Google has announced that Gemini Pro is available in Bard. This is described as the biggest upgrade to Bard since it launched. Google has also released a Gemini API and other AI tools for developers.

Gemini is Google's multimodal large language model developed by Google DeepMind. It has been released in several versions as a rival to OpenAI's GPT-4, and a successor to Google's LaMDA and PaLM 2.

Gemini has received mixed responses, and has been described as no better than OpenAI's GPT- 3.5 for most tasks in a paper by a team of researchers from Carnegie Mellon University. Google says that there's a forthcoming more powerful version called Gemini Ultra due out early in 2024 that has scored higher than GPT-4 on Google’s internal research.

Meanwhile, a video released at the launch showing Gemini in use has been criticized as while it looked impressive, a closer examination revealed that the video wasn't quite what it appeared at first sight.

The video Hands-on with Gemini: Interacting with multimodal AI, has received millions of views. Google described it as "highlighting some of our favorite interactions with Gemini," and said it showed how the multimodal model can be flexible and responsive to a variety of inputs.

Multimodal means Gemini can react to language and visual cues. In the video, Gemini is shown a sketch of a duck, starting from a very basic sketch through to a finished drawing that Gemini suggests is an unlikely color for a duck (pale blue). Gemini is then shown a blue toy duck and responds "What the Quack!". Other elements of the video show Gemini playing rock paper scissors with the narrator and suggesting what might happen next in a cat video. All seemingly very impressive.

The clue to the complaints from CMU comes from the fact that while Google says "latency has been reduced and Gemini outputs have been shortened," a more honest accompanying blog post said that Google researchers captured footage of Gemini in action working on a wide range of challenges, then prompted it using still image frames from the footage, and prompting via text, rather than the impressive live interactions of the video.

Meanwhile, Google is keeping going, with the news that Gemini Pro is available via the Gemini API to developers in Google AI Studio, and to enterprises through Google Cloud's Vertex AI platform. Google is also introducing other models in Vertex AI, specifically an upgraded Imagen 2 text-to-image diffusion too, and a family of foundation models fine-tuned for the healthcare industry.

Gemini is also being added to Bard, Google's conversational chatbot that uses artificial intelligence. Bard will use a fine-tuned version of Gemini Pro, and the feature will only be available in English to begin with. Google says it plans to expand to different modalities and support new languages and locations in the near future. Google says Gemini Pro will help Bard be far more capable at things like understanding, reasoning and planning.

More Information

An In-depth Look at Gemini’s Language Abilities

Gemini In Bard

Gemini API

Related Articles

Google Adds Code Generation To Bard

To be informed about new articles on I Programmer, sign up for our weekly newsletter, subscribe to the RSS feed and follow us on Twitter, Facebook or Linkedin.

Comments

Make a Comment or View Existing Comments Using Disqus

or email your comment to: comments@i-programmer.info