待翻譯：Wispr raises $280M to power up natural speech-to-text using AI

AI 服務暫時不可用，以下為來源正文，待恢復後補全翻譯。

Wispr AI Inc., the developer of a cross-platform, AI-powered dictation platform, today announced it raised $280 million in Series B funding at a $2 billion valuation. Menlo Ventures led the round alongside existing investors Notable Capital, NEA, Neo Ventures, 8VC and MVP Ventures. New investors Acrew, Forerunner, Goodwater, Peak XV, Together Fund, and PLUS Capital also joined the investment, bringing the company’s total raised to $361 million. Wispr’s primary product is Flow, a text-to-speech product that integrates into applications and lets people speak naturally into any text field to dictate. It takes spoken language, fixes grammar, misspoken language, filler words such as “ums” and “ahs,” and stumbles, and produces clean, readable prose for emails, memos and documents. This makes the tool extremely useful for mobile, desktop and other environments where someone might want to just sit down and talk to their computer instead of using a keyboard. Alongside the funding, the company is announcing a preview of its first proprietary AI model, Canto. Unlike most speech models, Canto was trained to detect speech across a wide variety of speech variances, in loud environments, with noise, interruptions, dogs barking, other voices, background noise and the sound of life happening in the background. The company said most voice models are trained on extremely clean voice assets to provide pure human vocal sounds. Canto was given examples that provide it with the ability to quickly isolate human voices from environmental noises – so that people can use their apps where they live, in their car, around traffic noise or in their office. When Canto is built into Flow, the company’s dictation tool, Wispr said the error rate in noisy environments falls from 30% to nearly 5% – 10%. In a quiet room, most modern AI transcription can be extremely smooth, with minor errors cropping up. In a loud environment, it can be catastrophic. Over the past 10 years, most text-to-speech apps have been used on the go for simple tasks such as checking the weather, asking for the time, or simple web searches, because that’s only a couple of words and very little can go wrong. With a system backed by Canto, it opens a whole new opportunity to sit in a car, even near a noisy venue, and talk about ideas and the model will get most of the words on digital paper. Something that, a year ago, a text-to-speech model would have struggled with. “I use Flow every day,” said Domantas Sabonis, three-time NBA All-Star. “I grew up rotating between English, Spanish and Lithuanian, and it keeps up with me no matter which one I’m speaking.” NBA players and business people aren’t the only users of transcription apps who will benefit as the technology advances. The deaf and hard-of-hearing community also uses AI transcription on mobile devices, where multi-participant conversations and noisy environments lead to interesting errors in the transcribed text; although many rely primarily on on-device transcription, the evolution of cloud models drives the industry forward. Wispr said people have written more than 60 billion words on Flow. Its product has also been used by almost all the Fortune 500 companies and over 10,000 enterprises, proving its use case across industries. Image: SiliconANGLE/Microsoft Designer A message from John Furrier, co-founder of SiliconANGLE: Support our mission to keep content open and free by engaging with theCUBE community. Join theCUBE’s Alumni Trust Network, where technology leaders connect, share intelligence and create opportunities. 15M+ viewers of theCUBE videos, powering conversations across AI, cloud, cybersecurity and more 11.4k+ theCUBE alumni — Connect with more than 11,400 tech and business leaders shaping the future through a unique trusted-based network Are you an AWS customer? Support SiliconANGLE financially by buying your AWS services from our Marketplace portal page and links: https://siliconangle.com/aws-marketplace/ About SiliconANGLE Media