Google Unviels, its secret AI weapon could eventually Outshines ChatGPT
Google is reportedly working on an AI system called Gemini, developed by its DeepMind laboratory, that aims to rival or even surpass ChatGPT in terms of capabilities, according to Wired.
The developers plan to incorporate AlphaGo, an older AI program based on reinforcement learning, into the upcoming language learning model (LLM). AlphaGo learns through trial and error, continuously improving its performance by taking feedback from failed attempts. DeepMind intends to equip Google's future LLM with problem-solving and planning abilities, combined with a generative AI's capacity to extract and present information from the internet in natural language.
The goal is for Gemini to become the most advanced AI in the world, with DeepMind's co-founder and CEO, Demis Hassabis, asserting that it could be the most beneficial technology for humanity if executed correctly. The development of Gemini is currently underway and is expected to take several months, accompanied by a substantial investment from Google, potentially costing tens to hundreds of millions of dollars. For comparison, the development of ChatGPT alone cost over $100 million.
However, skepticism remains regarding Gemini's potential. The main concern lies with AlphaGo, which gained recognition in 2016 for defeating a champion player in the complex board game Go using reinforcement learning. It is unclear how AlphaGo's proficiency in a board game would directly translate into solving complex problems or generating content. Additionally, relying on trial and error for a generative AI raises concerns about the possibility of AI "hallucinations." While AlphaGo can aid Gemini's faster improvement, it is crucial to address any potential issues that may arise during its development.
Furthermore, Hassabis' statement about the relatively short development time frame is worrying. Google faced criticism when it swiftly released its AI-powered chatbot Bard shortly after the rise of ChatGPT in early 2023. Bard received backlash due to its misinformation and was labeled as "a pathological liar" and "worse than useless" by some employees. Extending the development cycle to years and providing more extensive training for Gemini could help avoid similar issues. It begs the question: Why rush the process?
.jpg)
Comments
Post a Comment