StartupTalky

Tag: Llama 3

  • MeitY Requests Ideas for Developing India’s Own AI Foundation Paradigm

    Abhishek Singh, additional secretary, Union Ministry of Electronics and Information Technology (MeitY), stated during his visit to GIFT City in Gujarat on 27 January that India is seeking proposals for developing its own foundational model in order to compete with foreign AI foundational models that serve as the basis for numerous applications such as ChatGPT and Gemini. He added that how India creates a basic model has been a major problem.

    At the moment, every core model that a nation uses, including ChatGPT, Llama 3, Gemini, and Claude, is foreign. China has now developed Deepseek and other models as well. To create an Indian foundational model, India must invest and offer financial assistance. The foreign fundamental models can provide incorrect or unsuitable answers because they were trained on western data sets and are not adapted to Indian languages and contexts.

    Speaking at the opening of the event, where Chief Minister Bhupendra Patel was also in attendance, Singh went on to say that the government is putting out a call for proposals to create a foundational model in India so that India can compete with the best in the world. Singh continued without revealing a deadline for the submissions, saying, “I am confident that some of the start-ups and innovators present today will be joining this challenge to build an Indian foundational model.”

    Foundational Model and Risks of AI

    One kind of artificial intelligence model that can handle a variety of jobs is the foundational model. These models can be used in a wide range of applications because they are developed by training on large and varied datasets. While AI has a lot of potential, Singh also believed that there are a lot of risks associated with it. The nation has witnessed how deepfakes and false information operate, as well as how incorrect AI reactions can lead to issues.

    To do this, the government is developing instruments for creating AI that is ethical, responsible, safe, and trustworthy. Tools for identifying algorithmic biases, false information in AI, and deepfakes are being developed by researchers and organisations. CoEs and any start-ups developing it will also have access to these technologies. Prime Minister Narendra Modi will visit Paris from February 10–11, when France and India will co-host the AI Action Summit, Singh said, adding that India was engaging “globally in the development of AI governance guidelines.”

    India to Set up 18000 Graphic Processing Units

    According to the additional secretary, India’s IndiaAI Mission aims to establish 18,000 Graphic Processing Units (GPUs) in order to create a high-end, scalable AI computing ecosystem that will assist researchers, academicians, students, and start-ups. In comparison to other nations, India has relatively few GPUs.

    In order to achieve its goal of setting up 10,000 GPUs, the government has held bids under the AI Mission and collaborated with the industry. It is aiming for 18,000 GPUs at very competitive prices, and it has opened the financial bids last week. Additionally, all researchers and startups will have access to these GPUs at a further discounted price. This will also help the start-ups operating in the CoE, Singh continued.

    How Artificial Intelligence Is Transforming Business
    Artificial Intelligence is a critical factor in the strategy of those who want to expand their business impact in this digital era to make a win.
    StartupTalkyAvinash kumar mahato

  • Meta Launches Largest Llama 3 AI Model, Citing Language and Math Advancements

    On Tuesday, Meta Platforms (META.O) unveiled the most recent version of its Llama 3 AI models, which are primarily available for free. These models compete with paid models from competitors like OpenAI and offer superior multilingual skills and general performance metrics.

    Announcing the release of the new Llama 3 model, the parent company of Facebook stated in blog posts and a research study that it can talk in eight languages, produce higher-quality computer code, and solve more complicated math problems than earlier versions.

    Although it is still smaller than top models supplied by competitors, this new version outshines the previous one with 405 billion parameters, or variables, that the algorithm considers to produce answers to user questions.

    During his extensive promotion of Llama 3, Facebook CEO Mark Zuckerberg stated he expects that the future Llama models will surpass proprietary competition by the end of next year. He claimed that hundreds of millions of people were already using the Meta AI chatbot, which was driven by those models, and that it was on course to become the most popular AI assistant by the end of the year.

    Competition Is Getting Stiffer

    The release aligns with the ongoing competition among technology companies to demonstrate that their expanding portfolios of resource-intensive large language models can generate substantial improvements in well-established problem areas, such as advanced reasoning, to justify the substantial investments they have made in them.

    According to Meta’s leading AI scientist, these models will eventually run into reasoning limitations, necessitating the development of new AI systems.

    Amazon is reportedly working on a model with two trillion parameters, whereas OpenAI’s GPT-4 model reportedly has one trillion.

    Meta announced that it is updating its flagship 405 billion parameter model as well as its 8 billion and 70 billion parameter Llama 3 models, which debuted in the spring but are reduced in weight, with new versions coming out soon.

    Meta has announced that it will be releasing revised versions of its flagship 405 billion parameter model, as well as its 8 billion and 70 billion parameter Llama 3 models, which debuted in the spring. The models are lighter in weight.

    With an expanded “context window,” the three new models can handle larger user requests and are multilingual.

    Benefits Llama offers to its users

    Meta makes its Llama models available to developers for little or no cost, an approach that Zuckerberg claims will lead to better products, less reliance on potential rivals, and higher engagement with the company’s primary social networks.

    If developers choose the company’s free models over its paid ones, it would hurt its competitors’ business models and benefit the company too. Meta highlighted potential improvements in important mathematics and knowledge exams in its announcement, which may increase the attraction of that possibility.

    How to Use Meta Business Suite for Effective Management
    Meta Business Suite is designed to help businesses manage their presence on Facebook and Instagram. Learn what it is and how to use it.
    StartupTalkyShalini Rath