With the rapid growth in Artificial Intelligence (AI) applications, previous waves of technology are set to receive a boost, executives at Amazon Web Services (AWS), the dominant cloud and hosting provider, said. Yasser Alsaied, AWS’s Vice President of Internet of Things, pointed out that while IoT devices — small devices connected wirelessly as a part of large systems — would particularly benefit. Mr. Alsaied and other executives spoke to The Hindu from the sidelines of Amazon’s annual re:Invent conference in Las Vegas in November.
“Driverless cars have have a lot of sensors, over 160 of them,” Mr. Alsaied said. “We can ingest all this data and build simulations to make that product better for all markets, such as in India, Saudi Arabia and Canada.” Generative AI would, for instance, give car manufacturers more inputs and options from the data they already have. Mr. Alsaied also referred to increasing capabilities in ‘edge computing,’ which allow complex machine decisions to be made without having to send data to the cloud each time.
Manufacturing is a key area where the status quo can change drastically. “Nobody is ready to change manufacturing machines,” Mr. Alsaied said. But advances in edge computing allow factories to add small monitoring equipment to existing facilities, and they collect data that can then measure productivity and use AI to suggest process improvements. “In two to three years, a digital twin will become a must for factories,” Mr. Alsaied said, referring to simulated copies of places that can be used to test how an environment would respond to changes before they are done in the real world.
AI ‘models’ that are in use will also improve by leaps and bounds, Swaminathan Sivasubramanian, the company’s VP, Database, Analytics, and Machine Learning said. “In terms of their reasoning capabilities, you will continue to see more and more in a big way,” Mr. Sivasubramanian said. “But it’s not just about bigger models [that are based on more data]. Companies care about return on investment, and don’t want to spend a huge amount of money without justifying it with additional profits.”
“I still have the belief that these models will augment our intelligence instead of replacing it,” Mr. Sivasubramanian argued. “We as a society are going to figure out how to deploy these systems in a careful and responsible manner that is secure and privacy preserving.” He pointed to an invisible watermarking technology for AI-generated images announced by AWS, an alternative to visible watermarks “that can just be Photoshopped out,” as an example of taking thoughtful approaches to AI governance.
“In the future, we are going to see more and more work which can be done with machine learning and AI,” Anupam Mishra, who heads a team of ‘solution architects’ for the company in India, said. “We use a system called Code Whisperer to automatically suggest changes” when computer code is drafted, Mr. Mishra said, adding that in an A-B test, a team that used it was 57% more efficient than one that didn’t. “The more developer efficiency you have, the more of an exponential benefit there is for your company.”
“We see immense potential for generative AI across industries in India,” Mr. Mishra said. “We have so much data, so much potential available. We can transform industries.” Mr. Mishra said that AWS was keen to put its weight behind Indian startups that work in this area. “We recently had a program called ML Elevate, where 25 generative AI startups came together in a cohort, and spent a few months with AWS, with $200–$250,000 in credits to use on the platform, and provide mentoring. Many of them have built several applications out of India which are going to serve several enterprises.”
(Quotes have been edited for clarity)
(The writer was hosted by AWS at Re:Invent, a technology conference, at Las Vegas, in the United States.