​Sam Altman, head of OpenAI, says the main goal is to create “superintelligence”. AGI (Artificial General Intelligence) means the level of development in the field of artificial intelligence, when a system is able to perform any intellectual task that a human can perform. We’re still a long way from that point, but OpenAI says it’s possible, he writes Financial Times.

The company is working on Chat GPT-5, but has not said when the new generation of the chatbot might be released. The head of OpenAI says this new, ultra-high-performance version will need a lot more data to train on.

Microsoft is the largest investor in the company that created ChatGPT, with more than $13 billion, and Sam Altman says he wants to attract a lot of funding from Microsoft in the future to complete the task of creating a much more efficient AI model.

Altman says he spends his time in two categories: “researching how to create superintelligence” and how to gather enough computing power to do so. “Our vision is to build an AGI, see how we can make it safe and identify the benefits,” he told the Financial Times.

To train language models, OpenAI uses the most powerful processors on the market, Nvidia’s H100s, which ended up costing $40,000 apiece, are in huge demand and have boosted Nvidia’s value to over $1,000 billion.

Referring to the release of new versions of ChatGPT, Sam Altman says that more autonomous agents will be created that can perform various actions, such as writing code, making payments, sending emails or filling out various forms.

“We will make these agents more and more powerful, and their actions will become more and more complex in the future,” says the head of OpenAI.

What is AGI (Artificial General Intelligence) according to ChatGPT

AGI (Artificial General Intelligence) means the level of development in the field of artificial intelligence, when a system is able to perform any intellectual task that a human can perform. While most AI systems developed to date specialize in specific tasks or domains (e.g. speech recognition, natural language processing, strategy games), AGI must be able to solve a wide range of tasks, learn from experience, and adapt to new contexts in such conditions the way to the human mind.

It is important to note that at the moment most artificial intelligence systems are still far from reaching the level of AGI. They are often called limited artificial intelligence or specialized artificial intelligence because they are designed to solve specific tasks and are not capable of performing diverse and generalized actions like the human mind.

The development of AGI will involve solving serious problems related to understanding and modeling complex human cognitive capabilities, as well as managing the ethical and security issues associated with artificial intelligence at such an advanced level.