Adam Selipsky Talks Amazons 25-Year Generative AI Strategy and AWS Full Stack Approach
- Adam Selipsky, CEO of AWS, explains Amazon’s generative AI strategy. Amazon has been doing machine learning for 25 years, beginning with personalization on the website
- AWS launched Sagemaker in 2017 and now has over 100,000 customers doing machine learning on it
- LMS (large language models) are used in elements of Amazon’s retail website search and Alexa’s voice responses
- Amazon has had its own LMS in production for years
- AWS is taking a full stack approach to generative AI by designing its own chips to train models and do inference, but also providing GPU capacity for customers who need it
- Nvidia GPUs are immensely popular for machine learning applications including generative AI.
Amazon Makes $100 Million Investment in Generative AI Technology
- Amazon is investing $100 million into a new generative AI innovation center
- They are providing their customers with the tools to build their own models for secure, private, and consistent access
- AWS has been working on artificial intelligence for many years and have long accumulated experience and experts in the field
- Amazon is offering a full stack of capabilities when it comes to generative AI, including designing their own machine learning AI specific chips.
Amazon Pushes Boundaries with Codewhisperer, AI Taking Center Stage in Cloud Era
- Amazon has a managed service for accessing models
- They have launched Codewhisperer which has reduced the time to complete tasks up to 57% for Amazon developers
- Amazon is focusing on customers and building what they need in the generative AI space
- They have an expression “pioneers need to be willing to be misunderstood” in order to focus on customers and build what they want, regardless of perception
- The cloud is becoming traditional with growth slowing, but AI is proving to be a convenient distraction and monetization can make up for slowdown in enterprise spend.
AWS Continues to Lead Cloud Space Despite Economic Conditions
- Amazon Web Services (AWS) is in a strong position despite softening economic conditions
- Customers are cost optimizing, but still entering into long-term commitments with AWS to move to the cloud
- AWS has significant advantages over competitors and is focusing on customer value rather than competing against other companies
- Competition in the cloud space is robust, and there will most likely be a few providers that stand out.
Amazon Seeks to Expand Chip Design Options with Third-Generation ARM-Powered Graviton Chips
- Amazon is developing their own chips with the help of ARM and are in their third generation
- AWS is an $85 Billion a year business, and they will continue to see significant CapEx investments for generative AI
- Amazon has a relationship with ARM that has been expanding for years, and they are looking to have a diverse set of options when it comes to chip design
- The latest chips (Graviton three) are 60% more energy efficient than x86 based chips.