Amazon Web Services (AWS) has announced Amazon Bedrock and numerous new services built on advanced artificial intelligence (AI) services. Amazon Bedrock is the company’s new application programming interface (API) for developers to utilize Amazon’s new generative AI tools.
“Think of [Bedrock] as a cloud-based and configurable alternative to OpenAI’s ChatGPT and DALL-E 2 aimed at businesses and developers,” Engadget writes in its coverage of Amazon’s new AI technologies.
“If it seems like artificial intelligence (AI) is everywhere lately, it is, but AI has been powering our everyday experiences for some time. When you ask Alexa to play a song, when you stride out — sandwich in hand — from an Amazon Just Walk Out-equipped store, or when you press play on a movie recommendation from Amazon Prime, you are tapping into AI. More specifically, you are interacting with machine learning (ML) models,” Amazon explains.
Generative AI models, like ChatGPT, DALL-E, Midjourney, and Adobe Firefly, to name just some, are a subset of machine learning. Generative AI models are already having a profound impact on daily life across many industries and will continually affect health care, science, entertainment, education, art, and more, in increasing ways.
“At Amazon, we believe AI and ML are among the most transformational technologies of our time, capable of tackling some of humanity’s most challenging problems. That is why, for the last 25 years, Amazon has invested heavily in the development of AI and ML, infusing these capabilities into every business unit,” Amazon says.
Amazon has announced four new innovations, including Bedrock, to support generative AI applications.
Describing Bedrock, Amazon says it’s a “new service for building and scaling generative AI applications, which are applications that can generate text, images, audio, and synthetic data in response to prompts.” Amazon Bedrock provides customers access to foundation models (FMs) from top AI startup model providers, including AI21, Anthropic, and Stability AI.
Amazon also provides Bedrock customers exclusive access to its Titan family of FMs that AWS has developed. FMs are the ultra-large ML models that generative AI relies upon.
CNBC reports that clients can customize Amazon Titan models using their data. However, to address security concerns, the data clients enter won’t be used to train Titan’s models.
Bedrock is currently in a limited preview stage. AWS customer Coda is using Bedrock to speed up its workflow and productivity.
“As a longtime happy AWS customer, we’re excited about how Amazon Bedrock can bring quality, scalability, and performance to Coda AI. Since all our data is already on AWS, we are able to quickly incorporate generative AI using Bedrock, with all the security and privacy we need to protect our data built-in. With over tens of thousands of teams running on Coda, including large teams like Uber, the New York Times, and Square, reliability and scalability are really important,” says Shishir Mehrotra, CEO of Coda.
Amazon has also announced the availability of its EC2 Inf2 instances powered by AWS Inferentia2 chips. Ultra-large ML models require extensive computational power. AWS Inferentia chips promise “the most energy efficiency and the lowest cost for running demanding generative AI inference workloads at scale on AWS.”
Generative AI models must also be extensively trained, which requires massive computing power. Amazon has announced new Trn1n instances on its servers powered by AWS Trainium chips.
“New Trn1n instances (the server resource where the compute happens) run on AWS’s custom Trainium chips, and offer massive networking capability, which is key for training these models quickly and in a cost-efficient manner,” Amazon explains.
For individual developers, Amazon is providing free access to its CodeWhisperer technology. CodeWhisperer uses generative AI to help software developers work faster. The AI technology provides real-time code suggestions based on user comments and prior code.
Amazon’s Swami Sivasubramian explains Amazon’s new generative AI technologies in extensive detail on AWS’ Machine Learning Blog.
Image credits: Amazon