Smallest.ai Raises $13 Million to Improve AI Voice Agents
In brief
- Smallest.ai has raised $13 million to develop a small voice model that can mimic human conversation.
- The company wants to make speaking to an AI agent indistinguishable from talking to a human.
- The startup's model is designed to listen, think, and speak simultaneously, reducing response lag in voice conversations.
- This is important because even short pauses in voice conversations can feel unnatural.
- Smallest.ai's model can handle voice-specific nuances like diverse accents and noisy environments.
- The company has already raised over $21 million in total funding and has existing customers like RingCentral and Truecaller.
- Smallest.ai will continue to develop its voice model to improve AI customer support interactions.
Read full story at TechCrunch →
More briefs
AI Company Seeks Large Book Order
A Dutch antiquarian bookseller received an email from a company called 2077AI looking to buy over 3,000 books. The bookseller dismissed the email as spam, but it is part of a larger trend of AI companies looking for high-quality written material to train language models. This trend has raised concerns about the use of books for AI training, with one company having purchased millions of physical books, scanned them, and discarded the originals. The demand for books by AI companies could impact booksellers who specialize in selling secondhand books at scale. The future of book sourcing for AI training is likely to involve more large orders like this one.
Microsoft Chooses Affordable, Specialized AI Models Over Costly General-Purpose Ones
Microsoft's AI division is focusing on smaller, specialized models that are more affordable compared to expensive general-purpose ones. According to Mustafa Suleyman, the CEO of Microsoft AI, their new model, MAI-Cyber-1-Flash, outperforms Anthropic's Mythos in certain benchmarks and costs half as much. However, it still relies on OpenAI for handling complex tasks. The shift in competition is moving from individual models to orchestration software that manages and routes these models efficiently. This approach could make AI more accessible and cost-effective for developers and businesses, especially those with limited budgets. Looking ahead, expect more companies to adopt similar strategies, focusing on optimizing model performance within specific contexts rather than striving for general-purpose solutions.
Israel Pays Millions to Influence AI Chatbots
Israel is paying 46.5 million dollars to train AI chatbots. The goal is to influence what these chatbots say about Gaza. This matters because tens of millions of Americans use chatbots. If the plan works, these Americans will get answers that favor Israel. The plan is already showing results, with some chatbots giving answers that come from websites created by Israel's contractors. The impact of this operation will be seen in the future.
Fish Audio Raises $52 Million
Fish Audio raised $52 million in a seed round. The company offers a library of natural language controls for AI voice models. The company has over 8 million users and generates $21 million in annual revenue. Its models are used by indie developers and creators. The funding will help the company build on its traction. The company will use the funds to continue developing its voice models. Fish Audio will keep working on its technology to make AI voices more expressive and natural. The company will continue to grow with its new funding.
Amazon Overhauls AI Strategy
Amazon is winding down many in-house AI models and reorganizing its teams. The company is focusing on its highest priorities and concentrating engineering talent on a new strategy. Amazon has begun deprecating most of its in-house flagship Nova models, including the Premier and Omni models. This change affects thousands of customers who depend on these models. Amazon will debut a new flagship foundation model at its annual conference, which typically happens in the fall, and the company remains committed to investing in frontier models. Amazon will continue to develop new AI models.