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:

• Apple is planning an AI-focused M4 chip to overhaul its Mac line – Apple is preparing to update its entire line of Mac computers with a new M4 family of processors designed to highlight artificial intelligence capabilities.
• OpenAI partners with Color Health to build a GPT-4-based copilot for cancer screening – OpenAI is collaborating with health technology company Color Health to use GPT-4 to create personalized cancer treatment plans and improve care for patients.
• Google’s VLOGGER AI can create a realistic video of a person from a single photo – Google researchers have developed VLOGGER, an AI model that can generate a high-quality, controllable video of a person talking and moving from just one still image and an audio clip.
• Google and OpenAI reportedly used YouTube content to train AI models – A new report claims that OpenAI transcribed over a million hours of YouTube videos with its Whisper tool to train GPT-4, and Google also used YouTube transcripts, raising questions about data usage policies.
• Google announces Axion, its first custom Arm-based CPU for data centers – Google has unveiled Axion, a custom-designed ARM-based processor that it claims offers industry-leading performance and energy efficiency for AI and other data center workloads.
• Mistral quietly releases a powerful new open-source model, Mixtral 8x22B – AI company Mistral released a new 176B parameter Mixture-of-Experts model, Mixtral 8x22B, via a torrent link, which is considered one of the most powerful open-source models available.
• Hume AI launches the first “emotionally intelligent” voice AI – Startup Hume has released an API for its empathic voice interface (EVI), an AI that can understand the emotional tone of speech and respond with appropriate emotional inflections.
• Reviews for the Humane AI Pin are overwhelmingly negative – Early reviews from major tech journalists criticize the Humane AI Pin for being slow, unreliable, and having poor battery life, concluding it is not yet a viable replacement for a smartphone.
• Luma AI releases Genie, a text-to-3D model that generates 3D objects in seconds – Luma AI has launched Genie, a new text-to-3D tool that can create a 3D model with textures and mesh from a text prompt in under 10 seconds.
• Amazon CEO Andy Jassy says generative AI will be the company’s next major pillar of growth – In his annual letter to shareholders, Amazon’s CEO detailed how the company is investing heavily in generative AI at every level, from foundational models to AI-powered applications like Alexa and Amazon Q.
• Sam Altman is meeting with UAE investors to fund his AI chip ambitions – OpenAI CEO Sam Altman is seeking to raise trillions of dollars from global investors, including those in the UAE, to build a network of AI chip fabrication plants to meet future demand.
• A developer fine-tuned Mistral 7B to create a stock trading agent – A programmer successfully trained a Mistral 7B model to make stock trading decisions based on news headlines, with backtesting showing it outperformed a simple buy-and-hold strategy.
• Waymo’s fully autonomous ride-hailing service is now open to the public in Los Angeles – Anyone in Los Angeles can now download the Waymo One app and hail a ride in its driverless vehicles across a 63-square-mile service area.
• Microsoft Research unveils VASA-1, an AI that creates lifelike talking faces from a single photo – Microsoft’s VASA-1 model can take a single portrait image and an audio clip to generate a hyper-realistic video with precise lip sync, expressive facial nuances, and natural head movements.
• Intel launches Gaudi 3 AI chip to compete with Nvidia’s H100 – Intel has released its Gaudi 3 AI accelerator, which it claims offers 50% better inference and 40% better power efficiency than Nvidia’s H100 GPU at a fraction of the cost.
• Meta announces its next-generation custom AI chip – Meta has revealed the latest version of its Meta Training and Inference Accelerator (MTIA), a custom-built chip designed to power its AI ranking and recommendation models more efficiently.
• AI startup Reka launches Core, a powerful multimodal model rivaling GPT-4 – Reka has released its flagship model, Reka Core, a highly capable multimodal AI that can process and understand text, images, audio, and video, putting it in direct competition with top models from Google and OpenAI.
• A new AI agent can autonomously fix bugs and add features to codebases – Researchers have created SWE-agent, an AI system that achieved a 12.29% success rate in resolving real-world GitHub issues, demonstrating a significant step towards AI-driven software engineering.
• Meta is preparing to launch its Llama 3 open-source AI model within the next month – Meta is expected to release multiple versions of its next-generation Llama 3 model soon, aiming to be more open and competitive with proprietary models like GPT-4.
• Together AI raises $106M to build the leading cloud platform for open-source AI – Together AI, a platform providing infrastructure for developers to build on open-source AI models, has secured $106 million in a new funding round led by Salesforce Ventures.
• Perplexity launches Enterprise Pro, a secure AI research assistant for businesses – Conversational search engine Perplexity has introduced an enterprise-grade version of its service with enhanced security features like SOC2 compliance, data encryption, and user management.
• Stability AI appoints new interim co-CEOs – Following the resignation of founder Emad Mostaque, Stability AI has named its COO and CCO as interim co-CEOs to lead the company.
• Amazon Q, the AI-powered assistant for work, is now generally available – Amazon has officially launched Amazon Q, an AI assistant for businesses with new features like Q Apps for creating custom AI-powered applications and enhanced capabilities for developers.
• Adobe previews new AI tools for music creation and editing – Adobe is developing a new suite of generative AI tools that will allow users to generate, extend, and edit music within its video and audio software using text prompts.
• ElevenLabs launches a public beta for its new AI sound effects generator – Voice AI company ElevenLabs has unveiled a new tool that can create a wide range of sound effects from simple text descriptions.
• OpenAI expands its custom model program for enterprises – OpenAI is broadening access to its program that helps companies fine-tune and build custom GPT-4 models tailored to their specific industry and use cases.
• Klarna expands its OpenAI-powered customer service assistant to more countries – The fintech company is rolling out its AI assistant to 10 additional regions, which now handles two-thirds of all customer service chats and is equivalent to 700 full-time agents.
• Spotify tests an AI playlist creation feature for Premium subscribers – Spotify is beta testing a new feature in the UK and Australia that allows users to generate custom playlists by typing descriptive prompts into the app.
• Researchers explore how GPT-4V is learning to “see” the world like humans do – A new paper shows how multimodal models like GPT-4V can process sequences of images to understand dynamic scenes, enabling applications like AI assistants for the visually impaired.
• Study finds GPT-4 falls short of human doctors in complex medical diagnoses – Research comparing diagnostic accuracy found that while GPT-4 performed reasonably well, human physicians were significantly better at differential diagnosis, especially in more complex cases.
• An essay argues that AI could make human artists more valuable, not less – The piece suggests that as AI floods the world with synthetic content, the demand and appreciation for authentic, human-created art with a compelling backstory will increase.
• Researchers develop a method to remove specific concepts from large language models – A new technique called “representation zapping” can effectively erase concepts like copyrighted material or specific biases from an AI model without degrading its overall performance.
• Building a mental model of what Artificial General Intelligence (AGI) might look like – A blog post speculates on the development path of AGI, drawing parallels to how a human child learns, and explores the potential capabilities of such a system.
• xAI open sources the JAX implementation for its Grok-1 model – Following the release of Grok-1’s weights, xAI has now published the JAX-based code used to run the 314-billion parameter model.
• Vercel’s v0 is a generative UI tool that creates React components from text prompts – Vercel has launched v0, a tool that allows developers to generate copy-and-paste-friendly React code for user interface components using simple text and image descriptions.
• Replicate introduces simplified fine-tuning for open-source LLMs – The cloud AI platform Replicate has added a new feature that lets developers fine-tune popular open-source models like Llama and Mixtral on their own data with just a few clicks.
• Continue is an open-source tool that lets you chat with your entire codebase – This open-source extension for VS Code and JetBrains helps developers by allowing them to ask questions and get assistance from LLMs that have the full context of their project.
• LoRA Land is a new open-source hub for sharing efficient model adapters – Hugging Face has launched LoRA Land, a central place for the community to discover, share, and use LoRA models for efficiently fine-tuning large AI models.
• Diffus is a new open-source tool for creating AI-powered animations – This tool provides an interface for generating AI animations with precise control over camera movements, styles, and negative prompts.
• A developer’s guide to creating a custom GPT-powered Slack bot – A GitHub project provides the necessary code and instructions for building a personalized ChatGPT bot that can be integrated into a Slack workspace.
• LiveCodeBench is a new benchmark for evaluating real-time AI code generation – Researchers have introduced a new testing framework to measure the ability of AI models to perform code completion tasks in a live, interactive setting, similar to how developers actually code.
• A video tutorial explains how to properly test Large Language Models – This guide from the Weaviate vector database team covers various methods for evaluating LLMs, including testing for hallucinations, bias, and overall performance.

Trending AI Tools:

• Luma Dream Machine – A publicly available text-to-video model that generates high-quality, five-second video clips from text and image prompts.
• Claude 3.5 Sonnet – Anthropic’s latest and fastest AI model, outperforming competitors like GPT-4o and introducing a new “Artifacts” feature for interactive content creation.
• ElevenLabs – An AI platform for generating realistic text-to-speech voices and creating sound effects from text prompts.
• Krea AI – A creative suite that offers real-time AI-powered image and video generation, enhancement, and upscaling.
• Pika – An AI video platform that can generate and edit videos from text, and recently added a feature for automatic lip-syncing to any audio track.
• Suno – An AI music creation tool that generates songs with vocals and instrumentation from simple text prompts, recently updated to version 3.5.
• Udio – An AI music generation tool that creates high-quality songs from text prompts, complete with lyrics and vocals.
• Perplexity – An AI-powered search engine that provides direct, conversational answers to questions by synthesizing information from the web.
• Midjourney – A popular AI image generator known for producing highly artistic and stylized visuals from text prompts.
• Runway – A creative suite of AI tools for video editing, effects, and generation, including their new high-performance Gen-3 Alpha model.
• Sora – OpenAI’s highly anticipated but unreleased text-to-video model, known for its cinematic and realistic video generation capabilities.
• Apple Intelligence – Apple’s new personal intelligence system integrated into its operating systems to enhance writing, image creation, and on-device actions.
• Microsoft Copilot+ PCs – A new category of Windows PCs designed with powerful processors to run advanced AI features locally and efficiently.
• Kling – A high-fidelity text-to-video model from Chinese company Kuaishou, capable of generating videos up to two minutes long in high resolution.
• Viggle – An AI tool that allows you to animate static characters with motion from a reference video, maintaining character consistency.
• Arc Search – A mobile browser that uses AI to “browse for you,” creating a single, clean page that directly answers your search query.
• Recast – An AI tool that summarizes any online article into a concise audio podcast, allowing you to listen to content on the go.
• NVIDIA ACE – A suite of technologies designed to help developers create lifelike and interactive digital humans for games and applications.
• Ideogram – An AI image generator known for its exceptional ability to render coherent and accurate text within images.
• Vercel v0 – A generative user interface system by Vercel that creates React components and UI designs from text prompts.
• ChatGPT – OpenAI’s flagship conversational AI, capable of understanding and generating human-like text, images, and voice.
• HeyGen – An AI video platform that allows you to create professional videos with AI-generated avatars and voices.
• Firefly – Adobe’s family of creative generative AI models designed for creating images, vectors, and text effects safely for commercial use.
• Gamma – An AI-powered tool for creating polished and engaging presentations, documents, and webpages from simple text prompts.
• Consensus – An AI-powered search engine that extracts and synthesizes findings directly from scientific research papers to answer your questions.
• Explainpaper – A tool that helps you read academic papers more easily by letting you highlight confusing text and get an AI-generated explanation.
• Polycam – A 3D scanning application that uses your phone’s camera to capture objects and spaces and turn them into 3D models.
• Splash Pro – An AI music platform that allows you to create unique, royalty-free music and custom AI-generated vocals for your projects.
• GitHub Copilot – An AI pair programmer that suggests code and entire functions in real-time, right from your editor.
• Fine-Tuner – A no-code platform that allows users to create and manage their own fine-tuned AI models.
• Mindgrasp – An AI learning assistant that can instantly create accurate notes and answer questions from any document, presentation, or video.
• Notion AI – A set of AI features integrated into the Notion workspace that can help with writing, summarizing, brainstorming, and organizing information.
• Gemini – Google’s flagship multimodal AI model that can understand and operate across text, images, audio, video, and code.
• Llama 3 – The latest generation of Meta’s open-source large language models, designed to be highly capable for a wide range of applications.
• Prompter – An AI-powered tool that helps you write better prompts for large language models to get more accurate and desired outputs.
• Tome – An AI-powered storytelling and presentation tool that helps you build compelling narratives from scratch.
• Heyday – An AI-powered memory assistant that automatically resurfaces content you’ve seen before while you browse the web.
• FigJam AI – AI features within Figma’s online whiteboard that help teams brainstorm, summarize, and generate ideas more efficiently.
• Poe by Quora – A platform that provides access to a variety of AI chatbots, including models from OpenAI, Anthropic, and Google, in one place.
• Looka – An AI-powered platform that helps you design a custom logo and build a brand identity for your business.
• Saga – An AI-powered workspace that automatically organizes your notes, documents, and tasks, connecting related information for you.
• Leonardo AI – A comprehensive AI platform for creating a wide range of visual assets, including images, art, and game textures.
• Kaiber – An AI video generation tool that transforms your ideas, images, and videos into stunning visual stories.
• Visual Electric – An AI image generator designed for creative professionals to quickly explore and develop visual ideas.
• Gencraft – An AI-powered art and video generator that allows users to create unique visuals from text descriptions.
• Augie – An AI video creation platform for businesses and creators, making it easy to turn words, clips, and images into engaging videos.
• Defog – An AI tool that converts natural language questions into SQL queries, helping users retrieve data from databases without writing code.
• Tavily AI – An AI-native search API designed to provide real-time, accurate, and factual information for large language models.
• Co-pilot – Microsoft’s AI assistant integrated across its products to help with coding, writing, data analysis, and more.
• Superhuman – An AI-powered email client designed for speed, helping users get through their inbox twice as fast.

Sponsors:

• Masterworks – An investment platform that allows you to invest in shares of multi-million dollar paintings by artists like Banksy and Basquiat.
• Modal – Provides cloud infrastructure for running generative AI, batch jobs, and other demanding compute tasks.
• Vercel – A platform for frontend developers, providing the speed and reliability innovators need to create at the moment of inspiration.
• Together AI – Offers the fastest inference platform for running open-source models like Llama 3.
• CommandBar – Helps companies add a powerful AI user assistant to their product, improving user activation and feature adoption.
• AE Studio – A development, data science, and design studio that partners with founders and executives to build and ship innovative products.
• LM Studio – Allows you to discover, download, and run local LLMs from your laptop.
• Marketing Against The Grain – A newsletter and podcast by HubSpot CMO Kipp Bodnar and Zapier CMO Kieran Flanagan, covering marketing trends.
• VAST Data – Provides a data platform designed for the AI era, simplifying data management for enterprises and service providers.
• Vanar – A blockchain platform focused on entertainment and mainstream adoption, offering a fast and low-cost solution for brands.
• Midjourney – An independent research lab that produces a proprietary artificial intelligence program that creates images from textual descriptions.