In recent weeks, Amazon has found itself in an increasingly familiar position: playing catch-up in the AI space. The company’s late entry into generative AI technology, unveiled six weeks ago, has caused a stir among AWS (Amazon Web Services) cloud customers, many of whom are still waiting for an opportunity to test the product.
Amazon’s generative AI model mirrors OpenAI’s ChatGPT in its capacity to generate text or images by mining vast quantities of data. The announcement of Amazon’s venture into generative AI, however, has been met with skepticism, primarily because of the company’s uncharacteristically vague product launch.
AWS product launches usually garner glowing testimonials from multiple customers, while this launch included just one: Coda, a document-editing startup. Coda’s CEO, Shishir Mehrota, awarded Amazon’s technology an “incomplete” grade, highlighting that it was still “fairly early” in its development phase.
Amazon’s late entry into the generative AI arena has fueled speculation that the company launched the AI tools to quell perceptions of falling behind cloud competitors, namely Microsoft and Alphabet Inc.’s Google. Both these companies are already leveraging generative AI to enhance their products, revolutionizing the landscape of computing despite the technology’s current raw and error-prone nature.
However, Matt Wood, an AWS vice president of product, refuted the notion that Amazon’s AI software was incomplete or rushed. He stressed that the product was new and designed to allow select customers to test it and provide feedback.
Despite this, some industry insiders remain skeptical. Corey Quinn, the chief cloud economist at the Duckbill Group, described Amazon’s offering as seeming like “vaporware,” a term used to label products that are advertised before they are finished and may never see the light of day.
Despite these criticisms, it’s crucial not to underestimate Amazon’s potential in this space. The company boasts deep expertise in artificial intelligence, employing it across an array of tasks. AI is pivotal in managing stock levels, planning delivery routes, identifying faces in images or videos, and extracting text from medical records, among other applications.
However, some insiders argue that Amazon’s customer-focused approach often overlooks the kind of pure research carried out by competitors like OpenAI, Microsoft, and Google, which can lead to breakthroughs like ChatGPT.
In response to the ChatGPT hype, Amazon announced a new AWS service called Bedrock. This allows companies to access models built by Amazon partners and two of its own large language models, known as Titan. One model mirrors Amazon’s product search team’s approach, and the other resembles ChatGPT, summarizing content and composing emails or blog posts.
Despite this progress, most customers are still unable to test the technology, as they require Amazon’s approval—a departure from the company’s typical broad-access policy.
While some companies have secured deals to use the services, including Deloitte, 3M, and Royal Philips NV, it remains unclear when wider access to the technology will be granted.
In an interesting twist, Royal Philips NV has indicated it’s experimenting with ChatGPT, suggesting Amazon’s late entry into generative AI could potentially risk losing clients to more advanced offerings from other cloud providers.
Amazon’s journey into the generative AI market may be off to a rocky start, but with the company’s deep expertise and competitive spirit, it’s too soon to discount its potential impact in this space.