Shortly after launching Gemini, the big language model that Google hopes to bring to the top of the AI industry, the company is already announcing its successor. Today, Google is launching Gemini 1.5 and making it available to developers and enterprise users before a full rollout to consumers. The company is positioning Gemini as a business tool, a personal assistant, and a tool for everything that lies in between.
Gemini 1.5 Pro, Google’s general-purpose model, is apparently as good as Gemini Ultra, which the company recently released, and outperforms Gemini 1.0 Pro by 87% in tests. The model was created using a technique known as “Mixture of Experts” or MoE. This means that it only runs a portion of the overall model when you send a request, rather than processing the request with all of its power all of the time. This approach should make the model faster to use and more efficient for Google to run.
Gemini 1.5 has a huge context window, which means the model can handle much larger queries and look at much more information at once. That window is a whopping 1 million tokens, compared to 128,000 in GPT-4 and 32,000 in the current Gemini Pro. Google CEO Sundar Pichai explains that: “That’s about 10 or 11 hours of video, tens of thousands of lines of code.” The size of the context box means you can ask the AI bot about all that content at once. Pichai also said that Google researchers are testing a contextual window with 10 million tokens – that’s like having all the episodes of Game of Thrones at once.
Sundar Pichai also believes that a large contextual window will be very useful for businesses. “It’s useful in cases where you can add a lot of personal context and information at the time of query.” For example, filmmakers can upload their entire movie and ask Gemini what the reviewers have to say. Companies will be able to use Gemini to view masses of financial records at the same time. “I consider this one of the biggest breakthroughs we’ve made”.
For now, Gemini 1.5 will only be available to business users and developers through Google’s Vertex AI and AI Studio. It will eventually replace Gemini 1.0, and the standard version of Gemini Pro, available to everyone at gemini.google.com and in the company’s apps, will be 1.5 Pro with a contextual window of 128,000 tokens. You’ll have to pay extra to get a million. Google is also testing the model’s security and ethical boundaries, particularly regarding the new larger contextual window.
Google is now in a race to build the best artificial intelligence tool as companies around the world try to determine their own artificial intelligence strategy: enter into agreements with OpenAI, Google, or someone else. Just this week, OpenAI announced memory for ChatGPT and is about to introduce its own web search. So far Gemini seems amazing, especially for those already in the Google ecosystem, but there’s still a lot of work ahead on all sides.
Eventually, all those 1.0 or 1.5 versions, pros and ultras won’t matter to users. “People will just consume the experience. It’s like using a smartphone without paying attention to the processor in it,” says Sundar Pichai. But right now, he says, we’re still at a stage where everyone knows the chip in their phone because it matters.
Last year, Google launched its artificial intelligence-based chatbot Bard, which it has now decided to rename Gemini. This is the same name of the artificial intelligence model that the chatbot is based on.
Gemini is already running using the Pro 1.0 model in over 40 languages and more than 230 countries and territories. Now Google is introducing two new features to better engage with AI: Gemini Advanced and a mobile app.
The Gemini Advanced chatbot gives you access to Google Ultra 1.0’s advanced AI model. This makes it much better able to handle highly complex tasks such as coding, logical thinking, following detailed instructions and collaborating on creative projects. Gemini Advanced not only allows for longer and more detailed conversations; it also better understands the context of prior cues.
Gemini Advanced can become a personal tutor – creating step-by-step instructions, sample tests or two-way discussions customized to the user’s learning style. It can help with more advanced coding scenarios, serving as a source of ideas and helping to evaluate different coding approaches. Gemini Advanced can help digital content creators by generating new content, analyzing the latest trends, and brainstorming the best ways to grow audiences.
Gemini Advanced users will have access to enhanced multimodal capabilities, more interactive coding features, deeper data analysis capabilities, and more. Today, Gemini Advanced is available in more than 150 countries and in English; it will eventually become available in other languages.
Gemini Advanced is available as part of the new Google One AI premium pricing plan. It provides all the features of the current Google One Premium plan, as well as offering access to Gemini Advanced, and soon AI Premium subscribers will additionally be able to use Gemini in Gmail, Docs, Presentations, Tables, and more.
Additionally from now on, Google is starting to roll out mobile experiences for Gemini and Gemini Advanced with new apps on Android and Google on iOS.
With Gemini on your smartphone, you can type, talk, or add images for a variety of on-the-go assistance. For example, you can take a picture of a flat tire and get repair instructions, create a customized image for a dinner invitation, or ask for help writing a complicated text message.
By downloading the Gemini app or connecting through Google Assistant, users will be able to access it from within the app or anywhere Google Assistant is normally activated. Once launched, a new overlay window will open, providing access to Gemini as well as contextual help right on the screen. With this, you can create a caption for a photo or ask a question about an article. Many of the features of the Google Assistant voice assistant will be available through the Gemini app, including setting timers, making calls, and controlling smart home devices.
Google One AI is offered as a free trial for the first two months, and then will cost $19.99 per month. On iOS, access to Gemini will be available on the Google app in the coming weeks. All you’ll have to do is tap the Gemini Switch and start interacting with it.
Gemini will gradually appear on Android and iOS phones in the US in English today, and will become fully available in the coming weeks. Starting next week, you’ll be able to access it in more regions in English, Japanese and Korean, with other countries and languages to be added soon.
Source: The Verge Link