Google’s bet to compete with Open AI

For decades, Google has been a steward and pioneer of neural networks and machine learning, two fundamental elements of so-called artificial intelligence (AI). With Gemini, the digital giant is now presenting a very interesting scenario in front of an interesting battle to take over the global AI standard.

Following the success that Open AI has had with ChatGPIT and now Sora, honestly, apart from being two of the best AI architectures, it also has better media performance, Google sees the need to respond strongly and quickly to the battle skillfully fought by its rival with Microsoft.

(Also read: A new chapter in AI: The race for the best model between Google and ChatGPT)

And Gemini 1.5 is all that, a structural piece of the Lego of AI that completes Google’s offering with this update, along with Gemini Nano, Gemini Pro, Gemini Ultra, and Gemini Advanced, a suite of cloud services and mobile apps. The platform, which is free or under a subscription model, aims to provide generative and business AI experiences.

The new Gemini 1.5, currently in closed testing with developers, uses the MoE (Mixer-of-Experts) network, Which divides the model into smaller neural networks that activate according to the type of information given and requested, reducing latency and improving processing efficiency. Google clients are already testing it through AI Studio and Vertex AI, AI development and implementation platforms for businesses.

OpenAI is definitely not standing still. last week he showed Sora, your generative AI is capable of creating hyper-realistic videos with a simple gesture or text request. Although it is not yet open to everyone, the first tests surprised the whole world by the quality, sharpness and resolution of the generated content: except for a few details, which will of course be corrected, they are presented as if produced by AI. Difficult to identify.

(You may be interested in: What changes to Google Bard AI when it becomes a Gemini?)

How and why to use these AIs?

currently, Both ChatGPT and Gemini are free to use as generative artificial intelligence. That is, they can be asked about an infinite number of topics in different languages ​​and with different types of input (they can be given a text or context) so that they can generate very relevant answers.

In our experience, ChatGPT is more natural in its behavior towards the user, whereas Gemini provides relevant links from which to derive information to generate their responses.

If the purpose of using AI is for simple tasks, work, study or basic needs on a daily basis, both options are sufficiently powerful and useful, with natural care to be taken, because It is possible that ‘hallucinations’ or data that are not entirely true or concrete slip into their answers.

Request a summary of a text, create new editions, translations, recipes in another format or narrative style, recommendations on various topics, act as an expert in a profession or craft to explain topics, create documents, etc., Everything can be achieved adequately with ChatGPT and Gemini in the free versions.

Now, for more business uses, where both platforms offer advanced processing capabilities along with a variety of additional services, Gemini Advanced and ChatGPT Plus are an option for $20 per month.

In the case of Gemini, it includes everything that Google One brings with two terabytes of cloud storage and integration with workspaces (Docs, Gmail, Sheets, Slides), for the moment in English. The conversational chatbot comes in 40 languages.

in open ai ChatGPT Plus includes ChatGPT-4 and Dall-E, the first with a more powerful, efficient AI model with the ability to ‘understand’ images; While the other is an AI that generates images from text commands.

Copilot Pro is a good option because it includes everything from integration with ChatGept Plus plus Microsoft 365 productivity software to using AI directly in Excel, Word, Outlook, and PowerPoint through an annual or monthly subscription.

(Also read: Google introduces Gemini 1.5 AI model: These are its improvements)

Sora, a surprise

It is an AI from Open AI, similar to ChatGPT and DALL-E, which is capable of creating hyperrealistic videos with text (prompt) given as input.

What videos can you make? Any. Whatever comes to mind, real, futuristic, invented, anything, with stunning realism and visual quality, maximum resolution.

Can it be used now? It is not open to the public yet. Open AI testing behind closed doors. Last week they showed pre-made videos and they went viral.

what’s your News? Apart from the impressive quality and authenticity of the videos, the power of content creation is their natural language understanding. The better and more clearly, with many details, we ask for the video, the more he will produce it with utmost quality.

What will it be used for? Sora has wide versatility in use for creative industries, research, content development, education, communication, etc.

When will it be available? There is no exact release date for the public yet, but Open AI notes that it will be openly available soon.

Jose Carlos Garcia R.

multimedia editor

@JoseCarlosTecno on the network