Meta works on latest AI language model

That Meta Platforms – one of the most famous technology conglomerates – is using AI is hardly surprising. Photo filters, ad recommendations or blocking sensitive content are just examples of simple, widely available technologies. However, in the face of growing competition from OpenAI, Meta has taken decisive steps. Most noteworthy is their latest project, to be released in early 2024, which is expected to outclass OpenAI.

Meta Platforms, OpenAI – this is what you should know

With a word of introduction and chronicling reliability, we will briefly outline the profiles of the two companies. Meta Platforms is currently the largest social media-related company. If the name “Meta” itself doesn’t tell you anything, more light is shed by their former name: Facebook Inc. Meta Platforms’ portfolio consists mainly of instant messaging and social media: WhatsApp, Instagram, Facebook or Messenger. Together with Google, Microsoft, Apple and Amazon, they make up Big Tech – the “big five” in the IT industry. Joining this elite group is the increasingly bold step of OpenAI, a technology lab founded less than eight years ago. Their specialty, on the other hand, is broadly defined artificial intelligence to make people’s lives easier. The result of their work was the well-known ChatGPT and DALL-E, an image generator based on the inputted command. These projects have been such a huge success that Microsoft has decided to invest more than $10 billion in what it calls a “new” project. multi-year plan. This clearly posed a threat to the position of Meta, which uses its own artificial intelligence models on a daily basis.

Technological race in full swing

Less than a week ago, the Wall Street Journal and Reuters reported sensational news straight from Silicon Valley. Meta is working on a “new, much stronger AI model” that with its solutions “is expected to match and even surpass the tools made available by OpenAI.” The information is credible in that it comes from a trusted source directly involved in the project work, Reuters reports. The Wall Street Journal has tagged the news with a bold but, in our opinion, accurate claim: escalation of the technology race. At this point we know two rather vague pieces of information. The first is the date the project will be available for use, which is early 2024 (quote: early 2024). The second piece of information outlines the specifics of the project: a large language model, which leads us to believe that it will be a desire to outclass ChatGPT. It is worth mentioning that this is not the first attempt by the Meta conglomerate in the area of language models – their Llama 2 project was released in February 2023. Sources cited by Reuters and the Wall Street Journal say their new model will far surpass the currently available Llama 2 and prove even more powerful than any OpenAI project.

New AI language model from Meta – will it pan out?

It is clear that the beginning of 2024 will not be the date for the creation of a full-fledged tool – an unrealistic date given how complex the ChatGPT model is. Then we can expect something along the lines of “beta projects” that will learn from their mistakes and deepen the data library. However, some tentative enthusiasm can already be seen in the stock market. As of 8.09 (Friday, two days before the news was reported), one META share was worth $297.89. On the other hand, the day after the announcement of the new project, or Monday 11.09, the value rose to $307.56. And this is not a temporary trend – the price oscillating around $310 per share means a five percent increase in value, which allows us to talk about the interest in the new project. We do not have such data for OpenAI, as the company is not listed. One thing we can be sure of – the bold declaration from Meta Platforms is certainly worth watching for developments. Perhaps later in 2023 we will see a response from OpenAI, which will only work out for us – users of language models. We will be returning to the topic, so keep an eye on our company blog!