Asanearlyonline-onlyretailer, Amazonhad a uniqueabilitytocollectmassaggregatedataonshoppingbehavioranduseittocreatealgorithmstomaximizesalesandspeedylogistics.
Amazonhasfacedscrutinyinrecentyearsoveritsworkplaceinjuryrecord, withfederalcitationsforsafetyviolationsand a year-longSenateprobethatfoundPrimeDaywas a majorcauseofworkerinjuries.
I thinkthatwewouldseethattheywouldrequire a lotmoretraining.
もっと多くのトレーニングが必要だと思う。
AndsogenerativeAIhasbeenreally a step-functionimprovementinbeingabletoinferfromourvastproductcatalogabouthowtohandle a givenproduct, evenif I haven't seenthatproductbefore.
Howcantheyimplementautomationtoimproveefficiencyandmanagelaborexpenses, buthowcantheydoitin a waythatcomplementstheiruseofhumansanddoesn't replacethem?
Sowe'reabletousegenerativeAItocreate a linkbetweenproductswehaveseenbefore, wherewedohave a saleshistory, and a newproductwehaven't seenbeforeyetandgetitintherightplacethefirsttime.
Sowe'reabletousecomputervisiontogetherwithgenerativeAIandthatvastamountofproductandpackagedatatothendetectdamagesandbeingabletosideline a packageifwethinkitmightbedamagedbeforeweshipittocustomers.
ButtrainingandrunninggenAIisitself a carbonintensiveprocess, a factthatcouldmakeithardforAmazontoachieveits 2019 climatepledgetoreachnetzerocarbonby 2040.
Everybody I knewwasbucklingtheirseatbeltbehindtheirbacksbecausethetimeittookjusttobuckleyourseatbelt, unbuckleyourseatbelteverytimewasenoughtimetogetyoubehindschedule.
NowwithgenerativeAI, thatdatacanbeusedtogeneratebetterhyper-personalizedproductrecommendationsto a shopper, thankstonewdevelopertoolslikeAmazonPersonalize.