For several months, only one company has been regularly in the news: OpenAI. And for good reason, since it is the company that develops ChatGPT and Dall-E, two artificial intelligence tools that have met with huge success.
This logo, you must have seen it, if you have ever used or heard of ChatGPT. It’s actually the logo of OpenAI, the company that created ChatGPT and its language model, GPT. Who is behind OpenAI, one of the most powerful artificial intelligence companies in the world?
OpenAI, what is it?
OpenAI presents itself as “ an AI research and deployment companywith the aim of democratizing this technology. The company was originally a non-profit association, founded in December 2015, by two presidents: Sam Altman and the multi-billionaire Elon Musk, CEO of Tesla, Twitter and Neuralink among others, as well as several other members, all from the tech. Initially, the latter was to invest 1 billion dollars in total: he will not have invested ” that ” 100 million. The reason: after not good enough results according to him, Musk wanted to take control of OpenAI, which was refused by other founding members. He therefore left the board in February 2018.
In 2019, it became a capped-profit business: that is, the company can generate profits, but only at a certain level. A decision potentially due to the departure of Elon Musk and, by extension, of his investments.
The double interest of this status for OpenAI is to attract investors who want a return on the company, while limiting the potential excesses of the race for dividends. A ” capped profit“, but with a very high ceiling all the same: one hundred times the value of the initial investment.
How was the company created?
Sociologist Antoine Goujon tells the story of the creation of OpenAI in AOC . Elon Musk and Sam Altman commissioned Greg Brockman, former technical director of Stripe (an online payments company for businesses) to form a team of artificial intelligence specialists, mixing engineers and researchers.
To convince them, OpenAI, or rather its founding members, relied on the transparency of what was at the time an organization, particularly in terms of scientific research. At the beginning of 2017, OpenAI had 45 employees, but today that has changed a lot, since they are nearly 400.
The OpenAI philosophy: from openness to regulation
In its infancy, therefore, OpenAI wanted to be open (it’s in the name of the entity): publication of program source codes, results of scientific research disseminated, etc. But, since 2015, OpenAI has made a pivot in its philosophy: if the opening is not completed, it has suffered a brake.
The company recently acknowledged:we got our initial thinking about openness wrong, and we changed our minds: we no longer believe in releasing everything (although we do open source some things and plan to do so at the future), but rather that a way must be found to securely share access to the systems and their benefits.A philosophically important change, but intended to be secure: ChatGPT and Dall-E, the company’s two most important tools, can be used for unfair purposes. By their power, the responsibility of OpenAI is even greater.
Who funds OpenAI?
OpenAI’s starting cash was $1 billion, of which $100 million was brought in by Elon Musk and some by Amazon Web Services, the hosting and computing division.cloud computingfrom Amazon. But it is in 2019 that everything really accelerates: Microsoft then invests 1 billion dollars in what has become a company.
Then, with the insane success of ChatGPT, Microsoft and OpenAI extended their partnership in early 2023, with an investment from the Redmond firm of several billion dollars officially. The main rumor spoke of an amount of ten billion dollars, according to information fromSemafor. Microsoft would recover 75% of OpenAI’s profits, until it reimbursed these ten billion dollars. Although the firm does not own “that49% of the shares, 49% being to other investors and 2% to the non-profit foundation of OpenAI.
Today, however, we do not know precisely what is the shareholding of OpenAI. The valuation of this company is also not publicly known. But, in a $300 million seed round documented byTechCrunchin April 2023, the valuation had been assessed at between $27 billion and $29 billion.
The main artificial intelligence products of OpenAI
The popularity and power of OpenAI is mainly based on two AI tools: ChatGPT and Dall-E.
ChatGPT and GPT, its language model
Difficult to introduce it to someone for the first time as the chatbot has been talked about. What OpenAI has developed is more GPT, the great language model (or LLM in English), a sort of ChatGPT engine. The chatbot allows you to answer a whole bunch of questions, without really understanding what it is saying.
In fact, GPT is essentially an algorithm that predicts the word to follow according to a given context and according to previously generated words. This is why it cannot be said to have the reflection of a human being (or even of a living being).
Dall-E was especially talked about about a year ago, when Dall-E 2 was published, before seeing its popularity catch up with ChatGPT, but also with its competitors, such as Midjourney or Stable Diffusion . It is an image generation artificial intelligence program.
Using a textual description provided to him, Dall-E is able to generate images based on a massive database he has been trained with. For “to understand» the indications that we give him, he relies on GPT, a language model also used in ChatGPT.
Other OpenAI innovations
However, OpenAI is not a company known only for ChatGPT. We also owe him Microscope, an AI tool for visualizing neural networks, and Whisper, a voice recognition model.
Other artificial intelligence models are in development within the company. This is the case with Point-E, which allows you to generate 3D objects, and its successor, Shap-E.
Finally, we can mention Jukebox, described as “a neural network that generates music, including rudimentary vocals, as raw audio in a variety of genres and artist styles.The model was released in 2020, but nothing since.
What future for OpenAI?
With recent investments made by Microsoft and the trend towards artificial intelligence in the new technology industry, the future looks bright for OpenAI.
We can think that the company already has a next version of its language model used by ChatGPT in its boxes. Although at the moment OpenAI claims not to be working on GPT-5.
Towards general artificial intelligence
OpenAI has been a firm believer since the start in the “general artificial intelligence(AGI). We can define this term simply by an AI capable of learning and performing any task. At the moment, no AI model can be qualified as such, but the company is working towards it. She even imagines an AI that could be smarter than the human brain. But she sees it in a good light: the AGI would have “the potential to empower everyone with incredible new abilities; we can imagine a world in which each of us has access to help with almost every cognitive task, greatly multiplying the force of human ingenuity and creativity“, can we read in a recent blog post.
The company explains that “as we build more and more powerful systems, we want to deploy them and gain experience operating them in the real world.A statement which remains vague, but which makes it possible to understand that OpenAI wishes to integrate its tools and its AIs into the uses of humans to go beyond simple research and demonstration. OpenAI says it wants to be increasingly cautious in launching new AI models, especially since generative artificial intelligences will most likely be more heavily regulated by jurisdictions around the world.