DeepMind’s next AI model will be a serious competitor to ChatGPT

The ChatGPT chatbot may have caught the attention of the world’s public, but Google’s DeepMind Lab claims that their next big language model will be able to compete with or even surpass OpenAI’s model.

According to the Wired website, DeepMind is using AlphaGo techniques to create a new chatbot called Gemini Build to compete with ChatGPT. AlphaGo was the powerful artificial intelligence that was able to defeat a professional Go player for the first time.

If everything goes according to plan, Gemini will be able to plan, solve problems and analyze texts. “At a high level, you can think of Gemini as a combination of AlphaGo-like system capabilities and the amazing capabilities of large language models,” said Demis Hassabis, CEO of DeepMind. We also have some new innovations coming up that will be very exciting.”

Gemini DeepMind will be very powerful with reinforcement learning method

Wired reporter speculates that Gemini innovations in the field Reinforcement learning to be able to do things that today’s language models are not able to do. Reinforcement learning consists of a system that rewards artificial intelligence for performing desirable behaviors and punishes the system for undesirable behaviors.

DeepMind has gained good experiences with AlphaGo in the field of reinforcement learning, and it can undoubtedly have many uses of this technique in the field of generative artificial intelligence.

Gemini is not the first language model of DeepMind. Last year, this company created a chatbot named “Sparrow” (Sparrow) introduced in which various security measures were implemented in order not to give inappropriate answers to users’ questions. But Gemini is considered their most ambitious project in this field and can have something to say against ChatGPT.

Considering the great potential of the productive artificial intelligence market, companies are doing their best to have a greater share of this market. It is said that the value of this market will probably reach about 110 billion dollars by 2030.