All about OpenAI, from the birth of ChatGPT to the creation of an AI that is equal to humans

In less than 10 years, OpenAI has managed to become a leading company in the field of new forms of artificial intelligence, with ChatGPT, its conversational chatbot, as one of its best-known products.

During its short history, the company led by Sam Altman has counted among its investors – and founders – such important names as Elon Musk, although he left three years after its creation.

However, other multinationals such as Microsoft have maintained a close relationship with OpenAI, creator of ChatGPT and DALL-E.

How OpenAI was born

The origin of OpenAI, which is barely 10 years old (it was founded in 2015), has a lot to do with the concerns of the technology sector about the possibility of the existence of uncontrolled artificial intelligence.

Thus, several industry celebrities, including Elon Musk and Sam Altman, decided to join forces in a non-profit project, despite the fact that Musk decided to go his own way in 2018 in the face of the potential risks of this AI.

The year after Musk left, OpenAI pivoted to a hybrid model, maintaining its commitment to being a nonprofit but with a different focus that would attract investors with lucrative returns.

This was because GPT, the model that OpenAI was investigating, required numerous resources —very expensive— at a computational level, even more so when it comes to the search for general artificial intelligence, which will equal human thought.

Throughout 2018, OpenAI launched the first GPT model, although its first successes consisted of demonstrations of a bot in video games, such as the legendary Dota 2similar to League of Legends.

However, there were still many surprises in store, and OpenAI saved the best for last, bringing one of its products, ChatGPT, to the general public in a massive way.

The arrival of ChatGPT and DALL-E

By 2020, OpenAI’s AI model had already achieved great successes, such as writing quite remarkable journalistic texts, in a competition against professionals in the New York Times.

This media event was followed by the Introducing PlayGround, a feature that preceded ChatGPT —and continues to this day—, but is focused on developers who want to tinker with the APIs of the OpenAI model.

However, that tool was not aimed at the general public, to which Altman’s company responded by launching ChatGPT, a conversational chatbot that imitated natural language, as well as being able to respond based on it.

In November 2022, ChatGPT saw the light of day and, In just one month, the chatbot had 1 million users; Just 2 years later, ChatGPT has refined its model, with versions that have improved its capabilities.

Now, several experts signed a letter in March 2023 expressing their concerns about the next generations of learning models, calling for development to be halted for at least 6 months to study the risks.

Musk himself signed this letter addressed to OpenAI, in which the following was argued:“We are calling on all AI labs to immediately pause training systems more powerful than GPT-4 for at least 6 months. AI systems are coming close to competing with humans on general tasks.”

But ChatGPT is not alone, as another of OpenAI’s best-known products is DALL-E —in addition to Sora for video—, with different versions that allow generating artificial images based on a prompt or indication in natural human language.

This tool is based on the GPT-3 multimodal model, which was launched in 2020. More colloquially, it allows the interpretation of different data, such as images, texts or videos. Although there is still one more project in OpenAI, called Strawberry.

Strawberry, OpenAI’s secret project

Strawberry, formerly known as the Q* project —pronounced “Q star”— is one of OpenAI’s best-kept secrets.

According to what he shared Reutersit seems that OpenAI is testing an AI model that is capable of answer various questions about mathematics and other sciences at the same intellectual level as a human being.

This is what is known as artificial general intelligence or AGI, which aims to reason and interpret information in the same way your brain would, with complex neural networks.

So far, nothing more is known about the Strawberry project, beyond a warning letter about the dangers of the project written by a team of company scientists, according to a document to which this agency had access.

How and who funds OpenAI

On your own Web pageOpenAI already warns that any investment in its business is “high risk” and that, therefore, “investors could lose the capital invested, without seeing any return.”

One of the company’s biggest challenges is to reduce the cost of researching everything related to artificial general intelligence (AGI) and achieve a profitable service in order to be able to return the money.

In one way or another, OpenAI has great expectations among investors, since it has multinationals behind its financing such as Microsoftwith which it has maintained a strategic partnership — independently — since 2019.

In this regard, according to the organizational chart shared on the OpenAI website, both its non-profit organization and the additional entity remain separate from Microsoft.

In addition to the one led by Satya Nadella, there are many others who have shown interest in the future of OpenAI, such as Nvidia, Apple or Thrive Capital; as of the date of this article, it appears that OpenAI to get $150 billion valuation thanks to these.

For now, it is worth noting that OpenAI continues to operate independently, and not even Altman, its CEO, has direct capital.

Meet How we work in ComputerHoy.

Tags: Artificial Intelligence, Elon Musk