DeepMind, Google's research lab, has recently announced its ambitious project of developing a large language model that aims to rival or even surpass the capabilities of OpenAI's ChatGPT. While ChatGPT has garnered significant attention worldwide, with its creator Sam Altman revealing a development cost of around $100 million, Google's new innovation poses a formidable challenge in the field of artificial intelligence. With their upcoming language model, DeepMind aims to push the boundaries and potentially outperform the capabilities of OpenAI's ChatGPT, setting the stage for an intriguing competition in the AI landscape.
In a historic event in 2016, an artificial intelligence program called AlphaGo, developed by Google's DeepMind AI lab, defeated a world champion Go player, marking a significant milestone. Now, Demis Hassabis, the co-founder and CEO of DeepMind, has revealed that the techniques used in AlphaGo are being utilized by his team to develop Gemini, an AI system that is anticipated to surpass OpenAI's ChatGPT. With this new endeavour, DeepMind aims to continue pushing the boundaries of AI and create an advanced system that could potentially outperform existing models in various applications.
DeepMind's upcoming large language model, Gemini, is being developed with the intention of being on par with GPT-4, the engine behind OpenAI's ChatGPT. However, DeepMind CEO Demis Hassabis has revealed that their team plans to go beyond language processing by incorporating AlphaGo-specific strategies into Gemini. This integration aims to enhance the system's capabilities with additional skills such as planning and problem-solving, setting it apart from existing language models. The combination of text-based AI technology with AlphaGo-inspired methodologies holds the potential to bring forth a new level of performance and versatility in AI systems.
Demis Hassabis expressed, “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.” Hassabis further added, “We also have some new innovations that are going to be pretty interesting.” At Google’s developer conference last month, when the company unveiled a number of new AI projects, Gemini was first alluded to.
DeepMind's AlphaGo is constructed using a groundbreaking technique called reinforcement learning, developed by the team at DeepMind. This approach involves training computer programs to handle complex decision-making tasks, such as playing video games or mastering the game of Go, through iterative learning and performance evaluation. Furthermore, AlphaGo employed a method known as tree search, which allowed it to explore and store potential moves on the game board. This combination of reinforcement learning and tree search played a vital role in AlphaGo's success and breakthrough achievement in defeating world-class Go players.
The next significant advancement in language models may potentially involve increased utilization of computers and the internet. Gemini, currently in development, is expected to take several months to complete, as mentioned by Hassabis. The project's extensive development process may require tens or even hundreds of millions of dollars in investment.
ALSO READ: Great camera smartphones under the mid-range category
Gemini, upon its completion, is anticipated to emerge as a formidable competitor to OpenAI's ChatGPT and other generative AI technologies. Google's recent entry into the space with AI tools and services, including the introduction of its generative AI system called Bard, may seem somewhat hurried, coinciding with the launch of ChatGPT. Notably, Google has also integrated AI into its search engine and various other services, showcasing its commitment to advancing AI capabilities across multiple domains.
In the quest for advancement, Google unified DeepMind, led by Demis Hassabis, with its primary AI lab, Brain, forming Google DeepMind. Hassabis stated that the team is focused on synergizing these two influential entities, which have played pivotal roles in driving recent AI progress.
Hassabis expressed, “If you look at where we are in AI, I would argue that 80 or 90 per cent of the innovations come from one or the other. There are brilliant things that have been done by both organisations over the last decade.”
According to reports, Hassabis has a notable track record of orchestrating AI breakthroughs that can bewilder the tech giants. However, on his previous occasion, Hassabis might have caused quite a stir and generated a frenzy around him.
ALSO READ: WhatsApp Pink Scam for Android- Mumbai Police issues red alert for fake version
