HITS

Amazon Web Services (AWS) executives used their keynotes at the 12th annual re:Invent conference, in Las Vegas, Nov. 28-30, to highlight the company’s latest artificial intelligence (AI) and other innovations.

They also predicted that advances in generative AI (GenAI) will continue, while growth in GenAI and AI overall will continue to grow as more companies opt to support it.

During his Nov. 29 keynote, Swami Sivasubramanian, VP of data and AI at AWS, explored how you can use your company’s data to build differentiated generative AI applications and accelerate productivity for employees across the organization, among other topics.

“This year, I’m especially excited to stand on this stage and share our vision with all of you. That’s because this year we are standing at the edge of another technological era – an era in which a powerful relationship between humans and technology is unfolding right before us,” he told attendees and those viewing remotely.

“Generative AI is augmenting human productivity in many unexpected ways, while also fueling human intelligence and creativity,” he said. “This relationship, in which both humans and AI form new innovations, is rife with so many possibilities. This partnership is similar to many symbiotic relationships we observe in nature, where different species not only co-exist together but also benefit from one another…. Today, the world is captivated by generative AI’s potential to reinvent the way we work and form new innovations.”

Although AI and machine learning (ML) have “spurred rapid innovation over the last couple of decades, I believe this new era of human-led creativity with GenAI will shift our relationship with technology towards one of intelligence augmentation,” he said. “I find this co-evolution of humans and technology incredibly exciting because  there is also another symbiotic relationship happening under the hood that is key to driving our progress forward here: the relationship between data and Gen AI.  The massive explosion of data has enabled these foundational models to exist in the first place. The large language models that power generative AI [are] accelerating new innovations everywhere. With these new advancements, our customers can harness that data to build and customize these apps with models faster than ever before by using GenAI to make it easier to work with and organize that data.”

His take on what we’re seeing is that “data, generative AI and humans are all unleashing their creativity together” to develop GenAI apps, for one thing, he said.

What is needed for all this is a “powerful infrastructure to run these data intensive and ML intensive applications,” he said, noting AWS “has a long history of providing our customers with a comprehensive set of AI, ML data and compute stack just for this. And now, according to HG Insights, more machine learning workloads run on AWS than any other cloud provider.”

AWS thinks of its AI and ML offerings as a “three-layer stack,” he said, explaining: “At the lowest layer is the infrastructure tier. You will need to cost effectively train these foundational models and deploy them at scale, including our hardware chips and GPUs. This layer also includes Amazon SageMaker, our service which enables ML practitioners to easily build, train and deploy the ML models…. At the middle layer we have Amazon Bedrock,  which provides access to the leading” large language models (LLMs) … to build and scale generative AI applications. And, at the top layer, are the applications that help you take advantage of chained AI without the need for any specialized knowledge or having to write any code.”

Included among those apps is Amazon Q, the company’s new GenAI-powered assistant that he said is “tailored to your business.”

He explained: “Each of these layers build on the other and you may need some or even all of those capabilities at different points in your GenAI journey…. Amazon Bedrock has a broad set of capabilities for building and scaling GenAI applications at the highest levels of privacy and security. One of the main reasons customers gravitate towards Bedrock is the ability to select from a wide range of leading foundational models that support their unique needs.”

That customer choice is “paramount because we believe no one model will rule the world,” he said, predicting: “We are still in early days with GenAI, and these models will continue to evolve at unprecedented speeds. That’s why customers need the flexibility to use different models at different points for different use cases.”

AWS has also stepped up its efforts to cut down on false information and deep fake images, he said, noting it’s offering “built-in mitigations for toxicity and bias.”  Meanwhile, “to build on our commitments we made at the White House earlier this year to promote the responsible development of AI technology, all Titan- generated images come with an invisible watermark designed to help reduce the spread of misinformation,” he said. “By providing a discrete mechanism to identify AI-generated images, AWS is among the first model providers to widely release built-in invisible watermarks that are integrated into image outputs and are designed to be time progressive.”