Two months have passed since the launch of Gemini, a massive model of a language Google is hoping will take its position at the forefront of AI market The company is now unveiling Gemini 1.5, its successor. Google has launched Gemini 1.5 today, and is will be making it available to both enterprise and developers prior to a complete release to consumers scheduled for later. Google has made it clear that it’s all in to Gemini as a tool for business, a personal assistant and everything else between, and is making strides in that direction.
There are numerous improvements to Gemini 1.5 1. Gemini 1.5 Pro, which is the general-purpose model of Google’s system, appears to be in line with the premium Gemini Ultra that the company has only recently introduced and has beaten Gemini 1.0 Pro in testing for benchmarks at 87 percent. The model was created with a method that’s becoming more popular called “Mixture of Experts,” or MoE which means that it only runs a small portion of the overall model whenever you make a request instead of processing all of the data throughout the entire time. ( Here’s a good explanation of the topic.) This method should improve the speed of the model to you to use and better for Google to use.
There’s a new feature in Gemini 1.5 that makes the entire company, including the CEO Sundar Pichai, who is particularly excited with the news that Gemini 1.5 features a massive context window. That means it is able to handle more complex queries and display many more details at once. The window can handle 1 million tokens compared to 128,000 in GPT-4 from OpenAI and 32,000 in the Gemini Pro. Gemini Pro. Tokens can be difficult to grasp ( here’s a good explanation) So Pichai simplifies the process: “It’s about 10 or 11 hours of video, tens of thousands of lines of code.” The context window lets you inquire with the AI bot to answer questions about all the content in one go.
(Pichai adds that Google’s research team are currently testing a 10 million token contextual window — which is, like, the entire series of Game of Thronesall at one time.)
While he’s explaining the situation to me Pichai says in a casual way that it’s possible to put the whole Lord of The Rings trilogy within that window. This is too specific to me, so I ask him what he thinks about this: it’s already happened already, isn’t it? Someone from Google is just looking to check whether Gemini detects any continuity issues and trying to comprehend the intricate Middle-earth’s history and see if AI is able to finally understand the story the Tom Bombadil. “I’m sure it has happened,” Pichai declares laughing, “or will happen — one of the two.”
Pichai is also of the opinion that the larger context window could be extremely beneficial for companies. “This allows use cases where you can add a lot of personal context and information at the moment of the query,” Pichai states. “Think of it as we have dramatically expanded the query window.” He envisions that filmmakers could upload their entire film and inquire with Gemini what reviews might have to say and he can see businesses making use of Gemini to examine huge amounts of financial documents. “I view it as one of the bigger breakthroughs we have done,” the CEO states.
As of now, Gemini 1.5 will only be available to users of business and developers through the Google’s Vertex AI and AI Studio. It will eventually replace Gemini 1.0 and the basic Version that comes with Gemini Pro — the one that is available to everyone on gemini.domain.com and in Google’s apps it will become 1.5 Pro with a 128,000-token context window. You’ll need to pay an additional fee for the million. Google is conducting tests on the system’s security and ethical limits, especially with regard to the new context window.
Google are in a fast race to create the most effective AI tool currently in the midst of a time when businesses all over the world attempt to decide on what their AI strategy and whether or not to sign agreement to OpenAI, Google, or another. In the last week OpenAI released “memory” for ChatGPT and appears to be preparing for the launch of internet search. To date, Gemini seems to be remarkable, particularly for users already part of Google’s ecosystem However, there’s still a lot to be done from all angles.
In the end, Pichai tells me, all these 1.0s and 1.5s and Ultras and Pros and conflicts won’t matter much to the users. “People will just be consuming the experiences,” Pichai declares. “It’s like using a smartphone without always paying attention to the processor underneath.” At the moment the tech expert says we’re in a phase that everyone is aware of the chip within their phone, as it’s important. “The underlying technology is shifting so fast,” He states. “People do care.”
