Four keys to understanding GPT-4, a new version of artificial intelligence

GPT-4 It’s official. After strong growth chat At the end of 2022 and the emergence of this artificial intelligence on multiple platforms, OpenAI, the company that owns the project, announced a new version promising improvements in its performance.

Since 2020, the third version has been launched, but after the enormity of recent months, its update was highly anticipated by users and developers, especially after the investment it received from Microsoft and integration into the Bing search engine.

It may interest you: ChatGPT extension for Chrome is stealing information

The first key to understanding the significance of this new release is to understand exactly what it is about.

chat It is a program that provides answers and texts to requests, but there is technology behind it and that is what makes it work.

GPT is the system trained with data and algorithms to respond to requests in any language, with the ability to hold a conversation, make recommendations, provide information, create documents, and a long list of options that adapt to the tool.

So the arrival of a new version means that these platforms will have more advanced technology for their various functions. Especially in sections such as subtlety, complexity, naturalness, and proximity to what a human would do.

It may interest you: How to install ChatGPT on Apple Watch

currently, Open AI He was careful about the information he provided, leaving it to the developers to discover the system’s capabilities and differences compared to its previous release.

An important point in this context is the parameters, which are the amount of information that feeds into the learning. However, the company considers that it is not necessary to radically expand this number to make an important advance, but rather the essence lies in its work.

“In the future, many people will be surprised at how powerful AI can be without increasing standards,” said Sam Altman, CEO of the company.

You may be interested: ChatGPT: This is how you can create contracts, articles, and emails with AI

to achieve your improvement, GPT-4 It will allow a multimedia language, that is, requests will not only be made through text, but can also be associated with images, videos, and audio. Something that would work to summarize a YouTube video, for example.

Another important aspect is the accuracy of the information, as many of these chatbots have started to invent facts and data. But from the company they confirmed that this new version is “82% less likely to respond to requests for illegal content and 40% more likely to provide feedback”.

Since it is a technology that has come out recently, there are few options to use it. Initially, those who pay for ChatGPT Plus will have access to this new version, though for a limited time while the developers tweak the details of its operation.

In case you are a developer and want to bring this release to a platform, there is already a waiting list Open AI to get the API for this language model.

At first, everything that can already be done with current technology: order summaries, recipes, recommendations, search for information or start a conversation.

But part of the key will be in multimedia language and user creativity. For example, there will be an option to create texts for a video on social networks, and that there are more details in each element, by combining texts and images.

Overall, this technology will make requests more creative, higher quality, more accurate, and add variety to get closer to what a human would do in this case.