AlphaGo, an artificial intelligence program developed by Google’s DeepMind AI unit, made history in 2016 when it defeated a world-class Go player. Since then, the business has been working on a new project dubbed Gemini that attempts to bring together the linguistic capabilities of big models with the advantages of systems akin to AlphaGo to create an AI system that is more capable than that powering OpenAI’s ChatGPT. The new system will have new abilities like planning or the capacity to solve issues, according to Demis Hassabis, cofounder and CEO of DeepMind, making it a prominent player in the cutthroat AI market.
Gemini, which is currently in development, is a large language model that works with text and is similar in nature to GPT-4, which powers ChatGPT. However, unlike ChatGPT, the team behind Gemini is combining the technology with techniques used in AlphaGo, such as tree search to search and record possible moves on the board, and reinforcement learning, in which software learns to conquer issues that need decision-making on what actions to take, like in Go or video games. The combination of these two techniques is expected to result in an AI system that is more capable than any other system currently available.
Gemini is currently being developed, and it will likely take several months. According to Hassabis, the project might cost tens or hundreds of millions of dollars. However, when it is finished, it is anticipated to play a significant part in Google’s response to the danger that ChatGPT and other generative AI technology offer to its market share. The search company invented many of the methods that made it possible for the current onslaught of innovative AI concepts, but they carefully developed and implemented the resulting products.
Google has quickly released its own chatbot, Bard, and integrated generative AI into its search engine and many other products since ChatGPT’s launch. In April 2021, the business integrated Google’s principal AI lab, Brain, with Hassabis’ DeepMind section to speed up AI development. According to Hassabis, the new team would combine two titans who have contributed significantly to recent advancements in AI.
Sam Altman, OpenAI CEO, said in April that creating GPT-4 cost more than $100 million. However, the cost of developing Gemini is expected to be higher than that. According to Altman, the age of giant AI models is already over. To stay competitive, companies like Google need to invest heavily in research and development to create the next generation of AI systems. Gemini is Google’s answer to that challenge, combining the best of AlphaGo and ChatGPT to create an AI system that is more capable than either of them.
Reinforcement learning is a form of machine learning that involves learning from feedback to take actions that maximize rewards. In the case of AlphaGo, the program learned to play the game of Go by playing against itself and using reinforcement learning to improve its performance. The same technique was used to develop Gemini, which is still in development.
Tree search is a method used in computer science to explore and remember possible moves on a board. In the case of AlphaGo, the program used tree search to explore the possible moves on the board and remember them for future use. Gemini is expected to use the same technique to explore and remember possible solutions to problems.
A large language model is a type of artificial intelligence system that can generate human-like text. GPT-4, which powers ChatGPT, is an example of a large language model. Gemini is also a large language model, but it is expected to be more capable than GPT-4 due to the combination of language capabilities with AlphaGo-type systems.
ChatGPT is an AI system developed by OpenAI that can generate human-like text. It is powered by GPT-4, which is also a large language model. However, Gemini is expected to be more capable than ChatGPT due to the combination of language capabilities with AlphaGo-type systems.
Since the debut of ChatGPT, Google has been working on developing its own AI systems to compete with it. One such system is Bard, a chatbot that was rushed out to battle ChatGPT. Google has also put generative AI into its search engine and many other products. However, the development of Gemini is expected to be the company’s strongest response to ChatGPT yet.
Gemini is just one of many AI systems that are currently in development. As the field continues to evolve, it is likely that we will see many more breakthroughs in the coming years. The combination of language capabilities with AlphaGo-type systems is just one example of the new capabilities that AI systems are expected to have in the future. As companies like Google continue to invest in research and development, we can expect to see even more powerful AI systems in the years to come.
In conclusion, Google’s DeepMind is working on a new AI system called Gemini, which is expected to be more capable than any other system currently available. The system combines the language capabilities of large models with the strengths of AlphaGo-type systems, resulting in an AI system that is capable of planning and problem-solving. Although the system is still in development and is expected to cost tens or hundreds of millions of dollars, it is expected to be a major player in Google’s response to ChatGPT and other competitive threats through generative AI technology. As the field of AI continues to evolve, we can expect to see many more breakthroughs in the years to come.
First reported by WIRED.
Olivia is the Editor in Chief of Blog Herald.