Note from Matt: Below is an AI-generated list of AI news and tools. Our bot scours the internet daily to find AI news. Because it’s AI, it often finds older news or non-AI news and adds it to the mix. It also seems to break links often. We’re constantly working on improving the automation but this should, for the most part, ensure no AI news ever gets missed

The Latest AI News:

• xAI raises $6 billion to build out AI infrastructure and accelerate research – Elon Musk’s AI startup, xAI, has raised $6 billion in a Series B funding round to develop advanced AI and bring its first products to market.
• OpenAI forms a new safety committee to oversee its models – Following the departure of its key safety leaders, OpenAI has formed a new Safety and Security Committee led by board members to oversee the safety of its AI models.
• Scarlett Johansson says OpenAI’s ‘Sky’ voice is ‘eerily similar’ to hers after she declined to voice it – Actress Scarlett Johansson stated that OpenAI recreated her voice for its “Sky” chatbot without her permission, after she had previously declined an offer to voice it.
• Perplexity launches Pages to turn research into shareable articles – Perplexity AI has launched Pages, a new feature that allows users to generate and publish comprehensive, well-structured articles on any topic from a single prompt.
• Google’s AI Overviews are producing bizarre and dangerously incorrect answers – Google’s new AI Overviews in search have been criticized for providing strange and incorrect answers, such as advising users to add glue to pizza or eat rocks.
• Microsoft unveils Phi-3-vision, a small but powerful multimodal model – Microsoft has released Phi-3-vision, a compact and efficient multimodal AI model capable of analyzing images and text that can run locally on mobile devices.
• AI Pin maker Humane is reportedly looking for a buyer – AI hardware startup Humane is reportedly exploring a sale of the company for between $750 million and $1 billion following the poor reception of its Ai Pin device.
• Apple nears deal with OpenAI to use ChatGPT on iPhone – Apple is reportedly finalizing an agreement with OpenAI to integrate ChatGPT features into the upcoming iOS 18 for the iPhone.
• Adept AI releases Fuyu-Heavy, its new flagship multimodal model for enterprise actions – AI startup Adept AI, which is building a model to automate software tasks, released Fuyu-Heavy, its new flagship multimodal model designed for complex digital agent tasks.
• ElevenLabs launches Text to Sound Effects creator – AI voice company ElevenLabs has launched a new tool that can generate sound effects, short musical tracks, and soundscapes from text prompts.
• Sakana AI introduces Evolutionary Model Merging to create new open source models – Japanese AI lab Sakana AI introduced an automated, evolution-based method for merging different open source models to create new AI models with superior, combined capabilities.
• A new, potentially powerful Mistral model was leaked online – An unreleased 12B parameter model from Mistral AI appeared on Hugging Face, with early tests suggesting it may be a powerful new prototype rivaling models like Llama 3 70B.
• Microsoft introduces Copilot+ PCs for a new era of AI – Microsoft announced a new category of “Copilot+ PCs” with powerful neural processing units designed to run advanced AI features, like the new “Recall” function, locally on the device.
• EU countries give final nod to landmark AI Act – The European Union has given final approval to its landmark AI Act, establishing the world’s first comprehensive regulations for artificial intelligence.
• Legal AI startup Harvey raises $80M at a $2B valuation – Harvey, a startup that builds AI tools for law firms, has raised $80 million in a new funding round led by Kleiner Perkins and Elad Gil.
• Researchers bypass chatbot safety with ASCII art jailbreak – Researchers discovered a new “many-shot jailbreaking” technique that uses long, complex prompts filled with fake dialogues, including ASCII art, to bypass the safety features of large language models.
• Anthropic maps the mind of a large language model to improve safety – Anthropic researchers are studying smaller AI models as “model organisms” to understand how they represent concepts internally, aiming to make larger models safer and more interpretable.
• TikTok introduces new AI tools for advertisers, including virtual avatars – TikTok launched a suite of AI-powered advertising tools, including AI-generated scripts and virtual avatars, to help brands and creators produce ad content more efficiently.
• A Stanford study shows Generative AI’s mixed impact on freelancers – A Stanford study on the platform Upwork found that while generative AI helps lower-skilled freelance workers improve, it also tends to decrease job opportunities and earnings for all workers.
• Pollen Robotics unveils Reachy 2, an open-source humanoid robot – Pollen Robotics has launched Reachy 2, a new open-source humanoid robot with advanced mobility and expressive features, powered by a vision-language model for interaction.
• LangChain v0.2 focuses on production-readiness – The popular AI development framework LangChain has been updated to version 0.2, focusing on production-readiness by creating a clear separation from its experimental beta package.
• Why fine-tuning small models may be a waste of time – A blog post argues that fine-tuning smaller open-source models is often inefficient and costly compared to using prompt engineering with more powerful, general-purpose large models like GPT-4.
• How to use AI to automate your job search – A review of AI-powered job search tools like Kickresume, Teal, and Careerflow highlights how they can help optimize resumes, write cover letters, and track applications.
• A free course on fine-tuning LLMs with your own data – A new course from DeepLearning.AI and Lamini teaches how to fine-tune open source LLMs using a new technique that is faster and more efficient than traditional methods.
• Tutorial for building a multi-agent travel planner with GPT-4o – A tutorial demonstrates how to build a multi-agent travel planner application using GPT-4o, Tavily for web searches, and the LangGraph framework for coordinating AI agents.

Trending AI Tools:

• VASA-1 – Microsoft’s new AI model that can create lifelike talking faces from a single portrait image and an audio clip.
• Llama 3 – Meta’s latest generation of open-source large language models, designed to be integrated across its products.
• Claude 3 – A family of state-of-the-art large language models from Anthropic, often used as a benchmark for performance.
• Gemini – Google’s family of multimodal AI models that can understand and operate across text, code, images, and video.
• ChatGPT / GPT-4 – OpenAI’s flagship conversational AI, capable of a wide range of text-based tasks from writing to coding.
• Meta AI – An AI assistant powered by Llama 3, integrated into Meta’s family of apps like WhatsApp, Instagram, and Facebook.
• Grok – xAI’s conversational AI model designed to answer questions with a bit of wit and a rebellious streak.
• Devin – An autonomous AI agent from Cognition Labs designed to act as a fully-functional AI software engineer.
• Reka Core – A powerful multimodal model capable of understanding text, images, video, and audio, competing with top-tier models.
• Stable Audio 2.0 – A new model that can generate full songs with coherent musical structures up to three minutes long.
• Udio – An AI-powered application for generating high-quality music across various genres.
• Suno – An AI music creation tool that generates songs with vocals from a simple text prompt.
• FigJam AI – An AI assistant within Figma’s digital whiteboard to help generate, sort, and summarize ideas.
• HeyGen – An AI video platform that allows you to create studio-quality videos with AI-generated avatars and voices.
• Midjourney – A popular AI image generator known for producing highly artistic and detailed visuals from text prompts.
• Perplexity – A conversational search engine that provides direct, sourced answers to your questions.
• Pika – An AI video platform that can generate and edit videos from text prompts and images.
• Runway – A comprehensive AI video editing and generation suite with a wide range of magic tools.
• Opus Clip – An AI tool that repurposes long videos into viral short clips for social media with a single click.
• Descript – An all-in-one audio and video editor that lets you edit media as easily as editing a text document.
• Character AI – A platform that lets you create and have conversations with your own unique AI characters.
• Humane AI Pin – A new screenless, wearable AI device designed to be an ambient personal assistant.
• MemeCam – A mobile app that lets you create memes just by talking to your phone.
• DALL-E 3 – OpenAI’s advanced image generation model that creates detailed images from text descriptions.
• Gamma – An AI tool that helps create beautiful and engaging presentations, webpages, and documents.
• Notion – A popular all-in-one workspace that integrates AI to help with writing, summarizing, and organizing.
• Poe by Quora – An application that provides access to a variety of different AI chatbots and models in one place.
• Recast – An AI-powered tool that turns any online article into a short, conversational audio summary.
• Eightify – A tool that provides instant, AI-powered summaries of YouTube videos.
• Synthesia – A leading AI video generation platform focused on creating professional videos with AI avatars for corporate use.
• Wondercraft AI – A platform for creating captivating audio content, like podcasts and ads, using AI voices.
• AI Voice Generator by PlayHT – A tool for generating realistic text-to-speech voice overs online with AI.
• Grammarly – An AI-powered writing assistant that checks for spelling, grammar, tone, and clarity.
• Kickresume – An AI-powered builder that helps you create professional resumes and cover letters.
• Mindgrasp – An AI learning assistant that creates notes, summaries, and answers questions from any document or video.
• Clipnote – A tool that turns YouTube videos into notes, summaries, and social media posts with AI.
• Bento – A simple, elegant, and powerful personal search engine designed to organize and find your information.
• GummySearch – An AI-first customer research platform for finding problems and ideas within Reddit communities.
• Finalle – An AI-powered tool designed for performing deep financial analysis.
• AI Story Generate – A creative tool that allows you to generate personalized children’s books using AI.
• Shortform – A service that provides AI-powered, super-detailed summaries of articles and books.
• Promptsly – A tool designed to help you create the perfect, optimized prompts for any AI tool.
• Snowflake Cortex – A service that allows developers to build large language model applications directly within the Snowflake data platform.
• Together AI – A cloud platform that enables developers to build and run generative AI models.
• Fine-Tuner – A platform that allows you to finetune any open-source model with just a few clicks.
• Mindverse – An all-in-one content creation platform featuring AI-powered tools for writing, image creation, and research.
• TinyLlama – A compact, open-source 1.1B parameter language model trained on a massive dataset.
• Phi-2 – A small language model from Microsoft that demonstrates remarkable capabilities for its size.

Sponsors:

• Growth School – Learn how to use AI to become a 10X Product Manager.
• The Rundown AI – The Rundown is the world’s largest and fastest-growing AI newsletter, with over 600,000 smart and busy professionals reading it every day.
• HubSpot – HubSpot’s 2024 State of Marketing Report is your guide to navigating marketing and AI in 2024.
• Vercel – Vercel Ship is back for a special one-day conference on May 23rd, 2024, at the SFJAZZ Center in San Francisco.
• AE Studio – AE Studio’s B2B podcast booking service helps you connect with your ideal clients by securing you guest spots on relevant podcasts.
• Sling AI – Sling AI transforms lengthy videos into short, shareable clips, making content creation easier and faster.
• Brilliant – Brilliant is a fun and interactive way to learn real-world skills in AI, data, and more.
• FourthBrain – FourthBrain’s 12-week, live, online cohort-based programs are designed to help you level up your career with the most in-demand AI skills.
• Brave – Brave Search is a privacy-first search engine with its own independent index of the web.