By
Paula ParisiOctober 4, 2024
Nvidia has unveiled the NVLM 1.0 family of multimodal LLMs, a powerful open-source AI that the company says performs comparably to proprietary systems from OpenAI and Google. Led by NVLM-D-72B, with 72 billion parameters, Nvidia’s new entry in the AI race achieved what the company describes as “state-of-the-art results on vision-language tasks, rivaling the leading proprietary models (e.g., GPT-4o) and open-access models.” Nvidia has made the model weights publicly available and says it will also be releasing the training code, a break from the closed approach of OpenAI, Anthropic and Google. Continue reading Nvidia Releases Open-Source Frontier-Class Multimodal LLMs
By
Paula ParisiOctober 3, 2024
OpenAI unveiled major updates at its DevDay conference with the focus largely on making AI more accessible, efficient and affordable. Included were four innovations: Vision Fine-Tuning in the API, Model Distillation, Prompt Caching and the public beta of Realtime API. The approach underscores OpenAI’s effort to empower its developer ecosystem even as it continues to compete for end-users in the enterprise space. The Realtime API gives developers the option of building “nearly real-time” speech-to-speech app experiences, selecting from among six OpenAI voices. Vision Fine-Tuning for GPT-4o enables customization of the model’s visual understanding of images and text. Continue reading OpenAI Showcases Latest Updates for Voice, Picture and More
By
Paula ParisiOctober 2, 2024
AI startup Liquid, founded by alums of MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL), has released its first models. Called Liquid Foundation Models, or LFMs, the multimodal family approaches “intelligence” differently than the pre-trained transformer models that dominate the field. Instead, the LFMs take a path of “first principles,” which MIT describes as “the same way engineers build engines, cars, and airplanes,” explaining that the models are large neural networks with computational units “steeped in theories of dynamic systems, signal processing and numeric linear algebra.” Continue reading MIT Spinoff Liquid Eschews GPTs for Its Fluid Approach to AI
By
Paula ParisiOctober 2, 2024
Snap Inc. is leveraging its relationship with Google Cloud to use Gemini for powering generative AI experiences within Snapchat’s My AI chatbot. The multimodal capabilities of Gemini on Vertex AI will greatly increase the My AI chatbot’s ability to understand and operate across different types of information such as text, audio, image, video and code. Snapchatters can use My AI to take advantage of Google Lens-like features, including asking the chatbot “to translate a photo of a street sign while traveling abroad, or take a video of different snack offerings to ask which one is the healthiest option.” Continue reading Snapchat: My AI Goes Multimodal with Google Cloud, Gemini
By
Paula ParisiOctober 1, 2024
The Allen Institute for AI (also known as Ai2, founded by Paul Allen and led by Ali Farhadi) has launched Molmo, a family of four open-source multimodal models. While advanced models “can perceive the world and communicate with us, Molmo goes beyond that to enable one to act in their worlds, unlocking a whole new generation of capabilities, everything from sophisticated web agents to robotics,” according to Ai2. On some third-party benchmark tests, Molmo’s 72 billion parameter model outperforms other open AI offerings and “performs favorably” against proprietary rivals like OpenAI’s GPT-4o, Google’s Gemini 1.5 and Anthropic’s Claude 3.5 Sonnet, Ai2 says. Continue reading Allen Institute Announces Vision-Optimized Molmo AI Models
By
Paula ParisiSeptember 27, 2024
The European Commission has released a list of more than 100 companies that have become signatories to the EU’s AI Pact. While Google, Microsoft and OpenAI are among them, Apple and Meta are not. The voluntary AI Pact is aimed at eliciting policies on AI deployment during the period before the legally binding AI Act takes full effect. The EU AI Pact focuses on transparency in three core areas: internal AI governance, high-risk AI systems mapping and promoting AI literacy and awareness among staff to support ethical development. It is aimed at “relevant stakeholders,” across industry, civil society and academia. Continue reading Amazon, Google, Microsoft and OpenAI Join the EU’s AI Pact
By
Paula ParisiSeptember 26, 2024
As OpenAI gears up to become a for-profit company next year, it is releasing ChatGPT Advanced Voice Mode, which brings a humanlike conversation mode to ChatGPT 4o. All U.S. subscribers to ChatGPT Plus and Team plans will gain access to the new feature, which will also be made available to those paying for ChatGPT Edu and Enterprise plans in the coming weeks. The firm is also adding five new voices and allowing customers to save personalized instructions for the voice assistant, including memory behaviors. Concurrently, executives including CTO Mira Murati have resigned as the company pivots to commerciality. Continue reading OpenAI Rolls Out Advanced Voice Mode Feature for ChatGPT
By
Paula ParisiSeptember 26, 2024
Microsoft has released a suite of “Trustworthy AI” features that address concerns about AI security and reliability. The four new capabilities include Correction, a content detection upgrade in Microsoft Azure that “helps fix hallucination issues in real time before users see them.” Embedded Content Safety allows customers to embed Azure AI Content Safety on devices where cloud connectivity is intermittent or unavailable, while two new filters flag AI output of protected material. Additionally, a transparency safeguard providing the company’s AI assistant, Microsoft 365 Copilot, with specific “web search query citations” is coming soon. Continue reading New Microsoft Safety Tools Fix AI Flubs, Detect Proprietary IP
By
Paula ParisiSeptember 25, 2024
Alibaba Cloud last week globally released more than 100 new open-source variants of its large language foundation model, Qwen 2.5, to the global open-source community. The company has also revamped its proprietary offering as a full-stack AI-computing infrastructure across cloud products, networking and data center architecture, all aimed at supporting the growing demands of AI computing. Alibaba Cloud’s significant contribution was revealed at the Apsara Conference, the annual flagship event held by the cloud division of China’s e-retail giant, often referred to as the Chinese Amazon. Continue reading Alibaba Cloud Ups Its AI Game with 100 Open-Source Models
By
Paula ParisiSeptember 24, 2024
Amazon has joined the ranks of firms offering generative video tools, although its release is aimed only at advertisers, at least for now. Simply called Video Generator, it can turn a product image into a video that showcases the product and even demonstrates its features, “leveraging Amazon’s unique insights to vividly bring a product story to life.” At the company’s Accelerate 2024 conference Amazon also debuted Live Image, which lets brands create animated GIFs from stills, a customizable chatbot assistant for third-party sellers, and a new AI-powered recommendation engine based on customer interests. Continue reading Amazon’s Video Generator Turns Stills into Advertising Clips
By
Paula ParisiSeptember 23, 2024
BlackRock has joined forces with Microsoft to launch what will initially be a $30 billion investment fund to finance AI infrastructure — concentrating primarily on building data centers and developing energy projects. The amount could quickly scale to about $100 billion. Abu Dhabi-based tech investment firm MGX is also participating, as is Global Infrastructure Partners (GIP), which owns, operates and invests across energy, transport, digital and waste management. BlackRock announced it is in the process of acquiring GIP, and says a deal expected to close next month. The new fund is called Global AI Infrastructure Investment Partnership (GAIIP). Continue reading BlackRock Teams with Microsoft to Advance AI Infrastructure
By
Paula ParisiSeptember 18, 2024
The OpenAI board’s Safety and Security Committee will become an independent board oversight committee, chaired by Zico Kolter, machine learning department chair at Carnegie Mellon University. The committee will be responsible for “the safety and security processes guiding OpenAI’s model deployment and development.” Three OpenAI board members segue from their current SSC roles to the new committee: Quora founder Adam D’Angelo, former Sony Corporation EVP Nicole Seligman and erstwhile NSA chief Paul Nakasone. OpenAI is currently putting together a new funding round that reportedly aims to value the company at $150 billion. Continue reading OpenAI Bestows Independent Oversight on Safety Committee
By
Paula ParisiSeptember 16, 2024
OpenAI is previewing a new series of AI models that can reason and correct complex coding mistakes, providing a more efficient solution for developers. Powered by OpenAI o1, the new models are “designed to spend more time thinking before they respond, much like a person would,” and as a result can “solve harder problems than previous models in science, coding, and math,” OpenAI claims, noting that “through training, they learn to refine their thinking process, try different strategies, and recognize their mistakes.” The first model in the series is being released in preview in OpenAI’s popular ChatGPT and in the company’s API. Continue reading OpenAI Previews New LLMs Capable of Complex Reasoning
By
Paula ParisiSeptember 13, 2024
Adobe is showcasing upcoming generative AI video tools that build on the Firefly video model the software giant announced in April. The offerings include a text-to-video feature and one that generates video from pictures. Each outputs clips of up to five seconds. Adobe has developed Firefly as the generative component of the AI integration it is rolling out across its Adobe’s Creative Cloud applications, which previously focused on editing and now, thanks to gen AI, incorporate creation. Adobe wasn’t a first-mover in the space, but its percolating effort has been received enthusiastically. Continue reading Adobe Publicly Demos Firefly Text- and Image-to Video Tools
By
Paula ParisiSeptember 10, 2024
YouTube is introducing AI detection tools designed to allow people to learn when their face and/or voice are copied and used in third-party videos. As part of the effort, YouTube’s existing Content ID program that protects copyrighted music will expand to include more broad-based voice simulation detection technology. The new tools aim to protect “people from a variety of industries — from creators and actors to musicians and athletes,” according to the company. The Google-owned platform is also coming up with a way to address unauthorized use of its content for training AI models. Continue reading YouTube Adding Tools to Protect Against Unauthorized AI Use