News
AWS Takes Minority Stake in 'Claude' Chatbot Owner with $4B Infusion
Anthropic, in turn, has committed to expanding its support of the Amazon Bedrock foundational model platform.
Self-described "AI safety and research company" Anthropic now has a formidable backer: cloud giant Amazon Web Services.
AWS this week pledged to invest $4 billion in Anthropic, giving a significant boost to the creator of the "Claude" AI chatbot. The investment would make AWS a minority owner in Anthropic. In turn, Anthropic will tap AWS' extensive compute power -- particularly the Trainium and Inferentia machine learning-optimized chips -- to run the majority of its workloads.
Now in version 2, Claude differentiates itself from other chatbots in the market (foremost among them ChatGPT) by being "safer." Specifically, Anthropic uses what it calls "constitutional training" to make its chatbot "helpful, honest, and harmless."
Also, while ChatGPT is widely considered to have the advantage over Claude in terms of its integrations and the sophistication and accuracy of its output, Claude 2 can support much longer queries at a time -- as many as 100,000 tokens (roughly 75,000 words), compared to ChatGPT 4's 8,192 tokens. Additionally, Claude 2's training data extends to early 2023, while ChatGPT 4 has been trained to September 2021.
(ChatGPT 4, of course, is sterwarded by OpenAI, in which Microsoft -- a chief AWS rival -- has a 49% stake.)
"Training state-of-the-art models requires extensive resources including compute power and research programs," Anthropic noted in a press release Monday. "Amazon's investment and supply of AWS Trainium and Inferentia technology will ensure we're equipped to continue advancing the frontier of AI safety and research. We look forward to working closely with Amazon to responsibly scale adoption of Claude and deliver safe AI cloud technologies to organizations around the world."
As part of its deal with AWS, Anthropic is also committing to expand its support for Amazon Bedrock, a relatively new AWS service that gives developers access to foundation models from AWS partners like Anthropic to build generative AI-powered apps. Anthropic promises to provide AWS customers "access to future generations of its foundation models via Amazon Bedrock," according to this AWS press release, as well as "early access to unique features for model customization and fine-tuning capabilities."
"Amazon developers and engineers will be able to build with Anthropic models via Amazon Bedrock so they can incorporate generative AI capabilities into their work, enhance existing applications, and create net-new customer experiences across Amazon’s businesses," according to AWS.