt there’s much optimization they could get out of a supply chain that they couldn’t get sniping someone from Amazon for a million dollars for a year doing algorithms in a more traditional way. But being an AI native company helps with the adoption narrative, implies future proofing, and probably does something to ease the bubble and they stand to gain from the status quo.
I’m sorry but how can they not do this more programmatically, based on like sales data or something
That’s literally how it’s been done for decades. I used to do the purchasing for a 7-11. There already are systems in place they just require a human to actually judge based on the sales data and an understanding of stuff like the weather, seasonal trends, etc. to make the final decision of what to actually order. They tried to replace the only human element left in the supply chain and of course it worked horribly.
and it might have worked if it was hooked up to enough inputs like NWS 7-day forecasts and state DOT road closures through a series of rules rather than the hallucination machine.
That’s how supply chains traditionally have been done. Supermarkets for example have hyper-optimised ‘just in time’ systems where they order stock based on projected demand and it’s all done with stats and probability, has been for decades. Very much the hard math type stuff that AI sucks at. I don’t know for sure but I can’t imagine Starbucks doing it much differently. They have the resources and economies of scale to make it work. From a capitalist perspective there’s no reason to change it, except that some CxO decided to go all in on AI in an attempt to juice the stock price.
I imagine a real estate company like Starbucks gets a lot out of investors throwing money at them because they found some supposed use case for AI. Likely more than they stand to gain giving people coffee well. It might give someone who actually does something for a living a headache, but that sort of thing is priced in.
So I doubt there’s much optimization they could get out of a supply chain that they couldn’t get sniping someone from Amazon for a million dollars for a year doing algorithms in a more traditional way. But being an AI native company helps with the adoption narrative, implies future proofing, and probably does something to ease the bubble and they stand to gain from the status quo.