Tag: Language Models

Microsoft, one of the Big Five tech companies, is developing its own Artificial Intelligence (AI) chips to reduce their reliance on graphics card maker Nvidia. Reports from The Verge suggest Microsoft has been working on Athena AI chips since 2019. These chips are currently being tested by both Microsoft and OpenAI employees for GPT-4, language models and more. It's a cost-effective way for Microsoft to equip their already-established AI initiatives. Satya Nadella, CEO of Microsoft, has been affiliated with the tech giant since 1992 and has seen record growth in Office 365 and Azure under his leadership.

Discover ERNIE Bot, the revolutionary ChatGPT-style AI chatbot created by Baidu itself. Baidu's Chief Technology Officer (CTO), Wang Haifeng, developed it completely in-house. It has natural language, visual and biological computing models and can do incredible things like creating business correspondence and recognizing and producing text from trillions of words. Wang Haifeng believes that creativity and education should stay at the forefront of AI technology.

This article explores how Chinese tech veterans are joining the AI start-ups race to keep up with ChatGPT, the OpenAI product released in November 2020. With an ambition to support the Chinese government's plan to strive for excellence in the field of AI, entrepreneurs such as Meituan's Wang Huiwei, Sogou's Wang Xiaochun and Sinovation Ventures' Lee Kai-Fu are turning to AI start-ups. Though a lack of resources has posed a challenge, they are determined to succeed with large language models and industrial use cases. Will these start-ups be able to catch up with the Big Tech firms like Baidu and Alibaba or OpenAI?

RedPajama is a collaborative project between Together, Ontocord.ai, ETH DS3Lab, Stanford CRFM, Hazy Research, and MILA Québec AI Institute to create freely licensable open-source large models. The data and scripts are based off the LLaMA recipe, allowing organisations to freely use it to pre-train models and push the boundaries of AI. This project originates from the need for open access to large language models from organisations of any size. It seeks to provide ethical and commercially viable models with openly available data. With their combined experience of open source research, Chris Re and Vipul Ved Prakash lead this project.

German photographer Boris Eldagsen sparks global debate with his award-winning AI generated photo. He consciously decided to submit the photo in order to explore what the event can handle in terms of content. Will the world adapt to and accept this new reality of AI-image-making? Read and get the checklist to spot a generated image here!