Inflection AI is bringing conversational AI to the forefront — and into everyone's pockets — with its Pi chatbot. It sought to accelerate time to development and reduce downtime, better positioning it to be a leader in empathetic personal intelligence. Underpinned by reliable and stable Microsoft Azure AI technology, the Pi chatbot works with large language models to help people organize their days and offer them advice while matching their knowledge and interests on a growing range of topics.
Accessing powerful AI-optimized Azure virtual machines with InfiniBand networking through Azure AI infrastructure has helped Inflection AI gain superfast connections between GPUs and accelerate training of the largest language models available. Azure AI has been one of the first to get large-scale deployments of A100 GPUs in operation and to scale these GPUs by daisy-chaining them together in very large clusters, which is what Inflection needed to train the biggest models.
Personal assistants work to support people in various aspects of their lives, reducing the burden of simple and often monotonous tasks like scheduling appointments, sending reminders, and responding to emails. Pi is designed to engage in conversations, ask clarifying questions, and reflect back to ensure understanding. Pi takes an interest in the way users see the world and what they are interested in learning about. Pi is cautious by default, tries to be very factual, and if given feedback that it has got something wrong, will do its best to back down.
"It's been very valuable to us to work with Microsoft as we enhance and expand Pi because the reliability and scale of Azure AI infrastructure is among the best in the world," said Mustafa Suleyman, cofounder and CEO of Inflection AI. "Using turnkey Azure solutions to access GPU-accelerated virtual machines has sped up our time to development, but it's also put us in an amazing position with respect to our collaboration with Microsoft."
On the technology front, Inflection AI is confident that using Azure AI infrastructure has better positioned the company to lead in the years ahead, with dependable and stable infrastructure considerably speeding up development time and increasing the efficacy of its AI.
Details
- City
- Palo Alto
- Organization
- Inflection AI