Andy Jassy Calls Microsoft and Google A.I. 'Hype Cycle' and Amazon Starts a 'Substance Cycle'
– Amazon CEO Andy Jassy differentiates between the A.I. "hype cycle" and the "substance cycle."
– Jassy considers generative A.I. as a significant technical transformation and labels many current A.I. chatbots and tools as part of the "hype cycle."
– Amazon aims to become a major player across the entire A.I. supply chain, similar to its success in the cloud computing industry.
– Every business unit within Amazon is intensely focused on generative A.I.
– Amazon's A.I. strategy encompasses three macro layers: computing capabilities, underlying models, and the application layer (e.g., ChatGPT or Bard).
– The development of new chips capable of generating vast computing power is crucial for future A.I. technology.
– Amazon's AWS has developed Trainium and Inferentia chips with better price-performance ratios than competitors.
– Foundational models, powered by these chips, are the basis for generative A.I. applications. Only a few models are expected to dominate the field, but they are currently expensive and time-consuming to develop.
– To address this issue, Amazon created Bedrock, a service that sells large machine learning models to customers who prefer not to develop their own.
– Bedrock aims to become a go-to service for companies aspiring to develop their own generative A.I. applications.
– Amazon is building a generative A.I. tool to assist developers and improve customer experiences, but expects most applications to be built by other companies using AWS's A.I.-specific tools.
– AWS CEO Adam Selipsky highlights that the current A.I. race is comparable to being three steps into a 10K race.