Google's New AI Model Gemini: Revolutionary or Overhyped?
Summary:
Google unveiled its new AI model, Gemini, on December 6, claiming it to be the most advanced model, superior to OpenAI’s GPT-4. While Google's benchmarks demonstrated Gemini's performance, internet critics have questioned Google's assertions and promotional tactics. Some users have shared their dissatisfaction with Gemini, citing a preference for GPT-4. Despite these controversies, Google plans to launch Gemini more broadly to the public in early 2024.
On December 6, Google unveiled its newest artificial intelligence (AI) model named Gemini, touted as the most sophisticated AI model currently on the market, which is more advanced than OpenAI's GPT-4. Gemini, a multi-modal AI model, is designed to interpret and incorporate different types of data. It is available in three versions - Ultra, Pro, and Nano - to cater to various use cases. One area where Gemini outshines GPT-4 is in performing complex calculations and specialized coding tasks. Along with its announcement, Google revealed several benchmark tests that pitted Gemini against GPT-4, with the Ultra version of Gemini achieving top-notch performance in 30 out of 32 academic benchmarks. However, these claims have come under scrutiny over the internet, with critics questioning Google's methods and marketing strategies.
One such critic, who works in machine learning development, publicly questioned Gemini's proclaimed superiority over GPT-4. The critic argued that Google may be overstating Gemini's capabilities or selectively using examples of its success. Yet he concluded that Gemini could very well prove to be a strong contender against GPT-4 and lauded the competition in the AI space.
However, the same critic later posted that Google should feel "ashamed" of its promotional tactics, which he found misleading. His posts provoked responses from other users, some of whom expressed feeling tricked by Google's representation of Gemini. Others voiced their doubts over Google's assertion that Gemini would usher in the end of GPT-4, with one computer scientist labeling Google's depiction of Gemini as "dishonest."
Users also pointed out that Google's benchmarks included outdated versions of GPT-4, which made the comparisons irrelevant. Another concern involved the difference in prompts given to both models during their comparison. Users even suggested that the results could drastically change if Gemini's advanced model was compared to GPT-4's latest version, known as "turbo."
Instead of focusing on Google's benchmarks, some social media users shared their own experiences with Gemini and compared it to GPT-4. One such user, a web publisher and AI enthusiast, remarked that she was disappointed after using Gemini through Google's Bard tool and would continue using GPT-4 for now. She observed that Gemini/Bard refrained from answering political questions and made false claims about personal information.
Another user, involved in app development, asked both models to generate codes based on a picture using the same prompt. He shared his disappointment with Gemini/Bard's response compared to GPT-4.
Google has plans to launch Gemini more extensively to the public in early 2024 and to include the model in its suite of apps and services.
Published At
12/7/2023 4:25:26 PM
Disclaimer: Algoine does not endorse any content or product on this page. Readers should conduct their own research before taking any actions related to the asset, company, or any information in this article and assume full responsibility for their decisions. This article should not be considered as investment advice. Our news is prepared with AI support.
Do you suspect this content may be misleading, incomplete, or inappropriate in any way, requiring modification or removal?
We appreciate your report.