Amazon is working on an improved large language model to improve Alexa
Amazon’s CEO has claimed that with an improved LLM, they can achieve their goal of building “the world’s best personal assistant” faster.
Amazon CEO Andy Jassy announced during the company’s Q1 2023 financial session that they are building a “more powerful and general” Large Language Model (LLM) to power Alexa’s smart assistant capabilities.
According to a TechCrunch report, as Jassi explains, Amazon already uses an LLM for Alexa, but now plans to work on a more capable version of the technology. Amazon’s CEO has indicated that with an improved LLM, they are able to achieve their goal of building “the world’s best personal assistant” faster. However, he points out that this will be difficult to do in many areas.
Amazon’s CEO Says About Alexa-Enhanced LLM Development
Jassi said during this meeting:
“I think when people often ask us about Alexa, what we often share with them is that if we just made a smart speaker, it would be a much smaller investment. But we believe in a vision where we want to build the best personal assistant in the world. This is difficult to do; Because it is located in many domains and its surface is very wide. “However, large language models and generative AI make the underlying models much more effective, which I think really accelerates the possibility of building the world’s best personal assistant.”
The CEO of Amazon continued his explanation and pointed out that he believes that Amazon had a good starting point with Alexa; Because it has “hundreds of millions of endpoints” used in the entertainment, shopping, and smart home sectors. He also noted that there are many contributions from third-party partners.
During the meeting, Jassi also emphasized that Amazon has invested in AI and LLM for years and has the ability to invest heavily in LLMs, which is why it unveiled Bedrock’s cloud-based AI service last month.