Google is developing an artificial intelligence model that can imitate the ability of people to “think” by analogy with OpenAI o1 models. This was reported by Bloomberg.
As with OpenAI, Google is looking for the next big breakthrough in artificial intelligence. Models that will be able to make a logical chain of thoughts are exactly what they need. Over the past few months, many teams within the company have been actively working on an AI model that will be able to solve multi-step math and programming problems. According to sources familiar with the matter, they are making progress.
Such models will take longer to answer because they will process a number of related queries to the user’s question. After a few seconds of processing all these queries, the AI will summarize each result to give the best and most accurate result possible.
Bloomberg also writes that Google and OpenAI are actively competing in the AI market, but the new startup is still leading the race. Some employees of the DeepMind department within Google were worried after the release of the o1 models and thought that they had fallen far behind the competition. However, for the most part, the reactions were much calmer than after ChatGPT, as the company had already released several of its own AI products.
It is noted that several important steps have already been taken within the company to regain leadership. In particular, the leading research laboratories were merged to create the DeepMind division. Google is also strengthening relations between research and development teams. But despite this, the company still takes longer to bring new products to market, in part because of ethical concerns and the need to meet public expectations and trust, and other related factors.
Loading comments …