In this interview, Adam Selipsky, the CEO of Amazon’s cloud computing unit (AWS), talks about how companies are cutting back on cloud computing services, Amazon’s investment in AI company Anthropic, and the future of generative AI. The conversation has been edited for brevity and clarity.
Reducing Cloud Computing Spending
Selipsky discussed how companies have been reducing spending on cloud computing services in recent months. He noted that many customers have improved their costs by optimizing their use of cloud computing, and that there are other customers still in the middle of this process. He emphasized that the companies that will succeed are those that invest in this economically uncertain time, while others hesitate in their broader investments. He also pointed out the significant interest in Amazon’s generative AI offerings.
Amazon’s Vision for Generative AI
Selipsky talked about three different layers in generative AI technology. At the bottom layer of technology, the infrastructure necessary to deploy generative AI is provided. He noted that Amazon has a substantial business built on Nvidia GPUs and the design and provision of its custom chips like Trainium and Inferentia.
In the middle layer of technology, multiple foundational models are made available to customers through a service called Amazon Bedrock, including models from Anthropic, Meta, and Amazon itself. He affirmed that the idea that one company would provide all the models in the world is unrealistic, and that customers need to experiment with different models, which is what Amazon offers.
At the top layer of technology, applications built on generative AI are utilized. For this purpose, Amazon provides a companion for developers to ease the programming process.
Amazon’s Investment in Anthropic
Selipsky spoke about Amazon’s investment in the AI company Anthropic. He confirmed that there is a close and very beneficial relationship between the two companies, as Anthropic has chosen Amazon as its primary cloud computing provider for critical workloads. He noted that most of Anthropic’s workloads will operate on the AWS platform. He stressed that these relationships are not exclusive, as Anthropic requires significant computing power and has utilized other cloud service providers in the past. He highlighted the importance of having Anthropic within Amazon Bedrock, a reliable service for accessing foundational models.
He pointed out that the collaboration between Amazon and Anthropic in training and running its models on our chips will help us improve that technology. He confirmed that this relationship is indeed beneficial for both companies and their shared customers for a long time to come.
Fortifications in Amazon’s Technology
Selipsky discussed the importance of responsible AI, which Amazon takes very seriously. He noted that they have a number of principles for responsible AI and have created cards for their services that discuss model usage and the intended use of the model as well as how it was trained. He affirmed that they are striving to provide more transparency in how these services are built and utilized.
He mentioned that solutions related to responsible AI will require multi-stakeholder solutions. They need collaboration between cloud computing industry leaders like AWS and those who produce models like Anthropic, as well as governments, academia, and others. For this reason, they have actively participated in responsible AI forums at the White House and in the UK.
The Future of AI Competition
Selipsky shared his expectations for the coming year in the AI field. He noted that there will be rapid development and change, which partially reflects the fact that we are still in a very early stage of generative AI evolution. He confirmed that adaptability and flexibility are important traits for customers to achieve their business goals and delight their customers. They need to be flexible, agile, and able to adapt in how they utilize generative AI.
This
The interview highlights Amazon’s role in the development of technology and generative artificial intelligence, showcasing their vision and investments in this growing field. It underscores the importance of transparency and responsibility in the development of this technology and the collaboration between companies, governments, and academia to achieve success and mutual benefit.
Leave a Reply