Google's Gemini, a potential AI rival to OpenAI's GPT-4, faces launch setback: Report

Google's highly anticipated large language model, Gemini, is facing delays in its launch, according to reports. Initially expected in November, the tech giant has informed cloud customers and partners to anticipate its release in the first quarter of the coming year. This postponement coincides with Google's efforts to bolster its consumer offerings with the new technology before opening access to outside developers.

Despite Google's strong overall growth, its third-quarter Cloud revenue fell short of expectations, in contrast to rival Microsoft's robust growth in its Azure cloud services. The delay in Gemini's release for cloud customers suggests a strategic move to solidify its consumer applications before external access.

High expectations for Google Gemini, but I also wonder if they’re being out-executed majorly & lack focus on this. Not hearing encouraging stories from Googlers and after Thanksgiving is a poor time to launch, so they have 1-2 weeks max or will have to delay to q1 (a very… https://t.co/6hPPtS2zA6

— Arun Rao (@rao_hacker_one) November 7, 2023

CEO Sundar Pichai, addressing questions about Gemini during a recent earnings call, highlighted its development as part of Google's DeepMind team's efforts. Pichai emphasised laying the foundation for the next generation of models set to launch throughout 2024. Google envisions Gemini being available in various sizes and capabilities, to be integrated internally before extending it to developers and cloud customers.

Gemini, designed to handle diverse tasks based on complexity, has undergone testing by external developers using smaller versions. However, the primary and largest iteration is still in the finalisation stage, with the goal of matching or surpassing OpenAI's advanced LLM GPT-4.

One significant challenge is ensuring that Gemini outperforms OpenAI's model, known for its advanced language processing capabilities. Google's cautious approach to Gemini's integration into Bard, a less capable LLM, stems from a previous factual error made during a demonstration.

Google plans to leverage Gemini in advertising applications, generating ad campaigns automatically. Initially focusing on static display ads, the model's capabilities could extend to audio and video ads. This move aligns with industry trends, with companies like Amazon and Meta Platforms unveiling AI-powered ad tools.

Gemini distinguishes itself with a longer memory of user interactions, a feature absent in LLMs like Palm 2. This extended memory could offer advertisers valuable insights for campaign performance comparison over time, enabling the development of new ad variants based on past successes.

As companies worldwide join the generative AI race, Gemini's delayed release underscores the competitiveness of the field. Its potential applications in advertising and memory capabilities position it as a noteworthy player in the evolving landscape of large language models. The tech community eagerly awaits Gemini's official launch and its contributions to Google's AI portfolio.