- Work With AI
- Posts
- ElevenLabs lets you create AI-generated voices from text prompts
ElevenLabs lets you create AI-generated voices from text prompts
Plus new tools from Meta, Midjourney, and Coinbase
Today’s Highlights:
📰 News: ElevenLabs lets you create AI-generated voices from natural text prompts + new tools from Meta, Midjourney, and Coinbase
💰 Funding: Waymo raises over $5Bn+, its largest round ever
⚡️ Top News Stories:
1. ElevenLabs has introduced Voice Design, a tool that lets users create custom AI voices from natural text prompts, specifying characteristics like age, accent, tone, and character type, offering both realistic (e.g., conversational, professional) and creative voices (e.g., ogre, pirate) for versatile projects across media and ads.
2. Researchers have found that OpenAI’s Whisper transcription model frequently hallucinates or invents text not present in the original audio, with concerning implications for high-stakes applications.
Studies indicated hallucinations in up to 80% of transcriptions examined
The tool frequently invents text not present in recordings, including fabricated racial commentary, violent language, and even non-existent medical treatments, as reported by researchers and developers.
Whisper is widely used in healthcare, with over 30,000 clinicians employing Whisper-based tools, despite OpenAI’s caution against deploying it in high-stakes settings where hallucinations could have severe consequences.
Experts are urging OpenAI and regulators to address AI transcription issues, highlighting that even minor hallucinations can erode trust in AI and potentially endanger vulnerable groups, such as the Deaf community relying on accurate captions.
3. EngineAI Robotics unveiled its first full-sized humanoid robot, SE01, which features a groundbreaking end-to-end neural network allowing natural human-like walking motion—an achievement previously unachieved in humanoid robotics.
The SE01 robot integrates NVIDIA and Intel dual processors, along with stereo cameras, giving it exceptional visual processing and adaptability, suitable for both research and industrial applications.
EngineAI Robotics plans to release additional products in 2024 and targets a 1,000-unit annual production by 2025, aiming to expand humanoid robotics into daily life and various industrial applications.
4. Meta has secured a multi-year deal with Reuters, allowing Meta AI to provide real-time news answers to U.S. users on Facebook, Instagram, WhatsApp, and Messenger by summarizing and linking to Reuters content. While Reuters will be compensated for its journalism, it's unclear if Meta will use this content to train its Llama language model.
5. Meta has released NotebookLlama, a free, open-source AI tool that transforms documents and URLs into conversational audio summaries similar to Google’s NotebookLM, offering a customizable “recipe” format and using Meta’s Llama language models for flexible, podcast-style outputs.
6. Midjourney has introduced a new "Edit" tool that allows users to upload images and make AI-driven adjustments to style, texture, and composition while preserving the original core elements.
7. Coinbase has launched "Based Agent," a tool enabling users to create AI-powered crypto agents in under 3min to handle tasks like trading, staking, and swaps. Built with Coinbase’s SDK, OpenAI, and Replit, Based Agent requires API keys from Coinbase and OpenAI to interface directly with smart contracts for crypto transactions.
8. Apple is inviting researchers to test its Private Cloud Compute (PCC) system for vulnerabilities, offering bounties up to $1M for significant security findings; although many Apple Intelligence AI features run on-device, intensive tasks are processed on PCC servers built with Apple Silicon, maintaining strong privacy standards.
9. Google's AI-generated search summaries, called AI Overviews, are now available in over 100 additional countries, including Canada, Australia, South Africa, and the Philippines, supporting multiple languages like English, Hindi, Japanese, and Spanish to provide summaries matching the user’s search language.
10. Google is developing a new AI agent, codenamed "Project Jarvis," designed to automate web browser tasks like shopping, research, and travel booking by capturing and interpreting screenshots, with a potential launch as early as December.
11. Disney is launching a large-scale AI initiative involving hundreds of employees to streamline visual effects, post-production, and backend functions across its films, shows, and parks, though it won’t impact customer-facing elements directly.
12. Meta is reportedly developing its own search engine for integration with its Meta AI chatbot to reduce dependence on Google and Microsoft, having indexed the web for over eight months, partly motivated by a push for self-reliance after Apple’s 2021 App Tracking Transparency feature cost Meta over $10Bn in ad revenue.
13. NVIDIA announced that xAI’s Colossus supercomputer, featuring 100K NVIDIA Hopper GPUs, is the world’s largest AI supercomputer, and is expanding to 200K GPUs in Memphis, Tennessee.
14. OpenAI has dissolved its “AGI Readiness” team, which assessed the company’s preparedness for managing advanced AI, coinciding with the resignation of lead advisor Miles Brundage.
💰 Top Funding News:
1. Waymo, the autonomous robot taxi company, raised $5.6Bn (its largest funding round to date), led by Alphabet, w/ Andreessen Horowitz, Fidelity, Perry Creek, Silver Lake, Tiger Global, and T. Rowe Price.
2. Sierra, which specializes in AI-powered customer service agents that go beyond traditional chatbots, using a ‘constellation’ of generative AI models from OpenAI, Anthropic, and Meta to allow brands to customize the AI’s personality to align with their corporate identity, raises $175M at a $4.5Bn valuation, led by Greenoaks Capital, w/ ICONIQ and Thrive Capital
3. Path Robotics, which uses AI-enhanced autonomous robotic systems to handle labor-intensive welding tasks, particularly in small to medium manufacturing operations, raises a $100M Series D, led by Matter Venture Partners and Drive Capital.
4. Read AI, which provides an AI bot that automates meeting notes, summarizes conversations, and generates insights across a variety of platforms, raised a $50M Series B, led by Smash Capital, w/ Madrona and Goodwater Capital.
5. Nooks, which uses AI to streamline and automate routine tasks in sales, such as finding contacts, voicemail logging, and drafting emails, raised a $43M Series B, led by Kleiner Perkins.
6. Reflexivity, which develops AI solutions tailored to the investment industry, focusing on advanced financial analysis and decision-making tools, raised a $30M Series B, led by Greycroft and Interactive Brokers.
That's all for today's email! If you want more please follow us at the social channels linked below, or check out our website!
How'd you like today's email? |
Share our newsletter: If you like our work please share/forward this email with your friends, colleagues, and family. It's the best way to support us!
If this email was forwarded to you please sign up here to continue receiving them.
Want your content, product, jobs, or event featured in our newsletter? Reply to this email with the details, and our team will reach out to you.
Do you use AI for work? Tell us how, and you could be featured in our newsletter!
Check out our website for more resources, including a list of AI investors, products, events, and twitter follows.
For an archive of all our posts, click here.
We'd love to hear from you! You can always leave us comments or feedback by replying to this email!