It is the news of the moment. Google he just launched his response to OpenAI ChatGPT, long awaited by everyone. The failed test of Bard We are not going to tell it. The invention is called Gemini and it is, according to its developers, the most powerful artificial intelligence model ever created. Words that arouse a lot of expectation and excitement, but also a little vertigo.
The most enthusiastic are already talking about a new era. The "Gemini era«. The technology giant assures that This artificial intelligence model is going to be totally disruptive. Among other things, it has been trained to recognize, understand and combine different types of information: text, images, audio, video and even code.
Exaggerated? No, if we take as true the recent statements of the Google DeepMind CEO Demis Hassabis. He claims that Gemini possesses intelligence superior to any human brain in fifty different subject areas. This is apparently deduced from the previous tests that have been carried out in recent weeks.
Gemini will also have advanced programming skills. This includes the high quality code generation through an advanced system called AlphaCode 2. You will also be able to solve complex programming problems and assist the work of developers.
In any case, in Europe we will have to wait a bit to see if all this is true, since Google Gemini will not be available in EU member states due to current regulatory obstacles. How long will we have to wait? It seems like not much, really, since they are already working on a specific European legislation, (the so-called IA Law) to regulate the development and impact of this new technology.
The 3 models of Google's Gemini family
Following what Google explained during the Gemini presentation event, it seems that we are going to find three different models. Are these:
- Gemini Ultra, which will be the largest of all, intended for carrying out complex tasks.
- Gemini Pro, designed to carry out a broader range of tasks with an optimal level of performance.
- Gemini Dwarf, for use on mobile devices.
The truth is that a lot has been said about Gemini, although Google has refrained from delving too deeply into the details. We know, for example, that Google says that for its creation and during the testing phase, they have used the large Tensor Processing Units (TPU) v4 and v5e internal to the company. The heavy artillery, so to speak.
One of the most impressive news about this artificial intelligence tool is the development of an instrument (Gemini Nano) that will allow developers to create AI applications without the need for an internet connection, that is, with an additional guarantee of privacy. .
On the other hand, everything indicates that Gemini Pro will be available for developers and Google enterprise customers this month.
Try Gemini via Bard
Users in the United States can now test the new AI model through an adjusted version of Gemini Pro that has been implemented in the Google chatbot. We are talking about Bard, which was released in March of this year and was labeled a real failure. Now, thanks to Gemini, he gets a second chance.
In fact, Google is talking about a Bard update, the most ambitious that has taken place since its launch, which can serve as a gateway to Gemini for many users. It will be available in English in more than 170 countries and territories around the world. We have already explained that the citizens of the European Union will have to be a little more patient, what are we going to do.
The truth is that The entire Google universe is undergoing a spectacular transformation thanks to the introduction of Gemini. For example, a more modest version of the tool is already being implemented in the flagship of Google mobile devices, the Pixel 8 Pro smartphone. On the other hand, it is expected that in the short term it may also be present in other products and Google services such as Search, Ads and Chrome.
In addition to this, Google also announced at the same event the arrival of Cloud TPU v5p, a next-generation AI accelerator designed specifically for training cutting-edge artificial intelligence models. And the final step, one could say definitive, will be the landing of Gemini in the Google search engine. Its role will be to power your main web search engine and turn the process into a generative experience. That is something we are going to live.
In short, we are faced with a tool "more powerful than Chat00GPT", according to its creators, which comes with the intention of changing our world. While waiting for new news and to learn more about this exciting technology, we leave you a small appetizer in the form of a video: