Amazon joins Google and Microsoft in the AI Race

It has been announced by Amazon.com Inc. that is joining Microsoft Corp. and Google in the race for generative artificial intelligence, announcing both technologies aimed at Amazon's cloud customers as well as a marketplace where other companies can sell AI tools. 

‍

Amazon Web Services, the e-commerce giant's unit for offering large-language services, plans to announce on Thursday a pair of new large-language models. One of them will be designed to generate text, while the other will help power web search personalization, among other things. There are no plans for Amazon to release a chatbot like the ones that were introduced to mixed reviews by Microsoft and Google in the past few years. 

‍

Titan is Amazon's large-language model, which was trained on a large amount of text in order to be able to summarize content, write a draft of a blog post, or engage in open-ended question-and-answer sessions. Developers will be able to access them on an AWS marketplace called Bedrock in which they will be able to tap into models developed by other companies that are also working on generative AI, such as AI21 Labs, Anthropic, and Stability AI.

‍

Artificial intelligence, a form of computer software that can be used to generate text, images, or videos based on prompts received from a user, has captured the imagination of Silicon Valley, setting off fierce competition among companies to take advantage of the technology. The advocates of chatbots like ChatGPT and image generation tools like Dall-E believe that generative AI will revolutionize the kinds of tasks that software will be able to complete in the future. 

‍

Through a partnership with ChatGPT maker OpenAI, Microsoft has integrated generative AI technology into its Bing internet search service and plans to deploy those tools across the company's product portfolio in the near future. Google, the search giant owned by Alphabet Inc., is in a race to make similar moves. Meta Platforms Inc. has released its own large-language model and stated that similar work is being carried out throughout the rest of the organization. 

‍

AWS is one of the most prominent companies in the field of artificial intelligence, offering services such as on-demand computing power, software tools, and machine-learning applications. AWS has previously partnered with companies such as Hugging Face Inc. and Stability AI, which creates the image generator Stable Diffusion. However, the company hasn't previously announced plans to release a homegrown large-language version of the software.  

‍

As Swami Sivasubramanian, AWS's vice president of databases, machine learning, and analytics, explained, Amazon has been working on large-language models for a long time. They are already being used by Amazon to help shoppers find products on the retailer's website, as well as to power part of the Alexa voice assistant, among other things. 

‍

“In an interview with Sivasubramanian, he said that Amazon has taken an interest in this area for quite some time. 

‍

Customers of AWS will have the opportunity to apply to use the models during a preview period beginning on Thursday. The company has not yet decided on the price to access the tools, but Sivasubramanian did say that homegrown chips built by Amazon Web Services, such as Inferentia2 and Trainium, may be able to help customers keep costs down as they do their own machine learning. 

‍

As part of the announcement on Thursday, the Seattle-based company said that CodeWhisperer, which uses predictive tooling to proactively suggest code as developers type it, would be available free of charge to individual developers. 

‍

“As far as I am concerned, there is not going to be one model that will dominate the world,” Sivasubramanian explains.