$24M to Build the Future of Voice Intelligence
For the first time in human history, a rich understanding of language is available as a utility, delivered via API, useful enough to “be metered”. Intelligence has become a tool: something you can call, compose, and build on, divorced from any individual human mind.
This will change how humans interact with computers so profoundly that "interacting with computers" stops being the right description. Increasingly, we will be interacting with intelligence itself.
Practitioners of machine learning have begun to internalize this. So have the people putting these capabilities to work. But there's a second realization that fewer have reached: voice will be the primary way we interact with intelligence.
Today, on the back of rapid adoption across critical enterprise applications, we're announcing that Rime has raised a $24M Series A to do one thing: build the voice interaction models that will power the world's most important conversations.
The $24M Round
Our Series A was led by M13, with participation from Corazon Capital, Unusual Ventures, Cadenza Ventures, and Twilio Ventures. As part of this milestone, we're thrilled to welcome Morgan Blumberg, Partner at M13, to our Board of Directors.
"The design patterns that will define the speech-to-speech era of AI haven't been built yet. Rime is doing the foundational work by combining frontier AI with deep linguistic expertise to create the speech infrastructure the next generation of AI products will rely on. Lily, Ares and Brooke are building Rime with the rare combination of world-class AI research and deep linguistic expertise to deploy real time voice people can trust in complex, regulated environments." - Morgan Blumberg, Partner, M13
Why do we believe in continued investment in speech?
Speech is a native extension of our human capabilities for symbolic reasoning, and while it has become trite to say that we have been speaking and listening for hundreds of thousands of years, it still comes as almost shocking to suggest that typing will become relegated to history. But why?
Every era of computing has had an interface matched to where the work happened. When compute lived in a mainframe, we fed it punch cards. When it moved onto our desks and laps, the graphical interface changed the design pattern entirely. But the GUI's premise, the premise of the last forty years, was that you did the work and the machine waited for your next click.
Machine intelligence inverts this premise. As models take on real work, most compute will happen in the background, without our involvement. The design question is no longer "how do I operate this machine”, in the same way that we, very soon, will not be asking ourselves “how do I drive this car?”
And yet: the modeling primitives built for voice interaction haven’t been built.
We are also proud to announce that Rafael Valle has joined Rime as our Chief Scientist. Rafael has spent his career pursuing audio superintelligence. Prior to joining Rime, he led audio understanding in Meta Superintelligence Labs and worked at NVIDIA’s ADLR's (Applied Deep Learning Research) audio team.
Building Intelligence for Conversation
The primitives that will power voice interaction with intelligence are being built by Rime. While we are proud that our models power the voice applications for many of the Fortune 500, there is still so much left to build.
While LLMs have changed how voice applications are built, they haven’t changed how they feel to talk to. Fulfilling this promise requires solving incredibly complex engineering problems. It will also require using our linguistic expertise, our taste, and our craft as research instruments.
