Title: DeepMind’s Gemini Algorithm Set to Outperform OpenAI’s ChatGPT
Demis Hassabis, the CEO of DeepMind, announced that their latest AI system, Gemini, will surpass OpenAI’s renowned ChatGPT. DeepMind gained international recognition in 2016 when their AI program, AlphaGo, defeated board game champion Ke Jie. Hassabis revealed that techniques from AlphaGo are being leveraged to develop Gemini.
Gemini, currently in the development phase, is a robust linguistic model similar to OpenAI’s GPT-4, which powers ChatGPT. However, DeepMind’s engineers are integrating alternative techniques employed in AlphaGo into Gemini. The objective is to provide the system with advanced capabilities such as planning and problem-solving for complex scenarios.
During Google’s developer conference last month, Hassabis described Gemini as the fusion of AlphaGo’s prowess with the remarkable language capabilities of other leading models. AlphaGo relies on DeepMind’s pioneering reinforcement learning technique, where software tackles challenging problems by selecting appropriate actions and learning from the subsequent outcomes.
The development of Gemini will span several months and entail a substantial financial investment, potentially costing hundreds of millions of dollars. Upon completion, DeepMind anticipates that Gemini will be Google’s response to the competitive threat posed by ChatGPT and other generative AI technologies.
Google, since the introduction of ChatGPT, has responded by creating its own conversational AI known as Bard. Additionally, generative AI has been incorporated into Google’s search engine and other products. As part of an effort to strengthen AI research, DeepMind and Google merged their respective units in April, forming Google DeepMind.
Meanwhile, Google DeepMind researchers are actively exploring diverse fields ranging from robotics to neuroscience. Most recently, the company showcased an algorithm capable of learning manipulation tasks using various types of robotic arms. This approach, drawing lessons from the physical experiences of humans and animals, is considered a crucial step in enhancing AI capabilities. Some experts argue that solely relying on text-based learning, as linguistic models do, poses significant limitations.
DeepMind’s Gemini algorithm not only aims to surpass ChatGPT but also to address these limitations by incorporating the broader understanding of the world obtained through physical experiences. The upcoming system holds great potential to revolutionize various domains of AI application, powered by its combination of linguistic and problem-solving abilities.
In conclusion, DeepMind’s Gemini algorithm, heavily inspired by AlphaGo, seeks to surpass OpenAI’s ChatGPT by bridging the gap between language models and physical understanding. With continued advancements in AI research, the boundaries of what these systems can achieve are being pushed further, promising a future of innovative and more capable AI technologies.
