Google’s research lab, DeepMind, is building a language model that can compete with OpenAI’s ChatGPT. Google’s new innovation poses a significant challenge in artificial intelligence, after ChatGPT’s developer Sam Altman revealed a nearly $100 million development cost. DeepMind’s language model may surpass OpenAI’s ChatGPT, setting up a fascinating AI rivalry.
In 2016, Google’s DeepMind AI lab’s AlphaGo defeated a world champion Go player. Demis Hassabis, DeepMind’s co-founder and CEO, has stated that his team is using AlphaGo’s methods to construct Gemini, an AI system that will beat OpenAI’s ChatGPT. DeepMind hopes to develop AI and create a system that might surpass existing models in numerous applications with this new project.
DeepMind’s Gemini big language model aims to match OpenAI’s ChatGPT’s GPT-4 engine. However, DeepMind CEO Demis Hassabis stated that Gemini would use AlphaGo-specific tactics to go beyond language processing. This integration adds planning and problem-solving skills to the system, distinguishing it from language models. Text-based AI technology and AlphaGo-inspired methods might improve AI performance and adaptability.
“At a high level, you can think of Gemini as combining some of the strengths of AlphaGo-type systems with the amazing language capabilities of the large models,” said Demis Hassabis. Hassabis continued, “We also have some new innovations that will be pretty interesting.” Last month’s Google developer conference revealed many new AI initiatives, including Gemini.
DeepMind’s reinforcement learning approach builds AlphaGo. Iterative learning and performance evaluation educate computer algorithms to make complicated decisions like playing video games or Go. AlphaGo used tree search to investigate and store game board movements. Reinforcement learning and tree search were crucial to AlphaGo’s breakthrough victory against world-class Go players.
Language models may use computers and the internet more in the future. Hassabis expects Gemini, presently in development, will take months to finish. The project’s extended development may cost tens or hundreds of millions of dollars.
Gemini, when completed, will challenge OpenAI’s ChatGPT and other generative AI technologies. Google’s recent foray into AI tools and services, notably its generative AI system Bard, may appear rushed given ChatGPT’s debut. Google has incorporated AI into its search engine and other services, demonstrating its commitment to AI advancement across several sectors.
Google merged its Brain AI unit with Demis Hassabis’ DeepMind to progress. Hassabis said the team is focusing on synergizing these two key elements, which have driven recent AI advancement.
Hassabis stated, “If you look at where we are in AI, I would argue that 80 or 90% of the innovations come from one or the other. Both organizations have done great things in the past decade.”
Hassabis is known for engineering AI advances that confound tech companies. Hassabis may have made a stir and a commotion before.
Conclusion:
Google’s DeepMind AI lab is developing a language model called Gemini It aims to match OpenAI’s GPT-4 engine but incorporates AlphaGo-specific tactics for planning and problem-solving. The project aims to surpass OpenAI’s ChatGPT and create a fascinating AI rivalry. DeepMind’s co-founder and CEO, Demis Hassabis, believes that Gemini will combine AlphaGo-type systems with large models’ language capabilities, improving AI performance and adaptability.
The project’s reinforcement learning approach, which uses iterative learning and performance evaluation, is expected to take months to complete and may cost tens or hundreds of millions of dollars. Gemini, when completed, will challenge OpenAI’s ChatGPT and other generative AI technologies. Google’s commitment to AI advancement across various sectors is evident in its recent integration of its Brain AI unit with DeepMind, with both organizations contributing to recent AI advancements.
