Welcome to the latest dispatch from AI Hungry, where we delve into the evolving landscape of artificial intelligence. This edition brings to light the formation of the AI Alliance, a new coalition led by Meta and IBM, aimed at promoting open AI development. With heavyweight members like AMD and Intel, and backing from academia, this alliance challenges the dominance of proprietary models in AI.
In parallel, we explore the ambitious moves of Elon Musk's AI startup, X.AI, as it seeks a staggering $1 billion in funding. The launch of their new chatbot, Grok, positions them as a formidable player in the AI chatbot arena, competing with the likes of Google's Bard and Anthropic's Claude. These stories underscore the dynamic and competitive nature of AI innovation, with implications for the future of technology and society.
Tech Giants and Academia Form AI Alliance to Foster Open AI Development
Meta and IBM have spearheaded the creation of the AI Alliance, a coalition of over 50 entities, to encourage open AI development and offer an alternative to OpenAI's ChatGPT. Members include AMD, Intel, Oracle, and various universities, with AMD set to reveal new AI hardware to challenge Nvidia's market dominance. The alliance aims to foster a collaborative, open-source AI ecosystem, contrasting with the proprietary models of companies like OpenAI and Microsoft.
The AI Alliance emphasizes the importance of sharing and building upon AI advancements within a broad community. This initiative seeks to ensure the AI field remains inclusive, fostering innovation, transparency, and accountability. The move is seen as a counter to the concentrated influence of a few major organizations in the AI market, with the Alliance advocating for a future of AI shaped by a wider range of contributors.
Elon Musk's AI Startup X.AI Seeks $1 Billion in Funding Amid Launch of New Chatbot Grok
Elon Musk's artificial intelligence startup, X.AI, is making significant strides by filing with the SEC to raise up to $1 billion through an equity offering. The company, which is part of Musk's extensive portfolio including Tesla, SpaceX, Boring Company, and Neuralink, has already secured nearly $135 million from investors. The recent filing indicates that X.AI might have already achieved its funding goal, based on a binding agreement mentioned in the document.
In addition to the funding news, X.AI has introduced a new AI chatbot named Grok. This chatbot is set to compete with other advanced AI technologies such as Google's Bard and Anthropic's Claude. This development comes as Musk announced that investors in his other venture, X (formerly known as Twitter), would hold a 25% stake in X.AI. The launch of Grok marks a significant step in the AI field, aligning with the vision of online journalist Matthias, who predicts a transformative future for human-computer interactions.
💡 EnCharge AI Raises $22.6M to Revolutionize AI Chip Technology EnCharge AI has raised $22.6 million to develop a new chip architecture that could boost AI performance by 15x and reduce energy use by 90%. Their technology aims to make AI more accessible and cost-effective across various industries. (Link)
🎉 Microsoft's Copilot Turns One: Unveils GPT-4 Turbo and Enhanced AI Features Microsoft celebrates Copilot's first anniversary with new updates, including the faster and more efficient GPT-4 Turbo, integration of GPT-4V for Bing image searches, and a feature-rich Copilot sidebar in Edge. (Link)
💡 Mistral AI Nears $2 Billion Valuation with Major Funding Round French AI startup Mistral AI is on the brink of a $2 billion valuation after raising €450 million, led by Andreessen Horowitz. Their open-source language model, Mistral 7B, surpasses competitors like Meta's Llama 2. (Link)
⚡ ETH Zurich Innovates with UltraFastBERT to Accelerate AI Language Models ETH Zurich researchers have developed UltraFastBERT, a method that significantly speeds up AI language models, achieving up to 96% of the original BERT model's performance while using only 0.3% of the neurons. This innovation offers vast potential for large models like GPT-3, though software optimizations are needed. (Link)
🔒 Over 1,500 HuggingFace API Tokens Exposed, Meta Included Cybersecurity startup Lasso Security discovered that 1,500+ HuggingFace API tokens, including Meta's, are exposed, risking attacks on AI models and datasets used by millions. (Link)
⚖️ UK Court Advances Getty Images' Copyright Lawsuit Against Stability AI Getty Images is suing Stability AI for allegedly using copyrighted images to train its AI model without permission. The UK court has rejected Stability AI's dismissal request, citing the need for a trial to investigate the claims. (Link)
