Now is the time to experiment with generative AI says AWS

Last week Amazon Web Services (AWS) held its first local Summit since the pandemic, with Johannesburg playing host instead of the usual venue of Cape Town.

The hyperscaler used the event to reaffirm its commitment to cloud customers both in South Africa and across the African continent, with the likes of Capitec and BankServ highlighting how collaboration with AWS led to significant evolution at their respective companies.

AWS also used the event to highlight the newly announced cloud solutions primed to be made available to local customers, with generative AI being a major focus for the company moving forward, as it has been for the entire technology industry during 2023.

To that end, the recently unveiled Amazon Bedrock will be a key offering for the platform, and one that VP for Compute and Networking Services, David Brown, cites as being significant in future.

We had the opportunity to speak to Brown during a media-focused panel at the event, aiming to find out how AWS saw the use of generative AI platforms in South Africa panning out, especially as many organisations are still unsure of how best to implement said offerings given they are still very much in their infancy.

Brown acknowledged as much, but was equally keen to get generative AI solutions into the virtual hands of the hyperscaler’s customers as now is the perfect time to begin experimenting with the technology in his view.

“The reality is, it is early days for generative AI all over the world, so I don’t think South Africa is that far removed from what other regions are seeing,” the VP explained.

“One of the challenges we are seeing is that while there is enormous potential, we don’t really know how to deploy them (generative AI). In some cases we find a way to deploy generative AI, but maybe the cost is not something our business can support, and so these are things that we’re going to have to work through. I think companies are doing a lot of experimenting right now, so the thinking behind a service like Bedrock through AWS, is to start playing around,” he advised.

Here Brown emphasised the point that each business is different, with their data being the differentiator, a point he also mentioned during his keynote. As such, for true value to be wrought from generative AI, many companies are going to have to find out what does and what does not work, in the VP’s view.

“A few of my teams at Amazon went away and started playing with Bedrock, and brought back some amazing success stories, like allowing customers to use natural language to debug their application,” he shared.

“I think a lot of experimentation is going to be key, and then the other thing we’re focusing on is reducing cost associated with generative AI. Some of that will happen with model development, and some of it is going to have to happen in the hardware.”

While there is indeed some hesitancy in the industry, the Amazon VP is still of the opinion that curiosity around generative AI can unlock great value for local businesses.

“In spite of the uncertainty, the potential is just huge, which means we shouldn’t not lean into it. We should go and experiment, and if things work, we should double down on them,” he concluded.

With generative AI certainly here to stay, it looks like AWS is definitely targeting it as a way to give its customers a distinct advantage.


About Author


Related News