• AI KATANA
  • Posts
  • Meta to release commercial AI model in effort to catch rivals

Meta to release commercial AI model in effort to catch rivals

Also: Nvidia invests in biotech company Recursion for AI drug discovery

Hello!

In today's AI news roundup, we spotlight significant commercial advancements across sectors. Meta is set to launch an open-source commercial AI model, which aims to compete with OpenAI and Google, while augmenting its existing platforms with AI capabilities. Nvidia is investing $50 million into biotech company Recursion Pharmaceuticals to accelerate its AI-driven drug discovery process, reflecting growing interest in AI integration in pharmaceuticals. In a major alliance, KPMG and Microsoft commit to integrating AI and cloud services into professional services, promising digital solution development and acceleration for KPMG's global workforce. Tomasz Tunguz likens the role of large language models (LLMs) to 'green fracking equipment' in the data industry, emphasizing their potential to mine value from unstructured text.

Slicing on the chopping board:

  • ๐Ÿƒ๐Ÿปโ€โ™‚๏ธ Meta to release commercial AI model in effort to catch rivals

  • ๐Ÿ’Šย Nvidia invests in biotech company Recursion for AI drug discovery

  • ๐Ÿงพย KPMG and Microsoft enter landmark agreement to put AI at the forefront of professional services

  • โ„น๏ธ The Fracking of Information

๐Ÿƒ๐Ÿปโ€โ™‚๏ธ Meta to release commercial AI model in effort to catch rivals

Meta is set to launch a commercial version of its AI model, aiming to rival OpenAI and Google. The AI model will be open-source and widely available, offering a platform for start-ups and businesses to build custom software. This move is part of Meta's broader ambition to enhance its AI capabilities and utilise its large language models (LLMs) across its platforms, such as Instagram, WhatsApp, and Facebook, to power AI chatbots. Despite potential concerns about misuse and intellectual property issues with open-source AI, Meta believes this approach will help catch up with competitors and alleviate concerns about AI.

Chipmaker Nvidia has committed to investing $50 million in biotech firm Recursion Pharmaceuticals to accelerate the latter's AI-based drug discovery efforts. The investment led to a significant rise in both companies' stock values. Recursion, which collaborates with other drug manufacturers like Roche and Bayer, applies AI to identify and design new therapies. The company plans to use Nvidia's cloud platform to train its AI models using massive biological and chemical datasets. These AI models could then be licensed on Nvidia's BioNeMo, a cloud service for generative AI in drug discovery. Recursion, presently conducting human trials for five of its drugs, foresees using BioNeMo to support its internal drug pipeline and those of its partners. This deal highlights the growing interest in integrating AI into the pharmaceutical industry to expedite the delivery of life-saving treatments.

KPMG and Microsoft have unveiled a significant expansion of their global partnership, committing to integrate AI and cloud services into professional services like audit, tax, and advisory. The multi-year agreement involves a multibillion-dollar investment from KPMG into Microsoft's cloud and AI services over the next five years, expecting to trigger over US$12 billion of potential growth for KPMG. The alliance aims to modernize workforce, ensure secure AI usage, and innovate solutions for clients, industries, and society. Microsoft's technologies will empower KPMG's global workforce of 265,000, enabling faster analysis and strategic advice. As early access partners to Microsoft 365 Copilot and Azure OpenAI Service, KPMG professionals will pilot the technologies across the organization to enhance client engagements and accelerate digital solution development. The partnership also includes a commitment to supporting businesses with their ESG agendas, enhancing the commercial prospects of the agreement, and contributing to social and community impact worldwide.

โ„น๏ธ The Fracking of Information

Large language models (LLMs) have revolutionized the extraction of value from unstructured text files, akin to fracking in the oil industry. However, they struggle with answering quantitative queries in their current form. The proposed solution is a constellation of models, where initial models classify a query and delegate it to the right ML model for resolution. Out-of-the-box summarization, another powerful feature of LLMs, aids in transforming complex outputs into easily digestible formats. Further, the capability of these models to identify and cite source documents adds credibility and limits "hallucinations" or invented responses without data basis. The use of LLMs in knowledge management systems will be omnipresent, appearing in emails, presentations, blog posts, etc., with respect to the governance, permissions, and policies of a business. In essence, if data is the new oil, LLMs are the green fracking equipment, mining value from the dense substrate of unstructured text.

๐Ÿ› ๏ธ AI tools updates

Google's AI chatbot, Bard, is expanding its reach to Europe and Brazil, introducing new features and capabilities to challenge OpenAI's ChatGPT. The Bard update allows users to collaborate in over 40 languages, offers adjustable tone and style of responses, and enables users to pin, rename, export code, and use images in prompts. Despite facing initial privacy regulations in the EU, Google reassured local watchdogs about the transparency, choice, and control in the use of Bard. Google has also confronted a class action in the U.S. over potential misuse of users' personal information.

Canadian e-commerce company Shopify has announced its plans to roll out an AI assistant named "Sidekick" for its platform's merchants, joining other technology firms such as Alibaba, Zoom Video Communications, and Databricks that have recently launched similar AI systems. The assistant, to be embedded as a button on Shopify, will assist with merchant queries, provide information on sales trends, and help entrepreneurs update their online stores, including applying discounts if requested.

๐Ÿ’ต Venture Capital updates

Israeli startup Vendict, led by founders Udi Cohen and Michael Keslassy, has raised $9.5 million in funding to revolutionize security compliance processes in enterprise software. The firm aims to replace time-consuming manual security questionnaire completion with a unique generative AI model that uses the vendor's own data, potentially saving hundreds of work hours each month. The AI, trained in security language and combined with other leading models including Microsoft Azure's rephrasing LLM, responds professionally and accurately to security questionnaires, effectively reducing security assessments from weeks to mere hours.

San Francisco-based voice AI startup, Resemble AI, has secured $8 million in a Series A funding round, led by Javelin Ventures, with participation from Comcast Ventures, Craft Ventures, and Ubiquity Ventures. Resemble AI's generative AI technology enables users to clone voices, and currently has over 1 million individual users and 200 business clients, including music producers, game studios, and voice actors.

๐Ÿซก Meme of the day

โญ๏ธ Generative AI image of the day

Before you go, check out the new collaboration between AP and Open AI.