Meta Platform has upped the tempo to keep up with other tech giants in the artificial intelligence arms race. The giant social networking company has unveiled an upgraded version of its powerful Llama artificial intelligence model aimed at Microsoft-backed OpenAI and Alphabet’s Google.
Llama Upgrade
Llama 3 is the upgraded artificial intelligence language model version unveiled last summer. The large language model is designed explicitly for developers, researchers, and businesses to help them build, experiment, and scale various artificial intelligence ideas. Consequently, it is the bedrock of innovation amid his artificial intelligence developments.
The open-source artificial intelligence model has already been downloaded more than 170 billion times despite having some restrictions, all but underlining its popularity. Its popularity stems from the fact that Meta also uses it to run some of its popular products, including Facebook, WhatsApp and Rapid Ban smart glasses, to enhance user’s experience.
The unveiling of Llama 3 comes on the heels of reports that Meta Platform plans to unveil a much smaller version of the Llama language model to offer a more cost-effective way of developing AI models. Google has already unveiled a much smaller version of the Gemma family model, and the French company Mitral has also unveiled the Mistral 7B.
The smaller version of the AI large language models are not designed to handle long strings of information or instructions. However, they have proved faster, more flexible, and cheaper to run than the regular-sized model. The smaller models are proving effective as they can be deployed in specific projects like code assistance or small devices like phones and laptops.
The development of different forms of large language models comes as companies race against time to develop AI products and services. The companies are spending billions of dollars to build large language models to enhance their products and solutions and strengthen their competitive edge. For instance, Meta uses Llama to power its content feeds and recommendations.
Additionally, Llama has proved to greatly benefit in enhancing targeted advertising on the company’s platforms, therefore allowing it to attract more ad spending. Accordion Meta Chief Product Officer Chris Cox says they are no longer playing catch-up as Llama has emerged as an industry-leading benchmark in various AI models.
Just like Meta Platform’s Llama, Microsoft’s coding assistant Copilot, based on OpenAI’s generative intelligence, is also revolutionizing the working lives of software engineers. The AI-powered assistant has attracted over 1.3 million customers, including 50,000 businesses from small firms to big enterprises like Goldman Sachs, Ford and Ernst & Young. Microsoft Copilot stands out as it helps engineers easily handle tedious and repetitive tasks.
TSMC Warning
Meanwhile, Taiwan Semiconductor has cut its market outlook, warning that smartphone and personal computer markets remain weak. The warnings starkly contrast with initial confidence that artificial intelligence will allow chip giants to continue delivering impressive numbers. In the long run, investors expect AI-focused chips to account for the most significant share of the company’s revenues.
Revenues generated from the company’s artificial intelligence chips are growing at a 50% rate. Despite the robust growth, some investors are still determining whether the growth is sustainable in the long run.
